Highislandbook web 2014 by Vermillion

High Island (Ardoileรกn), Co. Galway: Excavation of an Early Medieval Monastery

ARCHAEOLOGICAL MONOGRAPH SERIES: 10

Frontispiece The monastery on High Island following conservation, viewed from the north-east (Photographic Unit, NMS).

High Island (Ardoileán), Co. Galway: Excavation of an Early Medieval Monastery

Georgina Scally

with specialist contributions by Jacqueline Cahill-Wilson, Claire Cotter, Ian Fisher, Rowena Gale, Michael Kenny, Stephen Lancaster, Christine Maddern, Margaret McCarthy, Meriel McClatchie, Brian McConnell, Clare McCutcheon, Barra O’Donnabhain, Sara Pavía, Tim Young.

Site drawings were prepared for publication by Martin Halpin. Artefact and cross-slab drawings are by Patricia Johnson.

Price: €35

Department of Arts, Heritage and the Gaeltacht Archaeological Monograph Series GENERAL EDITORS—ANN LYNCH AND CONLETH MANNING 1 2 3 4 5 6 7 8 9

Excavations at Roscrea Castle (Conleth Manning ed.) St Audoen’s Church, Cornmarket, Dublin: archaeology and architecture (Mary McMahon) Kells Priory, Co. Kilkenny: archaeological excavations by T. Fanning and M. Clyne (Miriam Clyne) The history and archaeology of Glanworth Castle, Co. Cork: excavations 1982–4 (Conleth Manning) Tintern Abbey, Co. Wexford: Cistercians and Colcloughs. Excavations 1982–2007 (Ann Lynch) Trim Castle, Co. Meath: excavations 1995–8 (Alan R. Hayden) Parke’s Castle, Co. Leitrim: archaeology, history and architecture (Claire Foley and Colm Donnelly) Clogh Oughter Castle, Co. Cavan: archaeology, history and architecture (Conleth Manning) Poulnabrone: An early Neolithic portal tomb in Ireland (Ann Lynch) These monographs are subject to international peer review.

© Government of Ireland, 2014, subject to the moral rights of the individual authors as established under the Copyright and Related Rights Act 2000. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any electronic, mechanical or other means known or hereafter invented, including photocopying or recording or otherwise without either the prior written consent of the copyright holders or a licence permitting restricted copying in Ireland issued by the Irish Copyright Licensing Agency Ltd., The Writers’ Centre, 19 Parnell Square, Dublin 1.

BAILE ÁTHA CLIATH ARNA FHOILSIÚ AG OIFIG AN tSOLÁTHAIR Le ceannach díreach ó FOILSEACHÁIN RIALTAIS, 52 FAICHE STIABHNA, BAILE ÁTHA CLIATH 2 (Teil: 01-6476834 nó 1890 213434; Fax 01-6476843) nó trí aon díoltóir leabhar. __________ DUBLIN PUBLISHED BY THE STATIONERY OFFICE To be purchased from GOVERNMENT PUBLICATIONS, 52 ST. STEPHEN’S GREEN, DUBLIN 2 (Tel: 01-6476834 or 1890 213434; Fax: 01-6476843) or through any bookseller. ISBN: 978-1-4064-2828-5 Copy-edited by Rachel Pierce at Verba Editing House Designed and typeset by Vermillion Design Cover design by Design for Market Printed by Castle Print

Contents

Acknowledgements 7 1. High Island 1.1 Introduction 9 1.2 Location 10 1.3 Topography and geology 10 1.4 Landing places 12 1.5 Flora and fauna 14 2. High Island and the Cult of St Féichín

Pádraic Moran

3. The Later History of High Island and its Antiquarian and Archaeological Rediscovery 3.1 Roderic O’Flaherty’s account, 1684 3.2 George Petrie’s account, 1820 3.3 The 19th-century copper mines and the miners 3.4 Later 19th-century accounts 3.5 Twentieth-century accounts

28 28 30 32 37

4. The Monastery and its Associated Structures before Excavation 4.1 The site of the monastery 4.2 The monastery 4.3 The church and its enclosure 4.4 The monastic enclosure 4.5 Structures outside the monastic enclosure

39 41 42 42 45

5. The Excavations 5.1 Introduction 5.2 The church enclosure 5.3 Cell B 5.4 The monastic enclosure 5.5 The sub-rectangular building outside the monastic enclosure

52 53 95 101 136

6. The Finds Claire Cotter 6.1 Introduction 140 6.2 The stone objects 145 6.3 Stone identification and X-ray analysis of selected small finds 159 6.4 Metal 160 6.5 Archaeometallurgical residues Tim Young 164 6.6 Coins 171 6.7 A rare Viking-age penny from High Island Michael Kenny 172 6.8 Bone 173 6.9 Glass beads 174 6.10 Miscellaneous modern 174 6.11 Pottery and clay pipes Clare McCutcheon 175

7. The Cross-slabs 7.1 Catalogue of cross-slabs

Christine Maddern 176 Ian Fisher (revised by 200 Georgina Scally and Christine Maddern)

8. The Burial Record 8.1 The human remains 8.2 Strontium and oxygen isotope analysis

Barra O’Donnabhain 228 J. Cahill-Wilson, C.A. Taylor, H. Usborne, P. Ditchfield and A.W.G. Pike 236

9. The Environmental Reports 9.1 The faunal remains 9.2 Non-wood plant macro-remains 9.3 The charcoal remains 9.4 Soil and sediment analysis

Margaret McCarthy 238 Meriel McClatchie 261 Rowena Gale 269 Stephen Lancaster 284

10. Discussion and Conclusions 10.1 Introduction 10.2 The Iron Age 10.3 The origins of the monastery 10.4 The monastery during Phase 1: 8th to mid-11th century 10.5 The monastery during Phase 2: mid-11th to late 12th/early 13th century 10.6 Phase 3: late 12th/early 13th century to mid-15th century 10.7 Phase 4: mid-15th to late 20th century

292 292 292 294 303 316 317

Appendix A: Analysis of mortar samples: visual and petrographic analysis of aggregate, binder and additions Sara Pavía 319 Appendix B: Analysis of samples of clayey soil Sara Pavía and Jason Bolton 331 Bibliography 334

acknowledgements My involvement with High Island can be attributed to Jenny White Marshall, who invited me to be the licensed director of the initial exploratory excavation and without that first step, none of this would have come to pass. From the first year of excavation, in 1995, through to the final year, in 2002, 45 people worked at the site and all, without exception, exhibited traits of uncommon adaptability to the somewhat unorthodox working, living and dining arrangements that became the hallmark of the excavations; many thanks to David Adams, Nora Bermingham, Jason Bolton, Jerome Cameron, Pádraig Clancy, Patrick Conway, Caroline Cosgrove, Claire Cotter, Ross Crawford, Claire Crowley, Abi Cryerhall, Niall Donald, Ed Donnelly, Alec Dundon, Brendan Fagan, Ken Fitzsimmons, Fran Green, Brian Hayden, Aidan Holmes, Mark Keane, Helen Kehoe, Shaun Keith, Eoghan Kieran, Bibi Larsson, Ellinor Larsson, Alan MacCarthy, Jenny White Marshall, Declan McCormick, Brian McDomhnaill, Gillian McLoughlin, Melanie McQuade, Maeve Moriarty, Franc Myles, Niamh O’Callaghan, Eavan O’Dochartaigh, Donal O’Flaherty, Rob O’Hara, Sinead O’Mahony, Dáire O’Rourke†, Sara Pavía, Dan Rhodes, Andy Shelley, RuthAnn Spike, Christine Wagner and Sharon Weadick. Two people, Abi Cryerhall and Franc Myles, were present for nearly every session. Abi and Franc also contributed to the initial period of post-excavation and many thanks to them for their considerable contribution. Many thanks also to Jenny White Marshall who, from the outset, established exceptional culinary standards without which, I am in no doubt, the entire excavation crew would have fled the island after the first season, never to return. When excavation began, the island was in the ownership of Richard Murphy, who was hugely supportive of the excavation and subsequent conservation works. Without his support in those early years there would have been no excavation, so many thanks to Richard for his permission to live and work on the island and for his encouragement and interest in the excavation results. The logistics of excavating on an off-shore island such as High Island are immense. During the course of the excavation and until his retirement in 2008, working life on the island was overseen by Pat Heraghty, District Works Manager of the Office of Public Works (OPW). Many thanks to Pat, foreman Jim O’Toole, and the staff at the Cong and Athenry National Monuments Depots for all their help and hard work.

Thanks also to Michael Heraghty of the Athenry Depot, who on a number of occasions had to jump into swift action when severe storms hit the island in the late 1990s and we were cut-off for some time. At times, these conditions warranted evacuation by helicopter, an occurrence not taken lightly, but deemed necessary for the safety of all involved. Since the very first day of excavation in 1995, when Féicín Mulkerrin took us out to the island on his boat, little did we realise the vital role he would play in the years to follow. Throughout the excavation, Féicín had a multitude of roles for which we, as a team, were all immensely grateful. Féicín was the lynchpin of the excavations and without his obliging and thoughtful character, our time on the island would have been considerably more harsh, more hungry, more thirsty and, in general, more anxious. Over the years, Féicín ferried us on and off the island in a range of boats and in conditions that were sometimes challenging, but never unsafe. He also bought, shipped and delivered food, etc. when supplies ran low; and in the early days, before mobile communication, he maintained contact with us by two-way radio, which sometimes entailed climbing onto the roof of his house, or driving to Aughrus Point to get sufficient reception to hear our list of requests. Weather permitting, he made impromptu evening visits to the island and took us on excursions around the surrounding seas and nearby islands. His input to the smooth running of the excavations was immeasurable. John McDonagh, a hugely skilled stone mason, played a vital role in the conservation of the church, the beehive cells and was working on the conservation of the monastic enclosure wall at the time of his sudden death, in 2010. Over the years John was helped initially by Tom Mulroe, and later by Gerry and Noel Mulroe, to whom also many thanks. Post-excavation work continued over several years and involved many people. To all the various specialists who have contributed to this volume, many thanks. The artefact and cross-slab drawings are the work of Patricia Johnson and many thanks to her for creating one of the finest illustrated collections of cross-slabs in the country. Many thanks also to Martin Halpin for producing the published site drawings, not an easy task when so much detail is required to be shown in such restricted space. Particular thanks are due to Con Brogan of the Photographic Unit, National Monuments Service, for the many photographs used in the earlier sections of this report, including

8 high island: excavation of an early medieval monastery

the aerial photograph used on the cover. John Scarry and Tony Roche of the Photographic Unit were also of assistance in compiling the photographic record. Along the way various other specialists helped out and contributed in many ways: Aubrey Flegg of the Geological Society of Ireland; Conor McDermot of the State Laboratory; Chris Stillman, Adrian Phillips†, Jeff Clayton and Jacqui Connolly all of Trinity College Dublin; Michelle Comber, University College Galway; Dr Gerry McCormac, Radiocarbon Laboratory Queen’s University Belfast, and Claire Owens, Radiocarbon Accelerator Unit, Oxford; Suzannah Kelly for conservation of the artefacts (to Adrian Kennedy for conservation of 478:1); Stephen Carter for discussion on soils; Tim Gordon for discussing some aspects of the island’s geology; Oscar Merne† for help identifying resident and migrant birds; and to Stephen Harrison and Sharon Greene for discussing various aspects of the site. Thanks also to Elizabeth O’Brien for discussion on the burials and to Claire Cotter for discussing ideas and interpretation of various aspects of the site over the years.

Thanks also to Andy Shelley for reading a first draft, to the anonymous referee for many useful suggestions and comments and to Miriam Clyne for pre-publication editing and for helping to make the text considerably more readable and user-friendly than had previously been the case. Every project like this has a project manager; in the early years that role was filled by Con Manning and Ann Lynch, overseen by the chief Archaeologist of the National Monuments Services, Brian Duffy. In later years, Heather King took over as project manager and stayed with it through a decade of post-excavation and conservation work. After Heather retired in 2009, Ann Lynch took over once again and has managed to guide it through to publication, for which I am very grateful; many thanks to Heather and Ann for supporting the project through the best part of two decades. Finally, I would like to thank my friends, especially Claire Cotter and Larry Nesbitt and the people of Claddaghduff and Omey, particularly Féicín Mulkerrin, his friends and family, who helped out in so many other countless ways.

1 high isl and 1.1 Introduction The first systematic surveys of the monastic remains on High Island were carried out by Michael Herity in the 1970s and 1980s. This work laid the foundation for a more detailed, multi-disciplinary study of the monastery and related structures that commenced in the late 1980s, undertaken by Jenny White Marshall and Grellan D. Rourke. As part of this study, a programme of excavation was proposed in order to better understand the relationships between the structures and to facilitate the conservation of the monastic buildings, which had seriously deteriorated in the previous decades. As the monastery is a national monument in State ownership, Pl. 1.1 High Island, looking north (Photographic Unit, NMS).

its care is the responsibility of the National Monuments Service and, with the collaboration of Paul McMahon, senior conservation architect, a schedule of conservation works and excavation was drawn up as excavations progressed. In 1995 an exploratory excavation was carried out, jointly funded by White Marshall and the National Monuments Service, during which it was discovered that the monastic structures had remained largely untouched since late medieval times. The subsequent seven seasons of excavation and post-excavation work were financed by the National Monuments Service. The preliminary results of the excavations carried out up to and including 1997 were incorporated into the publication, High Island, An Irish

10 high island: excavation of an early medieval monastery

CLEW BAY Clare Is.

Caher Is.

Inishturk

Inishbofin

CO. MAYO

Duvillaun

Inishark Aughrus Pt HIGH IS.

Cong

Omey Is.

Clifden

Slyne Head

Inchagoill

Chapel Is.

Lough Corrib CO. GALWAY

Galway

St. MacDara's Is.

20 km

Galway Bay

Fig. 1.1 Map showing location of High Island (Ardoileán), Co. Galway, and neighbouring islands with early medieval remains.

Monastery in the Atlantic (White Marshall and Rourke 2000), and the results of all eight seasons of excavation are now presented in this monograph.

1.2 Location High Island lies in the Atlantic Ocean, less than 4km off the north-west coast of Connemara in Co. Galway (Figs 1.1, 1.2; Pl. 1.1). This proximity to the mainland at Aughrus Point does not reflect the feeling of remoteness that the island imposes on all who land there. The early medieval monastery on the island was closely associated with the larger foundation of St Féichín on Omey Island, which can be reached on foot at low tide from Claddaghduff on the mainland. Omey Island was the most likely starting-point of the monks on their outward journey to High Island. The 6km route is hazardous, though the first part of the crossing is through safe water, sheltered by the low-lying land of Omey to the south and by Aughrus Point to the north. Once past the Point, however, the real sea journey begins, one not to be undertaken by the inexperienced or by those unfamiliar with this stretch of water. The islands of Friar and Cruagh lie to the north and south, respectively, and in between, along the most frequented route to High Island, the reef known as the

Carrickculloo Rock is visible in all tides and is a landmark to be avoided. On both sides, groups of rocks, the Trusses and the O’Maley Breakers (each an area of about six or seven dispersed rocks), are precariously low-lying, covered at high tide and difficult to locate even in calm seas, making them amongst the most dangerous along this route. Strong currents prevail in these waters, especially between High Island and Friar Island.

1.3 Topography and geology High Island (Ardoileán) is appropriately named as the cliffs on all sides are high and sheer, rising dramatically from the sea to a height of 63m OD (Fig. 1.3; Pls 1.1, 1.3). In 1878 George Kinahan of the Geological Survey of Ireland described the island as one of the partly submerged peaks of the Twelve Bens ridge (Na Beanna Beola), ‘bounded on all sides by high cliffs, which are for the most part perpendicular, or nearly so’ (Kinahan et al. 1878, 11, 53). The island, with its heavily indented shoreline, extends roughly north-east–south-west and covers 32 hectares (80 acres). It measures approximately 1.2km long by a maximum of 0.4km wide, and at the north-east narrows to just over 11m. The central part of the island is elevated, and its highest point lies at the eastern side, at 63.3m above sea level.

high island

The topography is varied, comprising a combination of hills and valleys with a large area (approximately 300m x 200m) of low-lying land located at the south-western end of the island (Fig. 1.3; Pl. 1.1). Smaller pockets of level, but elevated ground lie to the north, along the coastline, and at the eastern end. Two freshwater lakes occupy the area between the flat ground to the south-west and the more elevated areas in the centre. The larger, more northerly of the two lakes forms part of the monastic complex. The terrain is currently most suited to sheep-grazing. The bedrock is metamorphic, with schist being the predominant rock type. Some of the schist is rich in mica (silvered and foliated) and some is garnetiferous (T. Gordon, pers. comm.). Other rock types include bands of quartzite, psammite and marble (B. McConnell, pers. comm.). Limestone is shown on the 19th-century GeoFig. 1.2 Sea chart No. 1820 showing location of High Island, Co. Galway, and surrounds (courtesy of the UK Hydrographic Department â&#x20AC;&#x201C; licence no. 10765).

logical Survey of Ireland (GSI) map close to the shoreline on the north-eastern side of the island (Fig. 1.4, coloured blue), which belongs to the Lakes Marble Formation of the Bennabeola range of the Connemara succession. The GSI map is annotated to indicate the presence of gneiss but, in the modern understanding of the term, the rocks have not reached this metamorphic grade. Some small feldstone dykes that cut the metamorphic rocks are also recorded. Another rock type consistent with the underlying metamorphic geology is a fine-grained quartzfeldspar-mica-schist, probably originally a fine-grained acid igneous rock. The island was also subject to glacial activity, which brought with it the massive granite boulders that are scattered across the surface of the island in addition to the glacial till and beach pebbles in the coves. Since the ice flow direction was towards the north-west,

12 high island: excavation of an early medieval monastery

N. LANDING

Fig. 1.3 Contour map of High Island (5m intervals), showing ponds, spot heights and landing places (after White Marshall and Rourke 2000, Fig. 4, with additions).

40.5m

49.7m

45.5m

62.6m

S. LANDING

63.3m

33.5m

Monastery

31.0m

S.E. LANDING

Lake

the metamorphic and granite rock debris is likely to have been transported from the Connemara mainland. Much of the island is covered with peaty soil, which, where tested, near the central part of the island and in the valleys, was in excess of 1m in depth. Along the cliff edge, a thin layer of light grey soil covers the glacial till. The known mineral resources on High Island are copper, sulphur and iron pyrites (Kinahan et al. 1878, 53, 160). Remains of 19thcentury mining activities survive in the form of a shaft on the north-east of the island and two nearby accommodation huts (Fig. 4.1, nos 4–6). As the only recorded inhabitants of the island since the departure of the monks several hundred years earlier, the impact of the miners on the island and on the monastic remains is likely to have been considerable.

1.4 Landing places Landing on High Island with any degree of comfort is pos-

100

150

200

250m

sible only when the waters are calm and with favourable winds. There is no beach, pier or slipway and there is no sheltered cove or refuge within which to dock boats safely for long periods of time. Boats must be hauled onto the rocks for safety, although it is also possible to leave them at anchor when sea and wind conditions are favourable. There are three established landing places on the island (Fig. 1.3). Two are located at the north-east and the third lies at the south-east and is the closest (300m) to the monastery. Each landing place has its own merits, and the decision to use one in preference to the other is based on wind direction, the sea swell and the purpose of the journey. At all three landing places it is necessary to jump from boat to rock with split-second accuracy. When the wind blows from the north and the sea swell follows, the south landing is the preferred choice. Despite the short and relatively safe scramble to the elevated ground some 25m above, a gully in the rock at sea level poses the most danger. Any rough sea or incoming

high island

Fig. 1.4 Geological Survey of Ireland 19th-century manuscript ﬁeldsheet. Extract from 6-inch topographic map, Galway Sheet 21 (ﬁrst quarter). Field notes may have been written by G. Kinahan some time between 1870 and 1879 (courtesy of the Geological Survey of Ireland).

tide surges through the gully and spills into the cove with considerable force (Pl. 1.2). The north landing is generally used when the prevailing winds and sea swell come from the south. The landing is made close to the tip of an extended arm of rock projecting into the sea (Pl. 1.3). In dry weather, the climb of 30m can be achieved with relative ease, aided by small footholds carved into the rock by the miners who lived on the island in the early 19th century (Pl. 1.4). In wet conditions, however, this landing can be treacherous as the rock becomes slippery, like ice, making the climb considerably more difficult despite its relatively low height above sea level. The south-east landing is located at an inlet known locally as the ‘monastery landing’ (Fig. 1.3; Pl. 1.1). Previous writers have suggested that the landing was located on the northern side of the inlet, but they point out the dangers of climbing the steep and crumbling cliff-face and note also that landing is only possible at high tide because at low tide the ledge above cannot be reached from the boat (White Marshall and Rourke 2000, 34). However, if the other, southern side of the inlet is chosen, where the rock falls away gently towards the sea, landing is considerably less arduous. It is possible to land here at both low and high tides. At low tide, a boat can be pulled into the shore 40m from the mouth of the inlet, and at high tide the boat can be brought alongside the rock-face 10m from the mouth of the inlet, where a jump from boat to rock

Pl. 1.2 Sea surging through rock-gully, south landing (G. Scally).

14 high island: excavation of an early medieval monastery

Pl. 1.3 North landing (G. Scally).

can be made onto one of several ledges (F. Mulkerrin, pers. comm.). Once ashore, there is sufficient room to haul supplies up the incline of approximately 15m, and a little way inland there is level ground where it would have been possible to store supplies and boats. From here, it is an easy climb to the higher ground in sight of the monastery.

1.5 Flora and fauna

Pl. 1.4 Excavation crew scaling the north landing (G. Scally).

The island is dominated by maritime grassland vegetation that supports a range of plant species (Molloy et al. 2000, 234–5). In spring, the most colourful plant is the common orchid. Sea-pink, English stonecrop, tormentil and bird’s-foot trefoil are also widespread, particularly on the north-eastern and southern parts of the island. Various grasses abound, and black sedge and marsh pennywort are prominent towards the western and south-western areas, where they flourish in the deeper valleys. In late summer the island is covered by wild angelica, which spreads almost universally across all areas. Common nettle, greater plantain and alexanders, all plants associated with human disturbance or habitation, are notably lacking. At present, there are no trees on the island, with the exception of a localised area of dwarf willow at the north-eastern end. An abundance of bird life is sustained. In spring, great black-backed gulls, a smaller number of lesser blackbacked gulls and herring gulls breed, as do fulmars and oystercatchers. A small number of shags nest in the coves,

high island 15

and on the cliff-face on the southern side of High Island at least one pair of peregrine falcons breeds each spring. In summer, Manx shearwaters return to breed while large numbers of storm petrels nest in the stone walls and rubble around the island, in particular at the monastery. A small number of choughs and ravens also breed here, as do many songbirds, including skylarks, rock pipits, meadow pipits, wrens, robins, wheatears and stonechats. Passage migrants, such as curlew, whimbrel, dunlin, greenshank, green sandpiper, swallow, chiffchaff and starling, have all been seen on land. Out to sea, black guillemots, terns and gannets are relatively common. Occasionally, a lone puffin and, even more rarely, a wood pigeon have also been spotted. In late autumn, barnacle geese arrive to overwinter from northern Europe and when they leave in spring, the island is covered in a mantle of their droppings. The number of bird species found is a reflection of the diversity of habitats on the island. There are rocky cliffs for the fulmars, shags, peregrines and choughs; crevices for the petrels; burrows for the shearwaters; good grazing for the geese; short, grassy peat with invertebrates for the choughs and smaller birds; and fresh and brackish ponds for the dunlin, oystercatcher, greenshank and green sandpiper. The island is close to suitable fishing areas for the

seabirds and to the continental shelf, where the petrels feed on plankton (O. Merne, pers. comm.). A limited range of mammals has been recorded on High Island. The dominant species was the rabbit. Sir James Ware, writing in the mid-17th century, noted that rabbits were abundant on nearby Cruagh Island, which he described as insula cuniculorum (‘the island of the rabbits’) (O’Flaherty 1846, 114). It is possible, therefore, that rabbits also existed on High Island at this time, but it is more likely that they were introduced as a food source by the miners in the 19th century. When excavations began in 1995, rabbits were prolific; twelve years later, the entire population had disappeared. The reasons for this decimation are not immediately clear. The seas around the island support a variety of marine life. Pollock, cod and mackerel abound, as do lobster and crab. Seals thrive in the surrounding waters and the grey seal breeds each year in the rocky coves. Less frequently, basking sharks and sunfish have been spotted in the inshore waters. Records show that whales are not unknown in the area; a shoal of killer whales was seen in 1974 to the north of High Island, near Inishbofin (Figs 1.1, 1.2), while the last recorded sighting was a minke in 1991 (Irish Whale and Dolphin Group, www.iwdg.ie).

2 high island and the cult of saint féichín Pádraic Moran

High Island is one of very many early medieval monastic sites which furnish scarcely a mention in the historical record.1 It follows, then, that any attempt to sketch out the history of the island must inevitably be tentative and conjectural. What few references there are have already been collected and discussed in White Marshall and Rourke (2000, esp. 7–21, 215–28), including those collected earlier by Petrie (1845, 424–7). This section aims to supplement that material and provide a fresh assessment, giving in addition some account of the wider circumstances that may have shaped the monastery during the lifetime of its occupation. It also attempts to reassess the traditions surrounding the island’s patron, Saint Féichín, taking into account their focus on the apparent centre of his cult in Fore, modern County Westmeath. (For a recent survey, see Ó Riain 2011, 309–11.)

Saint Féichín and his cult

Traditionally, the foundation of the monastery at High Island is credited to Saint Féichín (also written Féchín or Fé(i)chíne), son of Caílcharn. The only record of his life in the annals is his death in 665, during the Yellow Plague (AU 665.3, CS 665, AT 665.4, AI 666.8, FM 664.1). AU enters his death twice, noting it again at AU 668.5 secundum alium librum (‘according to another book’). The martyrologies, records of the feasts of the saints, celebrate Féichín on 20 January, and in the Félire Óengusso (Martyrology of Óengus), c. 820 (Stokes 1905), he is referred to by his hypocoristic or pet name, Moécu (= Mo Fhécu, ‘my Fécu’, a shortened form of Féichín; see also CGSH §703.5). The later commentary on Félire Óengusso explains the name Féichín as a diminutive of fiach, ‘raven’ (Stokes 1905, 48–9). This explanation, though often-cited, may be called into question. It appears to rest on the assumption that fiach derived from an older form, *féch, which cannot have been the case given that fiach was disyllabic. Another etymological note in the same commentary, following directly, illustrates the unreliability of such material. Féichín’s name is explained as derived from the word feccaidecht,

meaning ‘bending’ or ‘stooping’, and a story explains that Féichín was angered at a perceived insult from St Ciarán, and refused to show his face to him. Ciarán dubs him a forḟeccaid, which appears to mean somebody who bends over. (Stokes’s translation ‘backslider’ is hardly appropriate here, as noted in DIL s.v. fec(c)aidecht.) Instead, Féichín may be a diminutive of the names Fiacc or Fiachu, from a word meaning ‘battle’ (Ó Corráin and Maguire 1981, 94). The Proto-Indo-European root is *weik- ‘fight, conquer’ (cf. Latin uictor, Irish fichid ‘fight’), giving Ogam vecrec (> Fíachrai) and archaic Irish Feec (> Fíac), Fēchrach (> Fiachrach), Fēchach (> Fiachach).2

Lives of Féichín

Our main sources of information on Féichín are a number of saints’ lives, one written in Latin and two in Irish, with another version composed in Latin, based on Irish-language sources now lost, by the Franciscan John Colgan in the 17th century. (The manuscripts and their relationships are discussed below.) Saints’ lives should not be regarded as historical biographies in any modern sense, but rather as texts written to proclaim the spiritual authority of their subjects, and by extension those monasteries and churches associated with them. Saintly authority is generally made manifest through a catalogue of stock miracles, including prophecies, cures and angelic visions, which rarely disclose convincing personal characteristics or historically plausible events. (This approach contrasts with the uncritical account of Féichín’s life in Coyle 1915; similarly G.T. Stokes 1892.) In general, their historical value lies in the circumstances of their authorship and transmission. They may seek to justify aspects of the relationships between monasteries, and between monasteries and secular authorities, or account for the origins of saintly relics and other venerated objects. Incidental details can also offer much in the way of social history: Plummer (1910, xcv–cxxiv) provides an invaluable list of references to agriculture, animal husbandry, bee-keeping, drying and milling grain, craftwork, ship-building, brewing, dyeing,

1 I am grateful to Denis Casey, who read a draft of this section and offered several suggestions. 2 See Thes. Pal., vol. 2, 259.31, 259.38, 260.9, 263.40, 264.24.

high island and the cult of saint féichín 17

the learned classes, kingship, fosterage, burial customs, ecclesiastical organisation, monastic labour, study, hospitality, ascetic practices, liturgy and the veneration of relics. The surviving versions of Féichín’s lives are as follows: V) The Latin Vita Fechini is found in the Oxford collection of saints’ lives. Rawl. MS B 485 was written in the early 14th century, and Matthew O’Dwyer made a copy, Rawl. MS B 505, in the late 14th century. The latter has been associated with Auguistín Magraidin (1350–1405), a canon of Saints’ Island in Lough Ree, who is said to have compiled lives of Irish saints, though the early date of its archetype would appear to rule him out as the author of the lives. Richard Sharpe (1991, 247–65, 368–83) has argued that the texts date from the late 13th century. A copy provided by Hugh Ward was the basis of the first published edition of Féichín’s Latin life in 1643 (Acta. Sanct., vol. 2, 329–33). Another copy, by Henry Fitz Simon, also available to the Bollandists, formed the basis of Colgan’s edition (1645, 130–33). Plummer re-edited the text in 1910 (vol. 2, 76–86), and regarded it as ‘an abbreviation of a longer work… clearly incomplete (there is no account given of Fechin’s death)’ (1910, vol. 1, lxv). B1) This is the first of two Irish lives (bethada) written consecutively in NLI MS G5, fol. 1r–8v, a 15th-century manuscript once part of the collection of Sir Thomas Phillipps at Cheltenham (Ní Shéaghda 1935, 31–5).3 Little is known of its earlier provenance, though Charles O’Conor of Belanagare (the elder, 1710–1791) signed the manuscript while a young scholar in 1731. Although Stokes (1891) printed both lives as a continuous text, they are clearly separate. The first ends with the colophon (fol. 5v): ‘The young Nicholas, son of the abbot of Cong, put this life of Féichín out of Latin into Gaelic, and Ua Dubthaigh took and wrote (it); and this is the year of the age of the Lord today, 1329, etc.’ The market cross in Cong may honour the names of these authors: it commemorates a Nicholas and a Gilliberd Ua Dubthaigh as abbots, and seems to date from the 14th century (see Macalister 1945–1949, vol. 2, §546–7). The same names are commemorated on the base and what Macalister regards as the original shaft. The base inscription is copied (imperfectly) on a modern shaft with the date 1350. Nichol Óg is described as the son of the abbot. His designation óg, ‘the young(er)’, suggests that his father was also named Nicholas, and the latter may have therefore been the abbot who was commemorated on the cross. Alternatively, given the hereditary nature of the abbacy at this time, Nichol could have succeeded his father and himself be the dedicatee. Ua Dubthaigh was the name of a prominent clerical family in Connacht, with connections to Cong. The annals include death notices for the following Ua Dubthaigh archbishops: Domnall (FM 3 This manuscript may be consulted online at <http://www.isos.dias.ie/>.

1136.2, AT 1136.4), Muiredach, an agent of Toirdelbach Ua Conchobair (CS 1149, FM 1150.1), Flannagán, who died in Cong (FM 1168.1), and Cadla, who ‘rested in Cunga Feíchín’ (AI 1201.9). Muiredach is also commemorated on the processional cross of Cong (Macalister 1945–1949, vol. 2, §552). Although 1329 seems to refer to the year of Ua Dubthaigh’s original copy (his colophon having been copied into the extant 15th-century manuscript), there is little in the language of the text to indicate that it was composed much earlier. B2) The second Irish life (fol. 6r–8v) is the shorter of the two, and its Latin introduction signals its function as a monastic lection: O uos fratres carisimi, audiuimus plura de uir[tu]tibus sancti Fechini abbatis et ancoritae (‘O dearest brethren, we have heard many things about the great deeds of Saint Féichín, abbot and anchorite…’). The life is clearly incomplete, missing the conventional accounts of the saint’s birth and death. C) Colgan followed his edition of the Latin life (V) with a supplementary life (1645, 133–45), compiled in his own Latin from Irish-language sources available to him, since lost. He described these as: C1) an Omey manuscript, translated from Latin into Irish; C2) an old text missing a beginning and end; C3) a metrical version in 74 couplets. (It is fortuitous that Féichín’s feast occurs in January, as only this first of four projected volumes of Colgan’s Acta ever appeared, featuring saints celebrated between January and March.) There is also a metrical life preserved in the Yellow Book of Lecan (Dublin, Trinity College Library, ms 1318 (H.2.16), cols. 1–2). The text is difficult to read, though a transcript by Eugene O’Curry is bound into the volume at the start. It begins with an eighteen-line poem, followed by the saint’s genealogy (traced back to Adam), and then a longer metrical work in 68 couplets. (Abbott and Gwynn 1921, 94, give references to the facsimile edition, but I have not been able to find the relevant pages in any copies I have consulted. They also write that the text was treated by Stokes in his edition of the Irish lives of Féichín, which is not the case.) This text remains unpublished and unedited, and a detailed study is still required to determine its date, historical value and textual relationships with the other lives (most notably the metrical version used by Colgan). Plummer observed (1910, vol. 1, lxvi) that Colgan’s description of his two prose sources (C1, C2) corresponds to that of the Irish lives (B1, B2): C1 is a translation from Latin (as is B1); C2 is missing its beginning and end (like B2). Moreover, V and B1 agree closely both in content and in their sequence of events, suggesting that a version of V was the basis of B1’s translation. There are few

18 high island: excavation of an early medieval monastery

episodes in V that are absent in B1. Of these, four (V §7–9, 15) were also absent from Hugh Ward’s copy, used by the Bollandists, suggesting Ward’s text is closer to the copy translated by Nicholas Óg. (Exceptions are V §11 and, significantly, the water mill story in V §14 – the latter is found in B2 §39, however.) The identity of B2 and C2 is supported by the fact that C2 parallels material occurring in B2 but not elsewhere (B2 §40–48, though not §39). The remaining episodes occurring uniquely in C (§§8, 14–21, 38–47) would therefore appear to come from Colgan’s metrical source (C3). These relationships are outlined in the concordance table below. A common narrative framework underpins all of the lives. Féichín was born among the Luigne in Connacht, a people inhabiting the area of present-day County Sligo, south of the Ox Mountains, whose name is reflected in the modern barony of Leyney. The place of his birth is Bile Féichín (or Bile Fobhair), associated with modern Bella, about 3 miles south of Ballysadare (see O’Rorke 1878, 220–22). His father was Caílcharn, his mother Lasair, ‘of the royal race of Munster’. Alternatively, the genealogies (CGSH §722.45) give Féichín’s mother as Sochlo or Sochla. As is conventional, the saint’s destiny was manifest from his childhood, with miraculous signs and the proph-

ecy of an authority none other than St Columba. Féichín’s education is entrusted to a local priest, Nathí, at Achonry (County Sligo), evidently an important church given its later status as a diocesan centre (V §4, B1 §§6–7, C §6). After establishing his authority over the local king (by cursing the king’s horses, who die and are then restored to life), an angel calls Féichín to Fore (just north of Lough Lene, in modern Westmeath), where he founds a substantial monastery (V §10, B1 §§9–11, B2 §§32–33, S §9). The Book of Leinster (c. 1160) contains a list of Irish saints paired with other saints of a similar nature, drawn largely from the Bible and Church Fathers (edited in CGSH §712). Féichín is associated there with St Anthony, the anchorite whose literary life became an inspiration for western monasticism. Féichín’s asceticism is not particularly emphasised in his lives (except perhaps C §46). Ó Riain (in a note in his edition) suggests that Féichín may have been exchanged with Kevin of Glendalough, who precedes him in the list and is described in his own life as an alter Antonius (‘another Anthony’). After various miracles involving provision of food, prophecy, resurrecting the dead, curing the sick and disfigured, and shining miraculous light, Féichín is called back to his native Connacht to convert the heathen population of Omey Island (Imaidh Féichín; V §12, B1 §17,

Table 2.1 Concordance of episodes in the extant lives of Féichín. Asterisks (*) mark sections of V not present in Hugh Ward’s copy. Square brackets [ ] denote partial correspondences. Sections common to all versions are in bold face.

Birth Childhood signs Miracles Churches among Luigne Foundation of fore Provides food Miracles Miracles Called to Omey Cures a leper (Jesus) Miracles Miracles Builds a mill Miracles Miracles Miracles Miracles Alliance with St Fintan Death

V 1–3 4–6 7–8*

— 10 11

— —

12 13 [17]

— 14

15*–16 [9]*

—

18–21

— —

B1 (§§1–28) 1–4 5–8

— —

9–11

—

12–16

—

17 19 20–25

— — — — — — —

26–28

B2 (§§29–48)

— — — —

32–33

— — —

35–36 37–38

— — 39

—

40–48

— — — —

C 1–4 5–7

—

8 9

—

10–13 14–21 22 23 25–27 28–30

— —

31–37 38–46

—

47 48–50

high island and the cult of saint féichín 19

B2 §§35–36, C §22). The pagans prove intractable. They cast the possessions of Féichín and his monks into the sea, but evidently not swayed by the miraculous restoration of these implements, they finally succumb to Christianity after Féichín restores to life two of his brethren who had died of starvation. This latter miracle came to the attention of Guaire Aidne, the king of Connacht, who sends them a plentiful supply of food, and his own cup from which to drink. The saint’s subsequent career seems mostly to have been played out at Fore. Aside from the usual cures and prophecies (described as ‘ecclesiastical whitewash’ by Plummer 1922, vol. 1, xxiv), the saint has a number of engagements with secular rulers Díarmait and Blathmac, the sons of Áed Sláine, joint kings of Tara and overkings of Brega and Mide (in which Fore was situated). Féichín obtains the release of hostages from Blathmac, after burning down his fortress with a fiery bolt from the sky, and then cures the king’s burns (B1 §21, C §20). He similarly frees Áedán, a warrior of noble race held captive by Díarmait (B1 §24, C §27). (Féichín then performs the miracle of reducing Áedán’s enormous appetite from that of seven men.) The saint also becomes involved in a great convention held between the northern and southern Uí Néill, the former headed by Domnall son of Áed (based in the north-west of Ulster), the latter by Díarmait and Blathmac, at Ráith Droma Nó (in Cenél Maine, east of Lough Ree). Domnall son of Áed was high king from 628 to 642/3. The name given to the hosting has caused confusion over its purpose. Stokes emended the name of the convention from the manuscript’s sluaiged in Meith to sluaiged in Meich, citing Colgan’s version. Colgan, however, was writing in his own words, and may have added his own interpretation. The text itself says that succession between the two branches of the Uí Néill (claochlod in da Niall), rather than boundaries, was the issue at hand, and there is no further mention of boundary marking in the text. However, it does say that Féichín provided food to sate the troop for three days and a night, and therefore the manuscript’s sluaged in meith (méth, ‘fat, rich’) is probably intended. Colgan’s interpretation may have been influenced by the annal entry AU 641.3, ‘Domnall m. Aeda castra metatus est i nDruim Náo’, which merely refers to Domnall measuring out his camp, not his territory. (The translation in Mac Airt and Mac Niocaill, ‘Domnall son of Aed changed camp in Druim Naó’, is not helpful either.) The lives have Féichín first feed the southern Uí Néill troops, and afterwards obtain the submission of Domnall (B2 §§44–45, C §§34–35). These events portray Féichín as closely engaged with the secular rulers of Mide and Brega, and although he emphatically asserts his superior authority, he is nonetheless their protector and patron. All versions of the life record the arrival of a leper at

Fore, seeking food and drink and a well-born woman to sleep with him (V §13, B1 §19, B2 §§37–38, C §23). Féichín makes the necessary arrangements, soliciting the wife of King Díarmait for the task. The leper then makes a further request: that the queen suck the mucus from his nostrils. She duly obliges and the following day, after the leper’s departure, his secretion has turned into a gold chain, and his identity is revealed as Jesus. Another arresting passage has Féichín being disturbed in his prayers by children playing hurling nearby (B2 §43). He then gives the children permission to drown themselves in the nearby lake, and their souls go to heaven. The lives conclude with an account of the saint’s death, which is attended by other holy men. Féichín receives communion and last rites from Mochoemóc (associated with Liath in Tipperary, east of Thurles), then Mochua (either of Balla in Mayo or of Timahoe in Laois) sees a sign in the sky and dispatches his own soul to heaven, while Moling (St Mullins, Carlow) converses with Satan, who reveals that Féichín has reached heaven unmolested by demons. This very conventional account of the saint’s death contrasts strikingly with that found in other sources. The life of Gerald of Mayo (Plummer 1910, vol. 2, 107–15) recounts that Díarmait and Blathmac summon a council to discuss the problem of famine in Ireland, at which it was decided to pray and fast for a plague in order to thin out the population. Gerald objected on behalf of the innocent, though he was opposed by Féichín, who won the day. When the plague came, Féichín himself died, as did Díarmait and Blathmac. The story appears to be reflected in the various annal entries recording Féichín’s death, which list Féichín first or early on among the dead, and sometimes specifically associate his name with the joint-kings. The story is also found in Colmán’s hymn (Stokes and Strachan 1903– 1910, vol. 2, 298–306), dated to the Old Irish period (i.e. before 900), thus pointing to the existence of an early version expurgated in the later tradition.

Féichín in Meath and Leinster

Despite being born in Connacht, Féichín is firmly associated with Fore, in the historical province of Meath (Irish, Mide; centred on the modern counties of Meath and Westmeath). His name is invariably given in sources as Féichín Fabair (or Fobair), Féichín of Fore. The lives locate the main events of Féichín’s career at Fore, where he has a close relationship with the ruling dynasts. (In the saints’ genealogies compiled by the Four Masters, Féichín is himself made a descendant of Áed Sláine; see Walsh 1918, 54). The episodes show him asserting his own higher authority, and invoking God’s power with impressive effect. Nonetheless, Díarmait and Blathmac are not made to suffer by his interventions, and indeed he nourishes their

20 high island: excavation of an early medieval monastery

army and safeguards their interests during the hosting by Domnall of the northern Uí Néill. Thus, the life asserts the greater authority of the Church over secular powers, while emphasising the benefits for local rulers of showing due deference. An independent entry in the annals supports his status as a patron of midland kings. Féichín is represented as making a personal appearance on the battlefield in 1069, to repel an attack from Murchad of Leinster (CS 1069, FM 1069.7). (The early medieval province of Leinster, the territory of the Lagin, covered roughly the modern counties of Wexford, Carlow, eastern Laois and Offaly, Kildare, south Dublin and Wicklow. The northern part of the modern province was controlled by a separate dynasty, the southern Uí Néill.) Murchad’s father, Díarmait mac Máel na mBó, had been raiding Meath intermittently since 1053, and was the first Leinster claimant to the high-kingship of Tara since the 6th century. By attributing the Uí Néill success to the intervention of Féichín, the annalist expresses the perceived special role of the saint as protector of the midlands dynasty. In another reference to Leinster in the lives, Féichín also has an encounter with Ailill son of Dúnlaing (B2 §41–42), whom he approaches at his royal residence at Naas (County Kildare) in order (once again) to secure the release of a hostage. When the king refuses, the saint invokes an earthquake, freeing all the hostages in the fortress and killing the king, who is afterwards restored to life, and who then grants the hill of Fore to Féichín, with freedom from tribute and the right to take tribute from Leinster. Stokes in his edition identified Ailill as the king killed by Norsemen in 871 (AU 871.4, CS 871, etc.), some two centuries after Féichín’s own death. However, another candidate may be found in the Ailill son of Dúnlaing who ruled Leinster from 527. His father was the founder of the Uí Dúnlainge dynasty, which was dominant in Leinster from the early 7th century down to 1042. In either case the chronology appears to be out of kilter. Equally puzzling is the donation by a Leinster king of land in a neighbouring province. Nonetheless, Féichín’s monastery was founded well after this period. However, an encounter with the earlier Ailill might be interpreted as an assertion of Féichín’s authority over the rulers of the neighbouring province. There is a further Leinster connection in an episode in Colgan’s supplementary life (C §47), where Fintan of Clonenagh (in County Laois) bequeaths his monastery to Féichín, although we are told that Fintan’s monks afterwards dissented. Two independent sources appear to corroborate this association. In one of Féichín’s genealogies (CSGH §550; see further below), he is descended from Echach Find Fuath nAirt, an ancestor of the Fothairt people, with whom Fintan of Clonenagh was associated. (Similarly in one of the genealogies compiled by the Four

Masters: see Walsh 1918, 75.) And a rather convoluted note in the commentary to the Félire Oengusso (Stokes 1905, 224 [Oct. 20]) explains that Fintan Máeldub of Durrow, an associate of Fintan of Clonenagh, was adopted by Féichín as a fosterling, and was made storekeeper of Féichín’s congregation. These connections may point to earlier traditions, suppressed for the most part in the extant lives. Curiously, Termonfeckin (Tearmann Féichín, ‘Féichín’s sanctuary’), in County Louth, is not mentioned in the lives at all.

Féichín and the Luigne

From the outset of the lives, Féichín’s connection with his own people, the Luigne, is made clear. This is reinforced in the saint’s genealogies (excepting CSGH §550, discussed above), which trace his descent from Tadg mac Céin, ancestor of all the Luigne (CGSH §§315, 421, 614). (Stalmans and Charles-Edwards (2004) write that the genealogies associate Féichín with the Gailenga, presumably because this group also trace their ancestry to Tadg mac Céin.) The Luigne people gave their name to the modern barony of Leyney in County Sligo; another branch in the midlands gives the barony name of Lune. Féichín is educated by a local priest, Nathí, for whom he secures the lands of his church, and it is here that he performs his first miracle, and makes his first stand against secular powers. A continuing relationship with his own people is further expressed by occasional return visits and miracles in the land of his birth, and the mention of local places (notably Ballysadare, later a parish) attests to the continued veneration of the saint in the area. The Luigne formed part of a patchwork of small territorial units (tuatha) whose coherence was maintained through a notion of common kinship. These groups fitted into a loose political hierarchy, at the top of which were the Connachta in the west and the Uí Néill in the north and midlands, who controlled the overkingship of their respective provinces and, in the case of the Uí Néill, the prestigious kingship of Tara. The Luigne were excluded from this dynastic structure, and were classified as an aithech-thuath (literally ‘payment people’; see e.g. CGH, 143 a 10–14), liable to render tribute and military support to their local overlords (in this case probably the Uí Aillelo, a minor branch of the Connachta situated on their eastern border). The Luigne were closely associated in the genealogies with certain other groups – the Gailenga, Cianachta and (sometimes) the Delbna – all of whom can be characterised as being of subject status, and by having branches geographically disparate. The location chosen for Féichín’s monastery at Fore is significant because he appears to have remained with his broader kin group, relocating to another, geographically separate branch of the Luigne.

high island and the cult of saint féichín 21

Fore, however, is not located within the modern barony of Lune, but gives its name to a barony of its own (Fore in Westmeath; there is a neighbouring barony of the same name in Meath). Borders are moveable, and the locations of modern baronies are not an exact guide to ancient territories. There are other indications that Fore was associated with the midlands Luigne: one annal entry (AU 993.5, FM 992.6) records that Máel Finnian Ua hÓenaigh was both successor of Féichín and bishop of the peoples of the Luigne, affirming not only the association between Fore and the Luigne, but also close ecclesiastical relations between the two branches. Féichín’s teacher, Nathí, is himself recorded in the earliest martyrologies (e.g. Stokes 1905, 175), and his prestige may be reflected in the fact that a diocese was later constituted around his church at Achad Conaire (Achonry). Perhaps Féichín sought to establish himself in a community away from his master’s shadow. In any event, remaining within his broader kin group would certainly have helped him carry out his work with greater legal and social mobility. Féichín is not the only saint who carried out his mission among a geographically divided people. Cianán of Duleek linked together the Cianachta of Brega and of Glenn Geimin. Brigit of Kildare was patron of the Fothairt, a people divided into at least four branches in Leinster. And there are further parallels. The Luigne, Cianachta and Fothairt are aithech-thuatha (subject peoples). The nature of Féichín’s miracles and interventions are also characteristic of the Kildare saint (see Charles-Edwards 2004a, 82–92). Many involve curing the sick and raising the dead. He confronts kings in order to secure the release of prisoners, or to maintain the peace. And rather than curse kings who are uncooperative, his encounters end with reconciliation. Like Brigit, Féichín champions underdogs and dependent peoples in the shadow of more powerful groups.

Féichín in Connemara

Kenney (1929, 458) was of the opinion that ‘the name of St Fechin is usually associated with that of his monastery of Fabar or Fobar… But his work seems to have been chiefly in the west of Connacht, where his most important foundation was the celebrated monastery of Cunga, or Cong. To him were ascribed also the churches of Omey island and Ard-oilen, “High Island”, off the most westerly coast of Galway.’ The evidence for this is not entirely clear. The founding of a monastery at Omey is recorded in all of the lives, and in every case occurs directly or soon after the establishment of Fore. There are few indications, however, in the description of the island or of the monks’ mission that the writer had any real knowledge of it. The account of the heathen inhabitants might best be regarded

as a literary trope. Charles-Edwards (2000, 240) suggests that ‘By the time of the first centenary of Palladius’ mission [ad 431], it was probably already clear that Irish paganism was a lost cause.’ If we take Féichín’s encounter with king Guaire as historical, it would place the foundation of the monastery within the latter’s regnal years of 655–663. However, this too appears to be literary, presenting Féichín’s positive interaction with the secular ruler of the day. Guaire developed a reputation for generosity in later literature (see Mac Eoin 1989, 172), which provides a context for the detail of donating his own drinking cup. The episode also provides an origin story for a relic later associated with the saint. C §22 records that Guaire’s gift is called Féichín’s cup ‘even today’ (usque in hunc diem). The leper episode functions partly as an origin story for Féichín’s crozier, doubtless another relic. (V §13 has Féichín encasing a piece of the miraculous gold chain in his crozier. In B2 §38 the crozier itself has been left by Jesus.) Another relic of Féichín, not mentioned in the lives, is noted in an annal entry (FM 1143.13), where Murchad Ua Maeleachlainn of Meath exchanges sureties with Turloch O’Connor, which include Féichín’s successor (comarbae) and Féichín’s bell. Colgan’s supplementary life alone furnishes any mention of High Island in all of the lives (C §22), almost as an afterthought following the foundation of Omey: Fundauit & vir Dei aliud Monasterium in vicina insula, quæ olim Inis-iarthuir hodiè ardoilen appellatur (‘The man of God also found another monastery in a neighbouring island, which was once called Inis Iarthuir and today is called Ard Oilén [High Island]’). Likewise, Colgan’s life records the only mention of Cong in the hagiography, when Féichín incidentally receives visitors there (C §21). This is particularly surprising given that B1 was translated from Latin by the young Nicholas, son of the abbot of Cong. Clearly, if he did know of any traditions associating Cong with Féichín, he did not see fit to include them. The persistence of his cult throughout Connemara is indicated by a number of holy wells mentioned by Roderic O’Flaherty in 1694 (Hardiman 1846, 106, 113, 120–21) at Killeen (Ballynahinch), Caramore and Gooreen (the latter, on Omey, ‘of late proves very miraculous for restoring of health’, p.113), supplemented in Hardiman’s footnotes by others at Teernakill, Cammanagh, Gowlaunlee and Dooghta. These wells are recorded in Gosling (1993) under sections 649, 674, 689, 692, 715 and 761, except for Caramore, which I have been unable to locate. The same survey also notes a well dedicated to Féichín at Drumsnauv (675), which had disappeared by 1899. Efforts to appreciate the context in which Féichín’s cult become established in west Galway are severely hampered by a general lack of documentary sources for the region. Indeed, the same may be said to apply to Connacht in

22 high island: excavation of an early medieval monastery

general, described by F.J. Byrne as ‘the poorest and more sparsely populated of the Irish provinces… a backwater whose affairs impinged little on the main course of Irish history until the spectacular and totally unexpected career of Toirrdelbach Ua Conchobair as high-king of Ireland in the 12th century’ (1973, 230). The extent to which a dearth of information on circumstances in Connacht reflects a state of political and cultural stagnation or merely the vicissitudes of transmission may be a matter for debate. If any significant annalistic activity took place in Connacht in the early medieval period, little of it survives. Our extant annals were begun on Iona in the early 7th century, expanded at Armagh and in or around Clonard in the 8th century, and continued in Clonmacnoise and Munster from the early 10th century. The contrast in treatment can be stark. For Féichín’s monastery at Fore, we have the names and death dates of nine ecclesiastics in the 8th century (abbots, bishops and others of unspecified office), five in the 9th, seven each in the 10th and 11th, and four in the 12th, with various additional references to attacks and other events. For High Island, we have one solitary entry, recording the death of Gormgal in 1018 (on whom see below). For Omey in the same period there is nothing. The area of modern county Galway west of Lough Corrib was inhabited by the Conmaicne in the early medieval period. Like the Luigne, Gailenga and other aithech-thuatha, these were a branched people, with the Conmaicne Réin situated south of Lough Allen on the eastern bank of the River Shannon and the Conmaicne Mara (‘of the sea’) giving their name to the Connemara. Lesser branches include the Conmaicne Cúile Tolad in the barony of Kilmaine, and the Conmaicne Dúine Móir in the barony of Dunmore. The Conmaicne Mara may have felt an affinity for Féichín as the patron of another subject people, geographically divided; perhaps they were introduced to his cult through communication with the Conmaicne Réin, situated as they were between the two branches of the Luigne. And we cannot rule out the possibility that Féichín was originally a distinct saint of local origin, whose identity later became confused with and merged into that of the more prestigious midlands saint. The name was not so rare: the Martyrology of Gorman (Stokes 1895), for example, in addition to Féichín of Fore, records the feasts of Féichín moccu Cáinche on 19 February, two priests named Féichín on 22 February and 2 August respectively, and a Féichín léir ua Lugba on 28 December. Whatever merited their inclusion in the martyrologies, nothing else is known of these men. One feature shared by the monasteries of High Island and Fore that may provide a clue for their common patron saint is the presence of a watermill. Colin Rynne (in White

Marshall and Rourke, 2000, 185–213: 201–2) notes that the vast majority of watermills before the 10th century in Ireland were non-ecclesiastical, though noting watermills described in the 7th-century life of Brigit by Cogitosus and the later life of Moling (p.197). In Irish hagiography, Féichín seems to be unique in building a watermill and then using miraculous power to create a mill-race (V §14, B2 §39). (Similarly, B1 §17 refers to Féichín creating a well on Omey using miraculous powers.) The story is retold elsewhere in the life of Mochua, where Mochua is present at Fore and partakes in the miracle (Stokes 1890, 283). This is a rare appearance of Féichín in another saint’s life, and the circumstances of him building a watermill might suggest he was particularly associated with this miracle. Féichín’s power is later associated with his mill in an episode told by Giraldus Cambrensis (Topographia Hibernica, 1186–7; Brewer et al. 1861–91, vol. 5, p.134). When an archer in Hugh de Lacy’s army attacks and violates a young girl in the mill at Fore, he afterwards suffers a dreadful inflammation on his limbs and dies. Féichín’s power over water remains prominent in the folklore of the area to this day, with the ‘seven wonders of Fore’ including water that flows uphill, water that will not boil and a mill without a race. If these reflect an early association of Féichín with water-milling, he would have been a natural patron to call on during the construction of the mill on High Island, and its successful operation would certainly have ensured his continued devotion.

Féichín in Scotland?

There is some evidence for a cult of Féichín further afield, perhaps even among the Picts of Scotland.4 The name of the parish of St Vigeans, on the east coast near Arbroath, is commonly regarded as a Latinisation of Féichín. Vigeanus for Féichín is a curious Latinisation, given that the name is rendered Fechinus in the Latin lives. However, it is possible that the spelling v/u for f might have been taken on analogy with other name pairs, such as Uinniau and Finnian (see Clancy 2001). Likewise, g for ch occurs also in Irish sources, as in the spelling Carthagus for Carthach in the Dublin collection of saints’ lives (though this may be a special case, influenced by word play with Carthage; see Harvey 1999). Nothing appears to be known about the early history of St Vigeans, but it is particularly significant that St Vigean’s feast is celebrated on the same day as Féichín’s (20 January). This would imply, at the very least, that Vigean and Féichín came to be associated historically, whatever of Vigean’s actual origin. How far this association goes back is difficult to say. St Vigean is listed in the Dunkeld Litany, a post-Reformation document that may, in parts, date from

4 The Scottish associations with Féichín are discussed in detail in Taylor (forthcoming). The author very kindly allowed me to read a draft.

high island and the cult of saint féichín 23

the reign of the Pictish king Giric mac Dúngaile in the late 9th century (878–9; see Woolf 2007, 120). If indeed Féichín’s cult was brought to Pictland, it was probably by the 9th century. The large body of elaborately carved Pictish stones that survives from St Vigeans suggests that the monastery flourished around this time. Alex Woolf (2007, 312) notes that the production of Pictish sculpture had seriously declined by the 10th century. Other placenames provide clues that Féichín was venerated in Scotland. Ecclefechan in Dumfriesshire might be interpreted as Eaglais Fhéichín, ‘church of Féichín’, although, given its location in the far south of Scotland, it might arguably reflect a Brittonic origin (cf. Middle Welsh eglwys fechan, ‘small church’). Dumfriesshire was within the historical Brittonic kingdom of Rheged, annexed to Northumbria by the 8th century. As such it would not have been a particularly Gaelic-speaking area, though may have been influenced by the Gaelic speakers in the kingdom of Galloway, to the west, that emerged between the 9th and 11th centuries. Lesmahagow, a town near Lanark, south of Glasgow, has been interpreted as Lios Mo-Fhégu, ‘enclosure of Mo Fhégu’, incorporating his hypocoristic name. (Mac an Táilleir prefers the explanation Lios Mo Chuda, from Eaglais Mo Chuda, ‘Mo Chuda’s church’.) Torphichen, in West Lothian, may be tórr Féichín, ‘Féichín’s hill’. The Scottish evidence remains uncertain. While some connection with Féichín seems very likely, how far this goes back is less clear. He may have been venerated in Scotland from an early date (say, at St Vigeans in the 9th century). Alternatively, his name may have been associated with originally local cults only at a much later date. What stands out however, in relation to sites such as High Island, Cong, Termonfeckin and elsewhere, is that the extant lives provide only a limited perspective on the extent of Féichín’s cult.

Historical contexts Ecclesiastical relations The nature of Church organisation in early medieval Ireland has been the subject of much recent discussion and revision (the most extensive treatment of the question is Etchingham 1999b.) The older prevailing view was that the earliest Christian mission in Ireland attempted to create a diocesan structure based on episcopal authority, modelled on the territorial organisation found elsewhere in Christendom, which was itself based on the late Roman administration. However, the authority of bishops is seen as having been eroded in favour of abbots during the 6th century, and this occurred due to the vigorous expansion of monasticism in the early days of the Irish Church, the high esteem afforded to ascetic life and the absence

of urban centres on which the Roman infrastructure was founded. Moreover, as new monasteries were founded from mother houses and others grouped together in common interests, broad federations were formed, united under devotion to a common patron saint (érlam; see Charles-Edwards 2003). Through this process, the abbots of major monasteries came to hold sway over ever larger territorial areas. More recent work has drawn attention to ambiguities and inconsistencies in the evidence, and suggests that the reality was more complicated. Both Latin and vernacular Irish law tracts acknowledge the superior authority of bishops, who are now recognised as having a central role in the organisation of pastoral care, and by the 10th century are forming their own hierarchy. There was clearly a good deal of variety with respect to roles and offices within the church. Abbots might sometimes also be bishops, although in other cases may not even have had clerical orders. A third office of importance was that of coarb (comarbae), the person responsible for the administration of the monastic estate. Although generally the function of a layman, the coarb could also be an abbot, and in principle there was no barrier to a single person fulfilling all three roles. (The head of a community is sometimes referred to as an erenagh, Irish airchinnech, ‘leader’, seemingly used ambiguously with regard to function.) All of these positions were liable to control by a single family group, and the office of coarb, in particular, came to be regarded as hereditary. Overall, the picture still emerging points to diverse sources of authority, whether deriving from clerical office, the prestige of a large and ancient monastic foundation, or the individual authority of a celebrated anchorite or scholar. We have very little evidence for affairs among the Conmaicne Mara in the early period, with no certain annalistic references before the death of king Muiredach son of Cadla in 1016 (AI 1016.8). An earlier reference may be from 663/4, with the death of Baetán moccu Corbmaic, who was abbot of Clonmacnoise (AU 664.5, CS 663, AT 664.4, FM 663.2). CS records that he was of the Conmaicne Mara in a gloss, while the later source FM includes this detail in the text. John Ryan (1940, 476) remarked that his designation moccu Corbmaic refers to the Uí Chorbmaic or Dál Corbmaic, ‘the name of a distinguished sept in Leinster and smaller septs elsewhere’. Given that the mention of the Conmaicne Mara is probably late, and there is no other evidence for any such group in Connemara, one wonders whether Baetán was in fact of the Leinster Uí Chorbmaic, with some confusion arising from the similarity of the population names. Far from being remote and isolated, however, the monastery at High Island was part of a network of early ecclesiastical sites along the western seaboard. Between

24 high island: excavation of an early medieval monastery

Sligo and Clare, Gwynn and Hadcock (1970) list foundations on Inishmurray, Inishglora, Inishkea, Duvillaun, Caher Island, Inishturk, Inishbofin, Inishark, Inishnee, St Mac Dara’s Island, the three Aran islands, Enniskerry (Mutton Island) and Bishop’s Island. Inishglora was devoted to Brendan, Inishkea to Columcille, Caher to Patrick, Inishark to Leo. The most eminent monastery, however, was Inishbofin, founded by Colmán, the former abbot of Lindisfarne in Northumbria, who brought relics to the island and established a monastery there in 668. Colmán was styled bishop of Inishbofin in his death notice of 676 (AU 676.1, CS 676) and his successor, Baetán, was also bishop (AU 713.1). These are the first two recorded bishops in Connacht. Thereafter the only bishops recorded in Connacht were at Mayo (Inishbofin’s daughter-house) in 732 and 773, until the designation ‘bishop of Connacht’ appears in 969 (contemporary with similarly territorial designations in other parts of Ireland). It seems hardly likely that Colmán’s episcopal jurisdiction was limited to an island as small as Inishbofin. Rather, he must surely have had authority within the territory of Conmaicne Mara at the very least, and perhaps even Connacht generally, given the absence of other bishoprics and the relatively small population in the region. In either case, Inishbofin would have had some significant authority over the foundation at High Island. Another source of authority over the island would have derived from its broader monastic affiliation, and the head of a monastic federation dedicated to St Féichín would certainly have been at Fore. Féichín’s mission to Omey is a significant event in the lives, and in hagiographical narrative the interactions between a patron saint and individual monasteries (founded or visited) are often read as expressing historical relationships at the time of writing. It seems then that, around the 13th century at least, Fore laid claim to authority over Omey. B1 (§17) refers to Omey in terms of its obligations: ‘For God hath granted to thee their tribute and their due…’. Given the proximity of High Island to Omey and its shared patron, we might infer that it too may have come under Fore’s influence.

Vikings The first phase of occupation of the monastery, until its apparent abandonment around the 9th century, coincides with the period of Viking raiding across Ireland and Europe. (See Etchingham 1996 for a detailed treatment of early Viking raids.) Following their initial raid on Rechru (probably Rathlin Island) in 795, Norwegian Vikings attacked Iona in 802 and again in 806 (killing 88 of the community), and in the following decade reconnoitred the coastline of Ireland, moving initially along the north and west. In 807 they burned Inishmurray, and rounded the west coast to invade Roscam (east of Galway city) in the

same campaign. We do not know what other monastic sites were targeted until attacks on monasteries at Howth, Cork and Skellig are recorded for 821, 822 and 824, respectively. We must presume, however, that the Nordic seafarers were thorough in their reconnaissance of the islands, encouraged by the opportunities for looting and the general abundance of monastic sites. Ongoing Viking activity in the region is reflected in episodes of local resistance. The Conmaicne were defeated in an engagement with the Norsemen in 812, and in the same year the Fir Umaill (based around Clew Bay) successfully defended their territory, though they were afterwards themselves routed in 813. The 830s mark a new phase of Viking activity, during which longphorts and other temporary settlements allowed the invaders to over-winter, and raiding gave way to campaigning, with incursions penetrating far inland. The intensive campaigns of the mid-9th century, renewed after a brief hiatus in the first half of the tenth, were centred on the rich monasteries of the eastern plain. Connacht during this time seems to have emerged relatively unscathed. The recorded raids on Connacht appear to have been mounted from bases on the Shannon, most notably Lough Ree, far from the western seaboard. That is not to say that the islands were immune. The Norsemen were active on Lough Corrib in 929, and ships must have sailed up the western coast periodically at least. During the 9th or 10th centuries, the monks of Inishmurray abandoned their island and united with the coastal monastery of Aughris (in Sligo), while the mixed community at Inishglora transposed themselves to Cross (north of Westport), presumably to avoid further slaughter (Gwynn and Hadcock 1970, 387). Other small monasteries and churches may have been abandoned, perhaps intermittently. The relative paucity of references to Viking activity on the western seaboard may in part be due to the neglect of contemporary chroniclers. More likely, however, Viking focus was diverted to the richer spoils available elsewhere, leaving their earliest targets relatively unharassed. High Island may have fared better than other island monasteries during the period, given the difficulty of landing there. Nevertheless, a small community with no possibility of escape would have been easily decimated in any attack, and with the loss of knowledge and experience recovery could have been slow. Indeed, the discovery of hearth debris within the church sanctuary may indicate that the pagan Norse themselves occupied the island for a time. (Note, however, that there were no finds at the site which corroborate this suggestion.) It is significant also that excavation at the monastic settlement on Omey indicated that burial activity there was not continuous, perhaps reflecting a shifting pattern of ecclesiastical stability in the area (O’Keeffe 1994b, 17).

high island and the cult of saint féichín 25

Gormgal If the foundation on High Island scarcely entered the historical record, either because of its lowly rank within a monastic federation or because of the absence of a bishop, the personal authority of one member of the community ensured a mention in the annals, in 1018: Gormghal in Ardailean, prim-anmchara Erenn, in Christo quieuit (‘Gormgal of High Island, chief anmchara of Ireland, rested in Christ’, AU 1018.1; cf. CS 1018, AI 1018.3, FM 1017.2). The name derives from gorm ‘(dark) blue’ (sometimes ‘illustrious’) and gal ‘valour, vigour’ (cf. the name Fergal, ‘manly valour/vigour’). Gormgal’s death is commemorated under 5 August in the Martyrology of Gorman, composed by Maél Maire hua Gormáin, abbot of Cnoc na nApstal (Knock, near Louth), around 1169: … Dunsech, Eche, Ernín, Gormgal minn nos-molab do domhan ’gá degrigh (Stokes 1895, 150), probably best translated: ‘… Dunsech, Echi, Ernin, venerated Gormgal, I will praise them to the world and its good king’. Stokes’s somewhat bizarre translation runs ‘Gormgal (who is) with her good King, a sacred thing I shall declare her to the world’ (Stokes 1895, 151). This would appear to be influenced by a Latin gloss on the manuscript, which Stokes ascribed to John Colgan (Stokes 1895, l): Sacramentum quod praedico coram mundo suoque optimo Rege (‘A sacrament which I predict in the world and with its great king’). This is to misunderstand minn, which can refer to a venerated object (sacramentum), but in this case applies to a revered person, ‘Gormgal minn’ (see examples in DIL s.v. 1 mind). In his index to the Martyrology of Gorman, Stokes suggests that the Gormgal there may be ‘the Gormgal mentioned by FM. A.D. 794, as the successor of Faendelach in the see of Armagh’ (Stokes 1895, 368). However, his date of 5 August corresponds to that of Gormgal of High Island, as cited in the Martyrology of Cashel. Gormgal’s date is confirmed in a reference to Gormgal of High Island in the Martyrology of Cashel (Ó Riain 2003, 162–84; see Stokes 1895, xvii), now known only through citations in John Colgan and Mícheál Ó Cléirigh. Ó Riain (2003, 163–5) regards its place of compilation as Lismore, no earlier than the mid-1170s. John Colgan also refers to a metrical eulogy for Gormgal which he had, but which is no longer extant. He gives the author as Corranus (1645, 141) or Cororanus (p.715), who lived around Gormgal’s time. (Colgan includes his notes on Gormgal with the life of Enda of Aran, as he confused High Island with Inisheer.) At his second reference, he gives the names of other saintly hermits of the same island who rest there with St Gormgal: Máel Suthain, Célechair, Dubthach, Dúnadach, Cellach, Tressach, Ultán, Máel Martain, Cormac, Condmach ‘and many more’ (et alii plures). Kenney (1929, 459) took this Corranus/Cororanus to be Corcrán of Lismore, who died in 1040. The eulogy

may account for the inclusion of Gormgal in the Martyrology of Cashel, in fact composed in the same location, and otherwise containing saints of earlier dates, c. 500–650. The term anmcharae (often translated ‘soul friend’) means ‘spiritual advisor’, and more specifically ‘confessor’. (Colgan translates prím-anmchara as Synderus, siue Spiritualis Pater, ‘Confessor or spiritual father’; 1645, 715; similarly p.141.) It is possible that by the 11th century the term may have referred to a specific monastic functionary. (Clonmacnoise, for example, has obituaries for successive anmchairdea in FM 1017.4, FM 1022.5, CS 1024, FM 1056.4, AU 1060.4, FM 1081.2, and so on.) However, there the term prím-anmcharae Érenn, ‘chief anmcharae of Ireland’, may best be taken as an expression of high esteem, rather than participation in any broader hierarchical structure. (Compare, for example, the similar designation (prím)ancharae Érenn, ‘(chief) anchorite of Ireland’, found in annals from the 10th century, which hardly implies the existence of any league of anchorites.) The emphasis on anmchairde may give some insight into the character of the monastic community during Gormgal’s time. Not only has the traditional organisational model of monastic over episcopal authority been challenged in recent years, but ideas about the nature of monastic communities themselves are also undergoing revision (see Etchingham 1999b, esp. 290–318, summarised in his 1999a article). The Irish term manach (Latin monachus) can mean ‘monk’ in the conventional sense, someone who has embraced religious life either as an eremite, living alone, or as a coenobite, member of a community following a strict rule. However, in Irish sources it can also denote a legal relationship similar to that between a client and lord, each with reciprocal obligations. (This duality in fact mirrors that of the erenagh, who may have the spiritual role of abbot and/or the secular functions of coarb.) The texts known as Penitentials lay down penances for a multitude of sins, and those guilty of the most serious (especially murder) could be prescribed a period of penance in exile. This would have entailed enrolment in a monastery, under the supervision of an abbot, who is sometimes styled anmcharae. (It has been argued that this type of penance was characterised in Irish as glasmartrae, ‘green/blue martyrdom’, in contrast to ‘white martyrdom’ (regular monasticism) or ‘red martyrdom’ (death from persecution); see Etchingham 1999b, 292–3.) Gormgal’s esteem as an anmcharae may well have derived, not from the quality of his spiritual direction, but from the number of penitents who submitted themselves to his authority. High Island would certainly have been an ideal location for austere penitential practice. Gormgal’s tenure seems to coincide with the re-establishment of the monastery from the late 10th century, and as such he may have been instrumental in its reinvigoration. The dating of two

26 high island: excavation of an early medieval monastery

of the skeletons recovered is compatible with his death date, and the elaborate treatment of those graves strongly suggests that one of them contained Gormgal himself. His far-reaching reputation may have led to the development of the island as a centre for pilgrimage (itself a form of penance). John V. Kelleher, in a personal correspondence to Richard Murphy (9 April 1971),5 suggested that Máel Suthain may have been the notary of Brian Ború, who recorded Brian as imperator Scotorum (‘emperor of the Irish’) in the Book of Armagh during the king’s northern circuit in 1005, and is sometimes described as Brian’s anmchara. Kelleher also thought that the names of the other hermits were found in significant concentration in the Dál Cais genealogies. On that basis, he suggested that Gormgal may himself have been Brian’s anmcharae, and that Brian probably made a pilgrimage to High Island. Kelleher’s speculations (which, it should be acknowledged, were not intended for publication) seem unfounded. The anmcharae who accompanied Brian is most likely to be Máel Suthain Ua Cerbaill, the ecclesiastic who died at Aghadoe (near Killarney, County Kerry) in 1010 (see Charles-Edwards 2004b). The other Máel Suthain who died in 1031 is credited as being Brian’s anmchara in one annalistic source only (FM 1031.2, not AI 1031.4). This may be a late gloss. Moreover, this Máel Suthain must have been considerably younger: he outlived Brian by 17 years, and Brian was already aged around 73 at the time of his death in 1014. In any event, the annals record two other men of that name (see FM 1031.2, 1125.3), and there is every likelihood that there may have been others, the Máel Suthain of High Island among them. Moreover, an examination of the genealogies published in CGL shows no special correspondence between Colgan’s names and the Dál Cais genealogies. Condmach, Dubthach, Dúnadach, Máel Martain and Ultán do not occur in the Dál Cais lists at all (though they are found elsewhere), while Cellach and Cormac are so ubiquitous as to be insignificant. Kelleher suggests that Máel Martain may be a devotee of the Dál Cais saint Martán; he could also have been named for St Martin of Tours, an important influence on Irish monasticism. Brian’s alliances in Connacht were with neighbouring groups in the south-east of the province: the Uí Fiachrach and Uí Maine (who both supported him at Clontarf in 1014). Any Dál Cais influence in the far west of the province seems inherently unlikely. Another tenuous connection may be found in Muirchertach Mac Líacc, a poet styled ard-ollam Éireann (‘leading chief poet of Ireland’) at his death in 1014, and sometimes associated with Brian Ború in later sources. Smith (2004) writes that a marginal note in an unspecified 5 I am grateful to Georgina Scally for putting this letter at my disposal.

chronicle records that ‘Mac Líacc was under a (monastic) rule in Ard Oilén (“high island”, Galway) when he died’. I have been unable to locate the note in question.

Change and decline Although we have no further reference to High Island after the notice of Gormgal’s death in 1018, the movement for Church reform that gathered momentum a few generations thereafter must have had a significant impact on the monastery. Ultimately precipitated by Pope Gregory VII (1073–85), the reform movement was introduced to Ireland initially through links between Hiberno-Norse settlements and the Archbishops of Canterbury. The main objectives of the reformers were to curb lay involvement in ecclesiastical affairs, and in particular the function of lay erenaigh, to impose a Roman model of territorial organisation, based on dioceses and parishes under an episcopal hierarchy, and to regularise various other matters, such as clerical celibacy and payments of tithes. In 1111 the Synod of Ráith Bressail established 24 new dioceses under two archbishops at Cashel and Armagh, afterwards increased to four, with Tuam and Dublin, at the Synod of Kells in 1152. These changes had a major impact on the established monasteries, the renewed and reorganised episcopacy precipitating a decline in abbatial authority and the redirection of monastic revenues. Moreover, the prestige of the ancient foundations was challenged by new monastic orders, recently established on the Continent and introduced into Ireland under the influence of St Malachy and others: initially Cistercians (Mellifont, 1142) and then Augustinian canons, followed in the 13th century by orders of Franciscan, Dominican, Augustinian and Carmelite friars. High Island was initially brought under the jurisdiction of the new see of Cong (1111), although after Kells (1152), Connemara came into the see of Tuam. In the intervening years the first major imported order was established at Cong, which Turlough O’Connor (d. 1156) refounded for the Augustinian canons (presumably after its burning in 1137; see Gwynn and Hadcock 1970, 146, 166). In the mid13th century the O’Flahertys, the ruling sept of the Uí Briúin Seóla, were displaced from their homelands east of Lough Corrib by Richard de Burgo, and resettled in Connemara. We can only speculate on what impact the political displacement of the Conmaicne Mara had on the patronage of monasteries such as High Island. The O’Flahertys established a new foundation of Carmelite friars at Ballynahinch in 1356, and later founded St Patrick’s Priory at Toombeola for Dominican friars after 1427 (Gwynn and Hadcock 1970, 287, 230). Omey, at least, was occupied into the 14th century given notices in the annals to an Ó Ferghusa, vicar

high island and the cult of saint féichín 27

of Imaidh (Imaidh Féichín? AU 1359.1) and an Ó Tuathail, a vicar of Imaidh Féichín, who kept a celebrated house of hospitality (FM 1395.3). Roderic O’Flaherty wrote in 1694 that of Féichín’s former monastery at Omey, only the parish church was extant (Hardiman 1846, 113).

Conclusion This attempt to survey some of the historical contexts and associations of High Island points up the very scanty and uneven nature of the extant sources, particularly with regard to Connacht. Our most extensive sources, the lives of St Féichín, have a devotional rather than historical perspective. It seems clear that their principle focus was Fore, and they may well have been composed there (they were certainly copied nearby in the area of Saints’ Island, County Longford). The interest in Connacht is mostly

confined to Féichín’s homeland among the Luigne. While Omey has an important place in the narrative, its treatment has little historical resonance, while other, more significant sites, such as High Island, Cong and Termonfeckin, receive at best incidental mentions. Nonetheless, High Island was clearly an important place during Gormgal’s time, and he must have been instrumental in the renewal of the community after its hiatus during the period of Viking raids. We may infer that by Gormgal’s time the island had become a centre for penitence and pilgrimage, which seems to have gathered pace after his death. The following century brought great change to the Irish ecclesiastical establishment, and the absence of any mention of the island in the 13th-century lives may indicate that by this time the monastery had already been abandoned.

3 the later history of high island and its antiquarian and archaeological rediscovery 3.1 Roderic O’Flaherty’s account, 1684 Ownership of High Island during the later medieval and early modern periods is unclear. On a number of maps of late 16th-century date, High Island is marked as ‘Ardelen’. The earliest of these maps appears to have been the 1591 map of Connacht by John Browne II (Andrews 2003, Pl.V). It is probable that at some stage the island came into the hands of the O’Flahertys, the secular rulers of Iarchonnacht (Connemara) until their defeat and dispossession by the Cromwellians in 1652 (Robinson 2008, 203). Roderic O’Flaherty (1629–1718), heir to the southern part of Iarchonnacht (the barony of Moycullen), was deprived of most of his ancestral lands during the Cromwellian campaign. This eminent scholar, historian and antiquarian prepared a description of Iarchonnacht for the Dublin Philosophical Society in 1684, though it was not published until James Hardiman’s edition appeared in print in 1846. O’Flaherty relies on John Colgan’s hagiographical text of St Féichín and Gormgal (see Section 2). His work includes the earliest known, albeit very brief, description of the monastery. He mentions that the church, the ‘large round wall’ of the monastic enclosure and a beehive hut, or clochán, were then upstanding (O’Flaherty 1846, 114– 15). He refers to the well, named after Brian Boru, and ‘a standing water, on the brook whereof was a mill’. Many parcels of land in Connemara were granted to Richard Martin in the Cromwellian and Restoration settlements (Mac Giolla Choille and Simington 1962, 14–22). Among those in the parish of Omey were ‘Island[s] in the sea not found in the plott’. Five islands are listed, of which the most easily identified is Turbot, assessed at 25 acres profitable and granted to Richard Martin. Four other islands are listed as follows: ‘Creelan alias Arctittant, Longbay, Lambay, Muckeogh’, as together comprising 50 acres unprofitable and 50 acres profitable, of which ten acres were granted to John Browne and 50 acres to Richard Martin. No names are given for the original proprietors of these islands (ibid., 22). The four islands involved are likely to be High Island, Cruagh, Inishturk and maybe

Friar Island, all of which are in Omey parish. The names given are problematic, however, some being badly garbled and some possibly even invented. Arctittant, otherwise a meaningless and highly unlikely name, must be a badly garbled transcription of something like ‘Ardillane’, where the d has been wrongly transcribed as ct, the ll as tt and the final e, possibly with a flourish, as another t. The original alphabetic list gives it as ‘Arclittan’ (ibid., 7). This all suggests that Martin ownership of High Island dates back to the late 17th century (C. Manning, pers. comm.).

3.2 George Petrie’s account, 1820 George Petrie (1789–1866) was the next person to record the antiquities on the island. Petrie was employed by the Ordnance Survey of Ireland and for a significant part of the period of his employment (1833–1846) was responsible for antiquarian and orthographical investigations in Dublin and other parts of the country. However, Petrie’s visit to High Island in 1820 was unconnected with his official duties and the publication of his private research did not appear until 25 years after his visit (Petrie 1845). Petrie’s account gives accurate descriptions and detailed measurements and he comments on the possible uses of the various structures that he encountered. His dimensions are remarkably precise, but his height measurements are incorrect as these would have been taken when the actual floor levels were obscured with masonry rubble. His survey is invaluable because he records structures which have since collapsed, suffered damage or been covered over. In Petrie’s time, the church was in a much better state of preservation and since then the east wall, including the only surviving window, has fallen down. He provides a sketch of the window and remarks that the opening, 1ft (0.35m) high and 6 inches (0.15m) wide, has a semicircular head (Fig. 3.1; Petrie 1845, 420). He records an important grave situated to the east of the church, constructed of large slabs of mica-schist, with the end-stones

the later history of high island and its antiquarian and archaeological rediscovery 29

Fig. 3.1 Sketch of remaining northern jamb of interior opening of the east window of the church by John O’Donovan, 1839 (courtesy of the Royal Irish Academy).

carved with crosses and a human figure. The defaced surface of the covering slab on the grave was also decorated and may originally have had an inscription, according to Petrie. He records that the enclosure surrounding the church led outside, via a covered passage 15ft (4.57m) long, northwards to a beehive hut, or clochán, which he suggests was the abbot’s dwelling (Fig. 5.1, Cell A). He notes that the beehive hut was nearly circular in plan, dome-roofed internally, 8ft (2.43m) high and that it was built without mortar, like those on the Aran Islands in Galway Bay, to the south of High Island. Of the larger beehive hut on the eastern side of the church enclosure, he describes it as round on the exterior but square on the interior, 7ft 6 inches (2.28m) in height, and he proposes that its purpose may have been as a refectory (Fig. 5.1, Cell B). The Pl. 3.1 The mine shaft opening, looking south (G. Scally).

fact that Petrie recorded the internal heights indicates that both cells were well preserved at the time of his visit. It is significant that he also refers to several other beehive huts on the other side of the church, which are no longer visible today, though they were already largely covered over in 1820. These huts, he suggests, were only large enough to accommodate a single person as they measured 6ft (1.83m) long by 3ft (0.91m) wide by 4ft (1.2m) high. He observed that they formed a Laura, in a similar manner to the earliest cells of hermits in Egypt. Petrie notes that the monastic enclosure wall is without mortar bonding and is nearly circular in outline (Petrie 1845, 420–21). He records the stone passage through the wall from the entrance at the south-eastern side (Fig. 5.1). He also notes that the buildings to each side of the entrance on the outside were probably for use by pilgrims and that, though their walls are built of stone, their roofs had been of another material. Petrie’s assiduous survey also describes a ‘covered gallery or passage’ within the enclosure wall at the west, its entrance doorway measuring only 2ft 3 inches (0.69m) in size (Fig. 5.1, wall chamber to south-west). Other structures of which there are now only scant traces were seen by Petrie. He notes that to the south of the lake ‘an artificial outlet is formed, which turned a small mill’ (Petrie 1845, 421). Beside the lake there was a stone path or causeway, 220 yards (201m) in length, leading to a stone house to the south of the monastery. This house, 18ft (5.49m) long and 9ft (2.74m) wide, had a small walled enclosure adjoining it, which he thought was most likely a garden, and close by there was a stone altar with a cross-slab on top.

30 high island: excavation of an early medieval monastery Fig. 3.2 Ordnance Survey Fair Plan 1838. Generally, red indicates structures, those in outline indicate in ruins. Faint pencil sketch profiles in margins denote heights in various parts of the island; profile E–F, orientated roughly N–S, notes a height of ‘about 200 feet’ (courtesy of the National Archives of Ireland).

3.3 The 19th-century copper mine and the miners In the late 18th century, High Island is known to have been owned by the Martin family, one of the fourteen great family tribes of Galway and one of the largest landholders in Connemara. The Land Registry Office records a lease of Friar Island and High Island (dated 1794) from Richard Martin (better known as ‘Humanity Dick’), of Ballynahinch Castle, to John Bodkin, the tenant (Marshall and Rourke 2000, 224, note 228). The island was still in the ownership of the Martins when the copper-mining campaign of the 1820s was undertaken. These miners are the only recorded inhabitants of the island since the monks’ departure several hundred years earlier. Whilst the impact of the mining on the island and on the monastery was considerable, this endeavour does not appear to have lasted long. Previous writers have indicated a date around 1828 for the mining campaign (Herity 1990a, 67; White Marshall and Rourke 2000, 12). More recent research by Tim Robinson has unearthed a letter written by a member of the Blake family, of Renvyle, who mentions that, while becalmed during a sailing expedition in August 1823, they had ample leisure to observe ‘the remaining traces of the copper-mine, which had been opened in High Island’ (Robinson 2008, 203). It would seem, therefore, that the mining campaign on High Island began shortly before 1823. No documentation has been found to indicate how long it lasted. It has been suggested that activities were restricted to the summer and early autumn on account of the difficulty of landing at other times of the year, and it is

likely that mining lasted no more than one or two seasons (White Marshall and Rourke 2000, 27). Small-scale copper-mining was not an unusual occurrence on the estates of 19th-century Ireland. The lode on High Island must have been of sufficient potential for the Martins of Ballynahinch to invest time and resources in establishing a mine there. The only published account of the mine, written by the GSI geologist George Kinahan, dates from his visit to the island in 1869, where he states that a north by south lode across the north-east of the island was sunk by Colonel Martin (Kinahan et al. 1878, 162). The following account was given by John Kelly, Glan, Oughterard, who had worked at it: ‘at nine fathoms there was in the lode stuff a good mixture of copper; also at fifteen fathoms; after this no good ground was met with, although the shaft was sunk to twenty-five fathoms’. Research has uncovered a previously unknown letter, written 17 November 1906 by P.H. Joyce to the GSI, which sheds some light on the difficult conditions the miners encountered and the reason for the abandonment of the mine (Joyce 1906). Joyce noted that activities yielded copper in profitable quantities until the shaft was sunk below sea level and the water ‘pushed in’, causing the abandonment of operations. It is also noted that a site further inland would have been more suitable than the location chosen, on the narrowest part of the island (Fig. 4.1, no. 4). The letter states that Joyce was writing on behalf of his wife, the owner of High Island, enquiring about the possibility of restarting the mining campaign of nearly a century earlier. Nothing came of this proposed endeavour. The mouth of the mine shaft (2.5m x 1.5m) is still

the later history of high island and its antiquarian and archaeological rediscovery 31

Pl. 3.2 The restored miners’ hut in foreground, with remains of the second hut in the background and cultivation ridges visible on the right (Photographic Unit, NMS).

visible close to the cliff-face on the north-eastern side of the island, just above the south landing (Fig. 4.1, no. 4; Pl. 3.1). The shaft descends almost vertically for 15m, but below this it veers to the north-west, presumably following the vein of copper. The base of the shaft is still visible as an opening in the rock-face, and this is presumably the level at which the water is reputed to have ‘pushed in’ and forced the works to be abandoned. The base of a smaller, secondary shaft is visible in the face of the cliff on the southern side of the main shaft. A number of blasting powder holes (approximately 20mm in diameter) are evident close to the north landing (Fig. 4.1, no. 7). The miners built two stone huts for their accommodation a short distance inland from the shaft (Fig. 4.1, nos. 5, 6). The huts were already in disrepair by 1838, as they are shown as ‘Ruins’ on the Ordnance Survey Fair Plan of that date (Fig. 3.2). The larger hut was originally a single-storey, three-roomed building (17m x 6m) (Pl. 3.2). The western end of the hut was shortened (by 3m), and the eastern room was repaired and re-roofed when Richard Murphy bought the island in 1969. He subsequently used the hut during his visits to the island, and more recently it was the base during the excavations from 1995 to 2002. The smaller hut (approximately 6m x 5m) has survived only as a mound of rubble. Several cut stones visible in the walls of the larger cabin suggest that the stones used to build the

huts may have originated from buildings located within the monastery. Therefore, it is likely that the miners in the 1820s inflicted the first significant damage to the monastic buildings. A well located close to the larger hut provided the miners with a supply of water (Figs 3.2 and 4.1, no. 5). The opening (approximately 1m x 0.6m) is defined at surface level by large boulders and the stone face of the well beneath is evident on all sides. The well itself has silted up and is very overgrown. The site for the huts was carefully chosen. It was located at the base of elevated ground to the south-west, which gave shelter from the prevailing winds. Furthermore, the miners built a stone wall and ditch on the southern and western sides, thus providing further protection for their huts from the wind and, probably more specifically, protection from run-off water from the elevated ground behind. The remains of the wall and ditch are visible today as a low, grass-covered mound within which there are cultivation ridges, probably for growing potatoes (Pl. 3.2). Beyond the cultivated area, traces of another collapsed wall (15m long and approximately 2m wide), built with large stones, traverse the width of the island at its narrowest point (Fig. 4.1, no. 3). These remains are enigmatic and it is unclear whether they are associated with the miners or if they are something much earlier.

32 high island: excavation of an early medieval monastery

3.4 Later 19th-century accounts The destructive impact that the mining activities and the miners had on the ecclesiastical monuments is apparent from the writings of the antiquarians who subsequently compiled accounts of their visits to High Island. The over-

all impression given by John O’Donovan, who came there on 15 June 1839 as part of his duties with the Ordnance Survey, is of a monastery in poor condition. To obtain any clear impression of the structures, he had to clamber over and root around in layers of rubble. O’Donovan’s account

Pl. 3.3 Cell B by W.F. Wakeman, 1839 (courtesy of the Royal Irish Academy).

Pl. 3.4 The church with bee-hive Cell B in the background. Drawn by W.F. Wakeman, 1839 (courtesy of the Royal Irish Academy).

the later history of high island and its antiquarian and archaeological rediscovery 33

Fig. 3.3 Working copy Ordnance Survey Fair Plan with handwritten notes by John O’Donovan, 1839. Pencil sketch in margin on left-hand side is the earliest known plan of the site, annotated as follows: ‘The cloghans are connected by a cyclopean wall in this manner’. At lower left-hand side the pencil note reads: ‘Site of a mill here, not shown, see my letter. JoD’ (courtesy of the National Archives of Ireland).

Fig. 3.4 Ordnance Survey plan of High Island, Co. Galway, 1841, based on the survey by John O’Donovan (courtesy of the National Monuments Service).

in the Ordnance Survey (OS) Letters makes no mention of Petrie’s 1820 record (O’Flanagan 1927, vol. 3, 36–43). O’Donovan gives a lengthy account of High Island and intersperses his comments with O’Flaherty’s work, despite considering it to have significant shortcomings. Writing of his visit some days later, he made a number of observations that help to reconstruct the process of dilapidation that had occurred. Petrie’s publication contains an illustration of the beehive hut by William F. Wakeman, who accompanied O’Donovan (Pl. 3.3; Petrie 1845, 128). His

drawings also impart an image of a monastery in a ruinous state. O’Donovan spent just one day on the island, but Wakeman returned the next day to complete his drawings of the cross-slabs and he drew ten in all. Four of these are especially interesting as they are now missing; they were probably taken off the island in the intervening years. His work provides the most abundant visual record of the cross-slabs until Michael Herity published them, 140 years later (Herity 1977, 1987, 1990a). O’Donovan’s notes on the church show that the gable

34 high island: excavation of an early medieval monastery

of the east wall, intact at the time of Petrie’s visit in 1820, had collapsed, but that the window partially remained in situ. Another clue regarding the condition of the church at this time comes from one of Wakeman’s sketches. The drawing of one of the headstones shows it leaning substantially, indicating that the east wall of the church, within which the cross-slab was set, had subsided severely. Four beehive huts are described by O’Donovan, two of which are identifiable (Fig. 5.1, Cell A, Cell B; Pls 3.3, 3.4). He refers to a cloghán ‘north and by west of the church’ that was level with the ground, and another ‘near the northeast corner, fifteen feet [4.57m] long and twelve feet [2.29m] broad’. O’Donovan records that the monastic enclosure wall was in many places destroyed to ground level, in particular at the north, but that the wall was upstanding on its other sides (O’Flanagan 1927, vol. 3, 40). The best-preserved section was near the north-western corner, where it measured about 10ft (3m) wide. Notwithstanding this, the monastic enclosure wall does not appear to have suffered considerable damage since the time of Petrie’s visit. In contrast to Petrie, who mentions one entrance through the enclosing monastic wall, O’Donovan mentions three. There are, in fact, four entrances; the one not mentioned by O’Donovan was in the north-west (Fig. 5.1). O’Donovan makes only scant reference to the penitential stations that were associated with pilgrimage, saying that there are several of these on the island, but that he could not learn their names from the people he met there (Figs 3.2, 3.4). O’Donovan was concerned also with the larger issue of recording High Island for the Ordnance Survey. In 1838 he produced a Fair Plan and in the margins sketched a number of cross-sections of the island (Fig. 3.2). He used a copy of this plan to mark up various alterations he wished to make (Fig. 3.3). Interestingly, in the margins of the working copy, he made a sketch plan of the monastic enclosure wall and some of the beehive huts he had recorded at the monastery. The sketch shows Cell A, Cell B and another cell or building at the south-east entrance to the monastery (Fig. 4.2). The plan is not a true reflection of the layout of the monastery as O’Donovan appears to have confused the two walls – the church enclosure wall and the larger monastic enclosure wall – most probably as a consequence of rubble obscuring their outlines at the time of his visit. Regardless of these shortcomings, this is the first rudimentary plan of the antiquities on the island. The first six-inch scale OS map of High Island was published in 1841 (Fig. 3.4). Wakeman (1863, 1867) published two articles describing his trip to the island and the antiquities he found there. His observations of the graves are informative. He details that they were built of slabs placed lengthways and that on the stone at the end of each grave there was a carv-

ing of a cross within a circle (Wakeman 1863, 220). Of the beehive huts, Wakeman records that three were in perfect condition, in particular their interiors, and that there were several others in ruins (Wakeman 1867, 368). The roof of the principal remaining cloghán, he notes, was formed by the gradual sloping inwards of the side walls and a single flagstone closed the opening (Pl. 3.3). Wakeman was the first antiquarian to remark so specifically on the large number of penitential stations around the island (Fig. 4.1, leachta; Wakeman 1867, 367). He writes that these structures of dry masonry are generally quadrangular in form, usually surmounted with a small cross-slab, and are believed to be penitential stations. Changing circumstances from the mid-19th century onwards further contributed to the disintegration of the monastic buildings. The Great Famine of 1845–1852 had a devastating effect on the coastal mainland around High Island. The Martin estate was broken up and ownership of the island changed. It is thought that around 1849, High Island came into the ownership of the Rev. Anthony Magee (Robinson 1990, 50). Magee purchased over 1,000 hectares (2,500 acres) of the Martin estate and though

Fig. 3.5 Plan of the monastery on High Island by George Kinahan, 1869 (courtesy of the Royal Irish Academy).

the later history of high island and its antiquarian and archaeological rediscovery 35

most of the land he bought was between Boolard and Cleggan village, it included High Island and also Cruagh and Friar Islands and a nearby islet, Malthooa (An Meall Thuaidh) (Robinson 1990, 50; 2008, 204). In the closing decade of the 19th century, the island changed hands again and was acquired by the Joyces of Lenaboy Park, near Galway, who also purchased Friar Island and Malthooa (ibid., 2008, 204). By 1906, Mrs Joyce of Oughterard had purchased High Island (see Section 3.3). George Kinahan had made his first visit to High Island with the Royal Society of Antiquaries of Ireland in 1868 (see Section 3.3). As the Hon. Provincial Secretary for Connaught, he wrote a report on the state of the antiquities on the islands off the coast of Connemara, in which High Island was included (Kinahan 1868–9, 348). His account gives the impression that there had been further damage to the monastic buildings since O’Donovan’s survey 30 years previously. He describes the structures as ‘all shattered and broken’. This, he states, had occurred since the Martin ownership, when the ruins were in good state of preservation, but had been allowed to deteriorate from then on. He subsequently published a detailed account of the antiquities on High Island, together with a map of the south-western end of the island and the monastery (Fig. 3.5; Kinahan 1869). This was the first map showing the ecclesiastical remains. Despite inaccuracies in orientation, measurements and some details of construction, together with more significant errors, such as conjoining the church enclosure wall with the monastic enclosure and locating the beehive huts outside this latter wall, Kinahan’s records are useful in that he mentions features not referred to by previous writers. He was the first to depict a wall running the full width of the island, which effectively separates the monastery from the rest (Fig. 3.5; Kinahan 1869, 552). He describes the remains of the wall as a line of upright slabs, conjecturing that there is a gateway through the northern end at the cliff edge. Kinahan remarks on the passage from the church enclosure at the north-east to the clochán (the covered passage that Petrie had seen) and he further adds that it seems to have led to a doorway in the monastic enclosure wall (Kinahan 1869, 552). The fact that this passage is recorded by two antiquarians suggests that some sort of rudimentary passage had existed here, but was destroyed by the time the excavations began. Kinahan adds to the observations made by Petrie on the small, upstanding clochán (Fig. 3.5, no. 3; Fig. 5.1, Cell A), noting that the inside of the walls are vertical from the floor before coving in to form the roof (Kinahan 1869, 552). Of the larger upstanding clochán (Fig. 3.5, no. 5; Fig. 5.1, Cell B), he notes that the outside is in a regrettable condition, but the inside is intact, commenting on the excellence of the masonry workmanship from the floor to the roof, the apex

Pl. 3.5 Cross-slab 19 placed upright by G. Kinahan at the station by the lake (Photographic Unit, NMS).

crowned ‘by three flags placed in steps’ (Kinahan 1869, 553). Kinahan describes the large chamber in the western side of the monastic enclosure wall, and he was the first to note the smaller, previously unrecorded chamber in the same wall (Fig. 3.5, nos 7, 8; Fig. 5.1, 3.8; Kinahan 1869, 553). He describes the latter as rectangular-shaped, 9ft (2.7m) long by 4.5ft (1.37m) wide, 4ft (1.22m) high, and covered by large slabs extending almost across the thickness of the enclosure wall. Its entrance from inside the enclosure was through a doorway, 2.5ft (0.76m) high by 3ft (0.91m) wide. Kinahan notes that many of the cross-slabs had been removed from within the monastic enclosure and had been positioned at stations or wells in other parts of the island. The secretary of the society, who accompanied him, searched through the rubble and found two crossslabs, which he placed in upright positions at two stations adjoining the enclosure so that they would not be damaged further (Kinahan 1868–9, 348). It is not known at which stations these cross-slabs were placed. Kinahan mentions how he relocated the cross-slabs at the southeast landing and at the well, both in good condition, but others that he found lying about the place he had fixed upright at the station in the south of the monastic enclosure and at the station at the north-eastern shore of the lake (Fig. 5.1; Pl. 3.5; Kinahan 1869, 554–5). Both crossslabs remain in the positions in which Kinahan placed them (Section 7.1, Cross-slabs 19, 20). The monastery received official protection under Section 25 of the Irish Church Act 1869, when the Irish Church Commissioners, the owners of the site, were authorised to transfer it to the Secretary, Board of Works, ‘upon trust for preservation as a National Monument …’. Monuments at 137 locations were transferred on 30 October 1880, which included the ecclesiastical remains at High Island (No. 52 on the list), described as ‘Stoneroofed cells and Ruins of Church’ (Cochrane 1892, 416, cited in Herity 1990a, 71).

36 high island: excavation of an early medieval monastery

On 22 July 1890, the Rev. John Healy, Bishop of Clonfert (Archbishop of Tuam 1903–18), visited High Island in the company of two other Catholic clergymen, Father Lynskey and his curate, Father Biggins. As a result of this expedition, Healy published in the Irish Ecclesiastical Record a moderately detailed, if occasionally flawed, account of the island and its antiquities (Healy 1890). He describes the church and notes the destruction of the east gable (Healy 1890, 684). He attributes the deterioration in the condition of the ruins on the island since earlier accounts to shepherd boys who had knocked down the walls of the buildings on their forays to the island to hunt rabbits (Healy 1890, 683). He also notes that vestiges of the monks’ garden are to be seen, but he does not give the location (Healy 1890, 677). Most significantly, he refers to traces of a graveyard on the eastern shore of the smaller lake (Healy 1890, 680). Visual inspection of the area between the lake and the cliffs did not lead to the discovery of the graveyard. In the summer of 1895, the Royal Society of Antiquaries of Ireland organised an excursion to a number of offshore islands on the west coast of Ireland. One of the islands visited was High Island and R.A.S. Macalister was on that trip. Like the antiquarians who had preceded them, the weather curtailed their visit and they had just one hour to examine the remains. Despite this, Macalister’s account is one of the most complete of the 19th century and he draws on the work of Petrie and Kinahan (Macalister 1896). Moreover, he was the first to include photographs and he also illustrated several of the crossslabs. It is Macalister’s knowledge of and reference to other similar early medieval ecclesiastical sites in Ireland that sets his account apart and makes it the first significant comparative work. The clochán (Cell B) is compared with Clochan na Carraige on Aran Mór, Co. Donegal. He compares the partly obscured decoration on the crossslab lintel on the church doorway on High Island to the

lintel of Teampull na Teinidh, Inishmurray, Co. Sligo. He also draws comparisons between one of the decorated cross-slabs (Section 7.1, Cross-slab 21) and a cross at Teampull na Naomh on Inchagoill, Lough Corrib, and suggests similar examples may be found at Ballindhoor, near Knockboy, Dunbelloge, Co. Cork (Macalister 1896, 201, 207). Macalister records that the eastern end of the church and its window were practically destroyed. No mention is made of the graves outside or of the decorated crossslabs associated with them, which suggests that they were obscured by rubble from the collapsed gable. He is the first to mention the large, rough, irregular stones of the masonry of the west wall and he surmises that they differ in date from the rest of the church walls. He comments that the roof of the church was probably of thatch as the walls are too thin to support a stone roof (Macalister 1896, 198–9). Of the beehive huts, he records their recent destruction and neglect. The two upstanding huts within the monastic enclosure had been damaged, with many of the external stones removed, leaving holes in the walls and roofs (Fig. 5.1, Cell A, Cell B). The other beehive huts at the monastery were almost entirely flattened. Macalister noticed (accurately) the three enclosure walls, comprising the inner wall around the church, the second surrounding the monastery and the outer wall separating the southern end of the island from the rest (Figs 4.1, 4.2; Macalister 1896, 200–201). Like O’Donovan, Macalister notes just three entrances through the monastic enclosure wall, but states that the buildings at the southeastern entrance noted by Petrie were not observed by his party (Fig. 5.1). He describes the small chamber in the northern end of the western side of the monastic enclosure wall, but acknowledges that neither he nor any member of the party noticed the larger chamber adjacent to the south-western entrance, which was described by Kinahan (Fig. 4.2; Macalister 1896, 202–3). Fig. 3.6 Plan of ‘hermitage’ (after Herity 1990a, Fig. 30b).

the later history of high island and its antiquarian and archaeological rediscovery 37

3.5 Twentieth-century accounts High Island, Friar Island and Malthooa were sold to Graham Lushington-Tulloch of Shanbollard, Ballynakill, in 1951 (Robinson 2008, 204). Richard Murphy bought the three islands from him in 1969 and sold Friar Island and Malthooa, but kept possession of High Island. In 1999 he sold the island and, with the exception of the monastery, it is retained in local private ownership. In 1955 Daphne Pochin Mould wrote a brief descriptive overview of High Island and its monastery (Pochin Mould 1955, 50–53). Although her description adds little new to the corpus of information already published, she was the first to draw an association between the monastery and pilgrimage. Significantly, she was also the first to mention that the land on the island might have been cultivated. The 1970s and 1980s saw a growing appreciation of the island and its monuments by Michael Herity, whose research has resulted in an invaluable body of work. He began by conducting the first detailed survey of the monastery and of other structures on the island (Herity 1977). He also prepared a reconstruction of the tomb of the saint on High Island, which he later revised and included in a comparative study of founder saints’ tomb-shrines in Ireland (Herity 1987; 1993b, 185–95). He published articles on other monastic sites – Caher Island, Co. Mayo, Gleanncholmcille and Rathlin O’Birne, Co. Donegal – drawing links between these settlements and pilgrimage (Herity 1989; 1993a; 1995b). In articles written in the intervening Fig. 3.7 Map of High Island showing penitential stations (S1 to S8), enclosures and huts (E1, H1 to H4) (after Herity 1990, Figs 32, 34).

years, he placed High Island in the context of pilgrimage (Herity 1990a; 1995a) and his detailed paper, ‘The Hermitage on Ardoileán, county Galway’, discusses the 19th-century sources and provides a comprehensive assessment of the monuments at the monastery and on the island (Figs 3.6, 3.7; Herity 1990a). He describes all of the previously recorded features, but in greater detail than before as he had more time to conduct a thorough survey, which his predecessors did not. He also identified three unrecorded penitential stations: one in the monastic enclosure outside the entrance to the church enclosure; another located north of the smaller lake; and the third, Brian Boru’s well, he suggests was part of the early medieval turas and would therefore have been a station of the pilgrimage round (Fig. 3.6, S1; Fig. 3.7, S1, S5, S6; Fig. 4.1, nos. 11, 23; Herity 1990a, 85). High Island is mentioned briefly in Peter Harbison’s seminal work on pilgrimage, where he suggests that the path or causeway running alongside the lake, described by Petrie, could be a section of a pilgrimage road (Harbison 1991, 146). The entries for High Island in the Archaeological Inventory of County Galway draw heavily on the work of Herity and, in relation to the many cross-inscribed stones, the work of Higgins (1987, vol. II). A total of fourteen archaeological monuments are included in the survey, all of which have been described by previous writers, with the exception of a possible promontory fort at the extreme eastern end of the island (Gosling 1993, vol. I, 209, no.120).

38 high island: excavation of an early medieval monastery

The most recent and most extensive publication on the island is Jenny White Marshall and Grellan Rourke’s book, High Island: an Irish Monastery in the Atlantic (White Marshall and Rourke 2000). This substantial publication, the culmination of nearly twenty years’ work, combines survey with the preliminary reports of the first three years of excavation, from 1995 to 1997. Their survey includes a description and analysis of both the monu-

ments at the monastery and on the rest of the island. It also includes a summary of all known previous written accounts of the monuments on the island, including a review of the corpus of cross-slabs identified up to 1997. Their work still forms the most comprehensive study of the island as a whole and, until this publication, of the excavations at the monastery.

4 the monastery and its associated structures before excavation 4.1 The site of the monastery The monastery is located in a sheltered valley at the southwestern end of the island (Figs 4.1, 4.3; Pls 4.1, 4.2).1 To the east and west, there is elevated ground (25m–30m OD) providing some protection from the harshest winter winds and the most violent sea storms. To the north and south, the land is lower, but sufficient to deflect the direct effect of onshore winds. The monastery is bounded on the southern side by the larger of the two freshwater lakes that would have provided an abundant, year-round supply of water. The presence of these lakes must have been a contributing factor to the choice of site for the monastery.

Pl. 4.1 The monastery located in a sheltered valley at the southwestern end of the island (G. Scally).

Pl. 4.2 Aerial view of the monastery, looking east (courtesy of Michael Herity). 1 SMR: GA021:026022; Townland: High Island; IGR: 50171 (E), 257348 (N); ITM 450154 (E), 757371 (N).

high island: excavation of an early medieval monastery

N. LANDING

4 5

8 8 9

S. LANDING

11 12

LIMIT OF VEGETATION

14 15

19 20 19 18

1. Ruined Structure 2. SE Landing Cross 3. Wall 4. 19th Century Mine Shaft 5. Murphy's Cabin and Well 6. Ruined Miner's Cottage 7. Blasting Powder Hole (19th Cent.) 8. Wall 9. Leacht / Penitential 10. Sub-Rectangular Structure 11. Brain Boru's Well and Cross 12. Circular Structure 13. Ruined Structure 14. Sub-Circular Structure 15. Northern Structure 16. Monastery 17. Leacht by Millpond 18. Enclosure 19. Wall 20. Cross Wall of Orthostats 21. Stone Wall? 22. Pre-Monastic 23. Circular Structure 24. Rectangular Structure and Enclosure 25. Wall 26. SW Landing Leacht 27. Mill Building

22 23 24

S.E. LANDING

100

150

200

250m

Fig. 4.1 Map of High Island showing all man-made structures and indicating the limit of vegetation growth (after White Marshall and Rourke 2000, Fig. 12, with amendments).

the monastery and its associated structures before excavation

PRE-MONASTIC FEATURE

RUBBLE

N CELL A 0

5.0m

ORTHOSTATS

EXTENT OF RUBBLE

W CH ALL AM BE

ENTRA NCE

ROCK

ROUGH MASONRY

CROSSINSCRIBED STONE

R LEACHT

CELL B

ALTAR CHURCH BURIALS

HIDDEN CROSS

RE ENCLOSU CHURCHWALL

CROSS-INSCRIBED STONE

LARGE WALL CHAMBER

GRANITE SPHERE

OSTA

CHAMBER

EXT. CASHEL WALL FACE

O NT

ORTH

LEACHT

ORTHOSTAT

STEPS

MAIN ENTRANCE

REVETMENT POSSIBLE GUEST HOUSE

ENTRANCE IR

REPA

MILLPOND

Fig. 4.2 Plan of the monastery on High Island before excavation (after White Marshall and Rourke 2000, Fig. 29a).

4.2 The monastery The focus of the monastery consists of a small church closely surrounded by an enclosure wall (Fig. 4.2). Outside the church enclosure wall, an upstanding beehive hut survives (Cell B), and also the collapsed remains of probably three other similar huts in the west of the monastic enclosure. All these structures are contained within the monastic enclosure wall. Four entrances lead through this wall, and another beehive hut (Cell A) is built within the thickness of its northern side. The main entrance is at the south-east outside where there are two buildings, one on

each side. Before excavation, the entire area within the monastic enclosure was obscured by masonry rubble and covered by long grass. White Marshall and Rourke describe in detail the monastery in the years before the excavations (White Marshall and Rourke 2000). The condition and appearance of the features within the monastery and those relevant to it outside the monastic enclosure are described below as they appeared in 1995, when excavations commenced, in order to provide a context for the discussion of the excavations. Dimensions and the relationships of the structures to one another described in this section are as they appeared before excavation.

42 high island: excavation of an early medieval monastery

Pl. 4.3 The church before excavations began in 1995, looking west (Photographic Unit, NMS).

4.3 The church and its enclosure The church The church is located within the northern half of the monastic enclosure (Fig. 4.2). The church, orientated east– west, was unroofed, however its side walls were upstanding to full height, thus making it one of the most intact buildings (Pl. 4.3). The north and south walls, especially their upper courses, were leaning a little towards the south, probably as a result of subsidence. The west wall was leaning inwards. The upper part of the east wall had collapsed and almost its entire length, with the exception of both ends, was leaning outwards at an angle of approximately 45 degrees. The church is entered by a doorway located in the west wall, slightly north of centre. The head of the doorway is formed by two lintels, with a packing of smaller stones between them. The inner lintel is a decorated crossslab (Cross-slab 16), showing signs of severe weathering. Inside the church, a stone altar abuts the east wall and a small aumbry is positioned in the north-eastern corner. The interior of the church comprised grass-covered soil and rubble.

The church enclosure Rubble up to 1.1m in depth covered the entire area between the church and its enclosure wall (Pl. 4.4). This rubble must have originated from the collapse of both the enclosure wall and the east wall of the church, and the graves described by Petrie in 1820 outside the east wall of the church were completely obscured. Prior to excavation, the drystone enclosure wall surrounding the church was barely discernible beneath mounds of rubble (Fig. 4.2; Pl. 4.5). The wall, enclosing an area of only 9m east–west by approximately 7m north–south, defines the most sacred space within the monastery. In places, the depth of rubble reached the top of the surviving wall, which made it difficult to recognise

Pl. 4.4 Rubble engulfing the church, the church enclosure wall and Cell B before excavation, looking west (Photographic Unit, NMS).

its outline (Pl. 4.4). It appeared, however, that the eastern side of the enclosure wall was the best preserved, standing to a maximum height of 1.4m. The southern and western sides both showed signs of inward collapse and survived to 1.1m and 0.6m in height, respectively. A low wall of rough masonry had been built as a buttress against the inner face of the western wall, to prevent further collapse. The northern side of the enclosure was the least well preserved and, though its outer face survived more or less intact, almost its entire inner face had collapsed inwards. Two entrances were visible in the church enclosure wall, one at the north-western corner and the other at the north-eastern corner (Fig. 4.2). The north-western entrance was free of rubble and a narrow path led towards the church doorway indicating that, at least in the recent past, this was the main point of entry. In contrast, the north-eastern entrance was filled with rubble and clearly had not been in use for a long time. Three lintels spanned the width of this entrance, suggesting that this was the covered passage leading northwards to the beehive hut, Cell A, which had been observed by Petrie and Kinahan (Fig. 4.2). In spite of the extensive collapse and resultant rubble, it was evident before excavation that the church was not centrally placed within its enclosure and that the walls of both structures were not on the same alignment (Fig. 4.2).

4.4 The monastic enclosure Cell B Outside the church enclosure, the most substantial structure to survive was the beehive hut, Cell B (Fig. 4.2). This cell had been built against the eastern side of the church enclosure wall. The lower half of the cell exterior and the ground surface in the immediate vicinity were covered by grass-covered rubble (Pl. 4.4). The visible upper part of the cell showed that the masonry of the wall faces had been

the monastery and its associated structures before excavation 43

Pl. 4.5 Aerial view of the monastery before excavation (Photographic Unit, NMS).

removed, leaving the rubble core exposed and partially eroded. Five projecting corbels were visible, spaced evenly around the cell walls. Small voids extending all the way through the walls were apparent. The apex of the roof had fallen in, creating an opening 0.6m wide. The cell is entered from the east via a doorway (W: 0.68m–1.58m) that splays widely outwards and is roofed with lintels, three of which had survived in place. The innermost lintel is a re-used cross-slab (Cross-slab 18). The floor in the entrance was covered with grass-covered rubble. The internal masonry of the cell was largely intact. All four walls, which were visible for 0.8m above the rubble, are corbelled, the excellence of the masonry resulting in smooth, gently curving faces on all sides. Masons’ tool marks are visible on the stones above the door lintel. Two small recesses, one in the north wall and the other in the west wall, are tentatively identified as wall cupboards. The amount of rubble within the cell was far in excess of what could have resulted from the collapse of the roof, indicating that some must have been brought in from elsewhere.

Other features within the monastic enclosure The monastic enclosure wall surrounds an area 32m east– west by 27m north–south. Within the enclosure, there were deposits of rubble that were extensive in the northern part in particular, where they extended almost to the height of the surrounding enclosure wall. Low, roughly circular mounds of rubble survive in the lee of the western wall (Fig. 5.1). These are probably the remains of beehive huts, possibly those remarked upon by Petrie. The number of huts located here is uncertain, but the outlines of rubble suggest three. Visitors to the island in the 1950s remember at least three upstanding huts in this location (O. and R. Fouéré, pers. comm.).

A cross-inscribed slab (Cross-slab 20) stands upright in the south of the monastic enclosure, where it had been placed by Kinahan (Fig. 4.2). The remains of a leacht are situated outside the western entrance to the church enclosure. Two courses of stone (H: 0.24m) were apparent on the southern and western sides, and slender upright stones abutted the south-eastern and north-western corners of the leacht. A large, round, granite boulder, known locally as the ‘granite globe’, lies on the grass in the south of the enclosure (Fig. 4.2). The stone has been shaped and its surface worked smooth so that it is almost a perfect sphere. Its function within the context of an early medieval monastery is uncertain. Pouchin Mould describes it as a ‘perfectly shaped ball of pale granite’ and suggests a connection with similar round stones, used as cursing stones, found at other sites, including Inishmurray, Co. Sligo, Ballyvourney, Co. Cork, and on Arran in Scotland (Pouchin Mould 1955, 53, 81). Such stones are also known from other island monasteries, such as the conglomerate boulder, known as Leac na Naomh, on Caher Island, Co. Mayo (Fig. 1.1). O’Donovan, during the course of his work for the Ordnance Survey, noted that the people of the mainland and of the neighbouring islands went to the stone on Caher Island because of its powers ‘to elicit the truth’ and to expose wrongdoers (cited in Herity 1995b, 101). Another cursing stone, called Leac Cholumb Cille, at Oughaval church (near Westport, Co. Mayo), has been linked to the death of a member of the O’Malley family, recorded in the annals, in the mid-12th century. Christiaan Corlett surmises that the use of the stone in its capacity to enact revenge can be traced back to at least this time (Corlett 2002). The High Island stone may have belonged to a similar local tradition, but no evidence for this has survived.

44 high island: excavation of an early medieval monastery

Pl. 4.6 Wall chamber (F721) before excavation, looking north (G. Scally).

The monastic enclosure wall The drystone wall, roughly ovoid in plan, measured 112m in circumference (Fig. 4.2). Much of it had collapsed and both the wall and the build-up of rubble on both sides were sealed for the most part by the grass-covered sod. In places, short stretches of the wall faces were visible (Pls 4.2, 4.5). Four entrances lead through the monastic enclosure wall, one each at the south-east, south-west, north-east and north-west ‘corners’ (Figs 4.2, 5.1). The main entrance (1.3m in width), at the south-east, was largely grass-covered. On the northern side of the entrance, the enclosure wall incorporates remnants of a small mural chamber. To the south of the entrance, the building, identified as a possible guesthouse, was built against the outer face of the enclosure wall. The entrance at the south-west of the monastic enclosure (L: 4m x W: 0.8m) is a narrow east–west passage through the wall (Fig. 5.1). This entrance permitted access to and from the southern part of the island beyond the lake and, importantly, to the monastic mill at the cliff-edge (Fig. 4.1). The entrance at the north-west was discernible as a shallow, depressed area (L: 4m and W: less than 1m). The entrance at the north-east, barely discernible at ground level, is shown by Kinahan on his plan of 1869 and it is clearly visible in an aerial photograph of the site (Fig. 3.5; Pl. 4.2). This entrance consists of a passage (L: 4m x W: 2.5m) between the end of the northern wall of the enclosure and a low (H: 0.5m), grass-covered wall on the southern side, which continues as the eastern stretch of the monastic enclosure wall (Fig. 5.1). The eastern wall of the monastic enclosure, located on higher ground, was the least substantial. The wall before excavation was visible as a low mound (H: 0.7m), traceable for 15m, and it was likely to have extended to the main south-east entrance of the enclosure (Fig. 4.2). Approximately mid-way along the eastern wall, the outline of a curved stone feature (L: 5m) was visible extending south-westwards inside the wall (Fig. 5.1). The southern

enclosure wall, which curves inwards to flank the lakeshore, was the best-preserved stretch. It extends 23.5m in length, varies 2.2–2.5m in width and survived to 1.2m in height. Examination of the visible parts suggested that this section of the wall was built exclusively of stone. The western enclosure wall (L: 34m, W: 3–4m and H: c. 1.5m) had largely collapsed and was discernible as a broad, grass-covered mound (Pl. 4.5). The faces of the wall were not visible and it was impossible to determine whether this stretch of the wall was built exclusively of stone or if it was a combination of stone and soil. The wall appeared to have collapsed inwards close to the northern end, at a point where the clochán had been recognised within the thickness of the wall by O’Donovan (Fig. 3.6, no. 3; Herity 1990a, 72). Two chambers, entered from inside the enclosure, lie within the thickness of the western wall (Fig. 4.2). The larger chamber, located close to the southern end, is orientated north–south, following the line of the enclosure wall. Prior to excavation, most of the lintelled roof of the chamber had collapsed and the chamber was full of rubble (Pl. 4.6). Only four lintels remained in situ at the northern end. The second mural chamber is situated 12m north of the larger chamber. This part of the wall had collapsed and the chamber was difficult to locate amongst the rubble. The chamber is orientated roughly east–west, so that it lies at right-angles to the wall faces, and it extends across most of the width of the enclosure wall (Fig. 4.2). The outline of the chamber was discernible as a small stretch of stone-facing and orthostats enclosing a narrow space (L: 3–4m x W: less than 0.9m). The northern enclosure wall (L: 16m) is the shortest of the four and also was the least well preserved. It appeared that the central section had been robbed out so that it was nearly level with the surrounding ground, as described by O’Donovan in 1839. Close to the entrance at its eastern end, a small beehive hut, Cell A, was visible within the thickness of the wall (Figs 4.2, 5.1; Pl. 4.7). The easternmost section of the northern enclosure wall was quite well preserved and was defined by a grass-covered mound (L: 6m, W: 2m and H: 0.8m), located on upward sloping ground. At the western end, it was identifiable as a grass-covered mound (L: 4m, W: 2.1–2.4m and H: 1m) traversed by the north-west entrance, which, prior to excavation, was discernible only as a slightly depressed area (Fig. 5.1). White Marshall and Rourke suggest that this was a pre-monastic feature (Fig. 4.2), however Herity correctly describes it as an entrance passage (Herity 1990a, 74).

Cell A The small beehive hut, Cell A, was built within the northern monastic enclosure wall (Fig. 4.2). The corbelled roof and the upper courses of the walls had collapsed inwards (Pl. 4.7). The cell is surrounded by a low wall abutting the

the monastery and its associated structures before excavation 45

Pl. 4.7 Cell A (in foreground) before excavation, looking south-west (Photographic Unit, NMS).

exterior, known as an annulus, which obscured all but the upper one or two courses of the walls. The visible parts of the external masonry appeared rough and uneven, and on the western side the masonry seemed to have slumped out of position. The cell is entered from inside the enclosure via a south doorway, which splays outwards. Originally, the entrance would have been roofed by three, or possibly four lintels, two of which had survived in position. The inner lintel had cracked and was precariously supporting the overlying masonry. Before excavation, the internal faces of the walls were visible to a height of c. 1.2m above the collapsed masonry. The walls are built with mica-schist interspersed with granite, both stone types that occur on the island. The stones are large with minimal evidence for having been worked, just enough to create the corbelled dome shape. In contrast to Cell B, the masonry appears rough and uneven, suggesting that Cell A was possibly built in haste and without great care or craftsmanship.

4.5 Structures outside the monastic enclosure Structures clustered together outside the main south-eastern entrance to the monastic enclosure comprise a large building on the southern side and a small chamber and a leacht on the northern side (Fig. 4.2). A stone path led

from the main entrance along the northern edge of the lake. Beyond the lake, a low boundary wall extends in a south-easterly direction to meet other structures close to the cliff-edge (Figs 4.1, 4.3; Pl. 4.8). To the north of the monastic enclosure, another boundary wall continues as far as structures located at the cliff-edge. These walls and structures, with the lake in between, effectively separated the monastery and the south-western end of the island (6ha/15 acres) from the remainder. The mill is located within this area and Brian Boru’s well, in the centre of the island, was also associated with the monastery (Fig. 4.1, nos. 11, 27).

The structures at the main entrance The structures at the main entrance were obscured by grass-covered rubble at the time of excavation and much of the descriptions below are based on earlier accounts (Petrie 1845, 420–41; Herity 1990a, 96; White Marshall and Rourke 2000, 49–50). Petrie suggests that the building at the southern side was used by pilgrims, while Herity and also White Marshall and Rourke postulate that, because of its size and position, it was probably a guesthouse. The structure (L: 5m x W: 4m) has an entrance in the east wall, and its west wall appears to abut the monastic enclosure wall (Fig. 4.2). The north wall (H: 0.7m) forms one side of the entrance passage. The south wall (H: 1.6m) is the best preserved of the four. All four walls appear to have been too narrow (W: 0.6–0.7m) to support a stone roof.

46 high island: excavation of an early medieval monastery

Pl. 4.8 Aerial view of the south-west end of High Island showing the monastery adjacent to the upper lake (on left) and low wall extending north–south (between the two lakes) across the entire width of the island (Photographic Unit, NMS).

This observation had been noted by previous writers, who suggest that the roof was of timber (Petrie 1845, 420–21; Herity 1990a, 73; White Marshall and Rourke 2000, 50). The location of this building ensured that visitors entering the monastic enclosure must pass through the south-east entrance and it also ensured that visitors staying in the guesthouse were accommodated outside the sacred area. The lake edge appears to have been artificially extended to the southern edge of this structure (Fig. 4.3; Pl. 4.5). The small mural chamber on the north side of the main entrance was almost totally covered over at the time of excavation, with just a triangular space visible (Fig. 4.2). It was not possible to determine its dimensions with certainty, but Herity suggests that the northern side measures 2.2m in length and he observed that the entrance faces south (Herity 1990a, 73). White Marshall and Rourke (2000, 49) noted that there are three steps at the entrance and propose that the chamber functioned as a gate-keeper’s or porter’s lodge. The presence of a guesthouse and a porter’s lodge indicate that the monastery, despite its relative isolation and inaccessibility, did entertain a sufficiently large population of pilgrims and other visitors to warrant investment in the erection of purpose-built structures. The existence of both a guesthouse and a porter’s lodge is not known at other small Atlantic island monasteries on the west coast of Ireland.

The mural chamber on the north side of the entrance was abutted on its eastern side by a leacht (Fig. 4.2). Herity described it as a square structure (1.2m2 and H: 0.45m), which had blocks of quartz on its surface (Herity 1990a, 73). White Marshall and Rourke also interpreted this structure as a leacht, describing it as a small, rectangular structure (1.14m x 1.06m and H: approx. 0.55m) with white quartz stones lying on top (White Marshall and Rourke 2000, 50).

Boundary walls and associated structures From the main entrance to the monastery, a path had run along the northern shore of the lake and was exposed for c. 200m when Petrie described it in 1820. A significant number of large paving stones can still be seen in the section face at the lakeshore and others, especially at the southern end, are partially visible through the grass-covered sod (Fig. 4.3). A boundary wall runs southwards from the lake to a curvilinear enclosure (approximately 17m north–south x 22m east–west) and a rectangular-shaped ruined building (c. 6.1m x 2.55m) with an entrance at the north, which are both located at the cliff-edge (Fig. 4.3; Pl. 4.8). The boundary wall stands two to three courses high (0.6–0.7m) and it extends into the lake for c. 2m. The boundary wall on the northern side of the monastery runs from the

LEACHT

ENCLOSURE

RECTANGULAR ENCLOSURE STRUCTURE

SMALL POND

SMALL POND (20.6m)

25m

CIRCULAR STRUCTURE

26.4m

WALLI

26.4m

25.3m

RIVULET

31.9m

WAY

PATH

MILLPOND (23.6m)

26.7m

ENCLOSURE

PATHWAY

MILLPOND

36.1m 1m

LEACHT

MONASTERY

REMAINS OF STONE STRUCTURE?

WALL

LARGE ERRATIC

SUBCIRCULAR STRUCTURE

NORTHERN STRUCTURES

COVE

Fig. 4.3 Western end of High Island: (a) plan showing the structures in the valley between the south-eastern landing and the monastery and the structures north of the monastery, contours at 1m intervals; (b) north–south section (after White Marshall and Rourke 2000, Fig. 22a, b with amendments).

(b)

(a)

COVE

the monastery and its associated structures before excavation 47

48 high island: excavation of an early medieval monastery

Fig. 4.4 Conjectural reconstruction of the mill building on High Island (after Colin Rynne in White Marshall and Rourke 2000, Fig. 146).

Fig. 4.5 A schematic, three-dimensional view of the mill system in the terrain on High Island (courtesy of Colin Rynne).

Pl. 4.9 The feeder stream channel, part of the mill complex still visible on the ground (G. Scally).

Pl. 4.10 Remains of the mill structure above Cuan a ’Mhuilin (G. Scally).

north-eastern entrance as far as the remains of structures situated at the cliff-edge (Fig. 4.3). These remains suggest a circular building (D: c. 6m) and include a possible short stretch of wall. At these structures, the wall turned sharply back to the south-west towards the monastery to enclose a triangular-shaped area of relatively level ground, possibly a garden, which is visible on an aerial photograph (Fig. 3.6; Pl. 4.2).

Rynne, pers. comm.). He demonstrates that the mill was a horizontal-wheeled watermill (Fig. 4.4). The water supply appears to have consisted of three main elements: first, an upper reservoir or feeder pond located on the high ground east of the upper lake, where water would have been held behind a gravity dam of earth or stone; secondly, a feeder stream that led from this to the upper lake or millpond beside the monastery; and finally, the upper lake or millpond itself (Fig. 4.5). Remains of the gravity dam and part of the feeder stream are still identifiable on the ground today (Figs 4.5, 4.6; Pl. 4.9). The lake or millpond, as Rynne refers to it, is a ‘natural lake, but one that has been modified and probably extended to serve as a storage reservoir for the operation of a watermill’ (ibid., 191). From the millpond, the water flowed out the headrace channel, which had been cut into the exposed bedrock on the western side of the lake (Fig. 4.5). Approximately 14m from the millpond, the headrace channel, running in a circuitous fashion, split into two. Surplus water was taken away in

The watermill The early medieval mill on High Island, mentioned by both O’Flaherty and Petrie, is situated beside the cliffedge 100m south of the monastery on the far side of the larger lake (Fig. 4.1). It has been the subject of an extensive study by Colin Rynne, who states that it is the earliest known example of an Irish monastic mill and that it is also the earliest survival of its type in Europe (Rynne 2000, 186). All the other recorded mills have been revealed by excavation and do not have upstanding remains (C.

MILL

25m

26.4m

26.7m

MILLPOND

26.3m

MILLPOND (23.6m)

LEACHT

ENCLOSURE

WALL

PATH

ENCLOSURE

PATH ALONG POND

31.9m

36.1m

RIVULET

WALLING

WALL?

ORTHOSTATS

WALLING

Fig. 4.6 (a) A 1m interval contour plan of the horizontal mill landscape on High Island; (b) west–east cross-section (after White Marshall and Rourke 2000, Fig. 142a–b).

(b)

POND No.2

(a)

COVE

24.6m

24.1m

MONASTERY

the monastery and its associated structures before excavation 49

50 high island: excavation of an early medieval monastery

Fig. 4.7 Brian Boru’s well (Patricia Johnson).

the bypass channel (L: 13m) to exit over the cliff-face. The other channel continued as the headrace for a further 30m as far as the mill building. The remains of the mill building consist of a single wall originally belonging to a rectilinear structure that is c. 8m from the cliff, overlooking a deep cove named ‘Cuan a ’Mhuilinn’ (the inlet or cove of the mill) (Pl. 4.10). Rynne estimates that the mill building (6.5m x 3m) would have had its long axis north–south, with an off-centre opening (W: c. 1.6m) at the west to accommodate the waterwheel assembly (ibid., 192). He also suggests that the mill differs from the more typical early medieval Irish horizontal-

wheeled mill in that it straddled its feeder channel and was provided with a bypass channel. Although other mills of this period are known to have used large structural timbers for the foundations and the superstructure, stone was used on High Island. Notwithstanding this, a large timber balk would have been required for the mill flume (Fig. 4.4; Rynne 2000, 193).

Brian Boru’s well An historical association linking High Island and Brian Boru, king of Munster and high king of Ireland, is not known (White Marshall and Rourke 2000, 30). The con-

Fig. 4.7

Pl. 4.11 Panoramic views of the mainland from Brian Boru’s well (G. Scally).

the monastery and its associated structures before excavation 51

nection goes at least as far back as the 17th century as O’Flaherty refers to Brian Boru’s well. In the 19th century it was considered a holy well and Kinahan reports that ‘the holy well is said to cure colic and all such complaints’ (Kinahan 1868–9, 554). Although it is not marked as a penitential station on the OS map of 1841, it has been suggested that the well was part of the early medieval turas and would therefore have been a station of the pilgrimage round (Fig. 3.7; Herity 1990a, 85). The site of the well, in the centre of the island, is spectacular on account of its elevation and the position it commands (Fig. 4.1; Pl. 4.11). Its elevated location (60m above sea level), set between the two highest ridges of the island, provides panoramic views of the mainland, encompassing the entire Connemara coast from Slyne Head to the south as far as Slievemore on Achill Island to the north, taking in the Twelve Bens and Croagh Patrick. The

well nestles at the base of a rock outcrop, thought to have been deliberately cut back (White Marshall and Rourke 2000, 31). The opening of the well is roughly horseshoeshaped (W: 0.65–0.85m) and the base is cut into bedrock (Fig. 4.7; Pl. 4.11). The upper part of the well was rebuilt in the early 1970s by Richard Murphy, who added the steps that lead down into the well. When digging a trench on the south-western side to allow excess water to flow down-slope away from the well, he uncovered a crossslab, which may originally have marked the location of the well (Cross-slab 38; R. Murphy, pers. comm.). In 1839, Wakeman drew another cross-slab, which was then at the church but is now beside the well (Figs 4.7, 7.12; Pl. 4.11; Cross-slab 21). It must have been relocated before 1868 because the reverse side of the same cross-slab was drawn by Kinahan after it had been moved to the well (Kinahan 1868–9, 554–5).

5 the excavations 5.1 Introduction

vation relating to the conservation works were undertaken between 2003 and 2009, amounting to a further three monthsâ&#x20AC;&#x2122; work. The excavations were carried out under Licence 95E124 and, since 2003, under Ministerial Con-

Eight seasons of excavations were carried out, one each summer between 1995 and 2002, amounting to a combined period of nine months. Shorter periods of excaN

NW entrance

AREA 4 Monastic Enclosure

5m Cell A AREA 5

NE entrance

AREA 1

Monastic Enclosure

AREA 8 Leacht Small wall chamber

Rubble

Altar

AREA 2

Cell B

AREA 3

Church

AREA 1 Church Enclosure

AREA 1

AREA 7

Wall chamber

AREA 6

'Granite globe' Chamber

Main SE entrance Possible guesthouse

SW entrance

Lake Structures Limit of excavation

Fig. 5.1 Plan of the monastery showing location of excavation areas 1â&#x20AC;&#x201C;8 (after White Marshall and Rourke 2000, Fig. 29a with amendments).

the excavations 53

sent C092. The findings of each season are recorded in detailed stratigraphic reports, which are lodged in the Archive Section of the National Monuments Service (NMS) and in the National Museum of Ireland (NMI) (Scally 1995–2002). A substantial photographic record of the excavations is retained in the Photographic Unit of the NMS; black-and-white and colour prints and colour transparencies form the bulk of the collection. Digital photography commenced in 2007. Excavation was carried out in eight areas within the monastery (Fig. 5.1). The location of seven areas was determined by the conservation programme, which stipulated that only structures capable of being stabilised and conserved should be excavated. The depth of excavation in these areas was limited to the uppermost surface of paving, or other suitable surfaces, upon which conservation could take place. For this reason, in most areas archaeological deposits remain in situ at lower levels. Only one excavation area, located in the south-western part of the site between the church enclosure and the monastic enclosure, was opened for research purposes (Fig. 5.1, Area 6). However, excavation was terminated at the end of the first season as it became apparent that a larger, more open area (as opposed to a narrow trench) would require investigation if the features uncovered were to be interpreted in any meaningful way. The exposed features were recorded and the area was backfilled. All features have the prefix F and were allocated numbers according to the excavation area number, e.g. in Area 1 numbers begin at F1 and continue to F198; in Area 2 they begin at F200 and continue to F298, etc. Where additional feature numbers were required, for example in Area 1 above F198, the numbering sequence recommences with F1000 and the same sequence was applied, where required, in Areas 2–8. Plans, sections and elevations were drawn on-site at a scale of 1:20, the only exceptions being the elevations within Cell B, which were drawn at a scale of 1:10. Levels shown relate to an arbitrary temporary benchmark (TBM) of 10.00m, and all drawings are shown relative to this height. The finds are registered in accordance with NMI specifications, where the license number of the excavation is followed by the feature number and then by the individual find number, for example, 95E124:10:1, 95E124:209:1. Where finds are referred to in the text, the license number is omitted, for example, 10:1, 209:1; all dimensions quoted in the finds catalogue refer to maximum surviving size. Stray finds were also allocated a unique number, depending on where the find was recovered. Those from Area 1 (or from rubble/soil removed from Area 1) are prefixed by F199; those from Area 2 by F299; from Area 3 by F399; and so on to Area 7. The one exception is Area 8, where stray finds were prefixed by F8099. Cross-slabs

recovered from the excavation and those known prior to excavation have been allocated a unique cross-slab number from 1 to 66. Soil samples were registered by the year collected, followed by a letter of the alphabet, e.g. ’95A, ’97A, ’01A, etc. When more than 26 samples were collected in one season, the sample was prefixed with a duplicate letter, for example ’95AA, ’97AB, ’01AC. In 2002, samples with a duplicate letter followed a slightly different sequence, for example ’02AA, ’02BB, ’02CC. The results of the excavations are presented in this section, commencing with the church enclosure in the centre of the monastery and then moving outwards to Cell B and the monastic enclosure wall, finishing with a description of the subrectangular building (Fig. 5.1). The chronology of the site is divided into four phases based on the broad date ranges provided by radiocarbon analysis. This shows the development of the monastery from its probable foundation sometime in the 8th to mid-11th century (Phase 1), through its most intensive period of occupation from the mid-11th century to its abandonment in the late 12th or early 13th century (Phase 2), to a period of reoccupation between the late 12th or early 13th century and the mid-15th century (Phase 3) and finally to its neglect and destruction from the mid-15th century onwards (Phase 4).

Conventions and terminology All calibrated radiocarbon dates are referred to at 2 sigma (95.4% probability). The following chronological outline of the medieval period is used in this report for discussion purposes: early medieval: 5th–12th century; high medieval: 12th century–c. 1400; late medieval: c. 1400–16th century.

5.2 The church enclosure This section describes the excavation of the church enclosure. It begins with an account of the features uncovered beneath the church and of the church itself (Fig. 5.1, Area 2). This is followed by a description of the paved surface in the enclosure, the graves outside the east wall of the church and the enclosure wall (Fig. 5.1, Area 1). The section also describes related features outside the enclosure, including a wall (F91) and Grave 11 outside the north-east entrance (Fig. 5.1, Areas 4 and 8).

5.2.1 The church Phase 1: 8th to mid-11th century The pre-church features The earliest ecclesiastical evidence was located beneath

54 high island: excavation of an early medieval monastery

1 Paved area (F245) 2 Charred surface (F232) RD

3 Burnt debris (F232a) 4 Burnt debris (F246) 5 Burnt debris (F267)

A 1

6 Clay (F205)

7 Post-hole (F250)

Upright stones

The later church 6

Rabbit disturbance (F243) Conjectural

Later feature Obscured Limit of excavation

Stone 0

Fig. 5.2 Pre-church features associated with the paved area (F245). For sections B–B1 and C–C1, see Fig. 5.3 (a) and (b).

the upstanding stone church (Fig. 5.1, Area 2; Fig. 5.2). The surface of the undisturbed boulder clay, which was cream in colour, was stiff and cohesive and full of stone and was uncovered at the lowest excavated level (Fig. 5.3 (a), (b), no. 12, F271). A thin lens of fine mid-brown soil with inclusions of mica-schist and substantial pockets of almost soil-free, fragmented schist covered the boulder clay (Fig. 5.3 (b), nos. 8, 9, F247, F248). The overlying layer (depth: 10–14mm) was an organically rich soil that contained particles of charred vegetation identified as fragments of heather, juniper, oak and birch (Fig. 5.2, no. 2, F232). The remains of domesticated animals consisted predominantly of sheep/goat and also cattle bones (Section 9.1, F232). There was also a large quantity of bird bones, almost half of which were woodcock, indicating that this wild food source was exploited. The same charred material

was revealed beneath the centre section of the east wall of the church, which was dismantled (Fig. 5.3 (a), (b), no. 7, F232). This deposit was also discovered further to the east, around the graves outside the church, indicating that this scorched ground surface may have been extensive (Section 5.2.3). Three areas of burning (each less than 25mm in depth), which were possibly hearths, lay within the charred vegetation (Fig. 5.2, nos. 3–5, F232a, F246, F267). These deposits were composed of oxidised soil, orange ash and organic material and fine grey ash, which when analysed were found to be peat ash (Section 9.2, F232, F267). Embedded into the soil and fragmented schist (F247, F248), the remains of a small area of paving (1.2m x 0.75m) were exposed, comprising five flat slabs of schist (Fig. 5.2, no. 1, F245; Pl. 5.1). A socket where a sixth slab had been removed was also evident. The slabs were thick (one meas-

the excavations

East B

West C

West B1

East C1

Top of Grave 3

2 3

F67

3 4

Grave 3 12

50 cm

(b)

7 8

8 (a)

1 East wall of church

10 Kaolinite clay ball (F270)

2 Stones (F256)

11 Prop stone behind headstone (CS 5), Grave 3

3 Clay and schist (F216/F262)

12 Boulder clay (F271)

4 Burnt debris (F221)

Paving

5 Silver/grey clay (F224)

Unclear

6 Grey clay (F205)

Obscured

7 Charred surface (F232)

Limit of excavation

8 Soil and schist (F247)

RD Rabbit disturbance

9 Fragmented schist (F248)

CS Cross-slab

Fig. 5.3 Sections through the east wall of the church and underlying deposits: (a) south face; (b) north face. Location of sections B–B1, C–C1 on Fig. 5.2.

Pl. 5.1 The paved area (F245) (under ranging rod) and posthole (F250) in left foreground, looking south. Paving (F29) contemporary with the church visible on upper right (G. Scally).

ured in excess of 0.18m) and they were secured in place by small stones placed in the joints. Other slabs lying loose to the north may also have formed part of the same paved floor (Fig 5.2; Pl. 5.1). A post-hole (D: 0.12–0.14m; depth: 0.4m) lay adjacent to the east side of the paved area (Fig. 5.2, no. 7, F250). It was partially stone-lined and filled with charred vegetation, which produced a similar range of species as that identified from the surrounding charred

surface (Section 9.3, F232, F250). Patches of compact, light-grey clay (depth: 70mm) surrounded the paving slabs and more extensive areas of the same type of clay survived, especially towards the east and north-east where it extended beneath the walls of the stone church (Fig. 5.2, no. 6; Fig. 5.5, no. 4, F205). This redeposited clay may have been used as a rudimentary mortar (Section 9.4, F205, ’00AK; Appendix B). Smaller patches of the same type of clay were also found outside the east wall of the church at Graves 6 and 7 (Section 5.2.3, F65). Silver/grey clay overlay the paved surface and surrounding area (Fig. 5.3, no. 5; Fig. 5.5 (b), (c), no. 3, F224). In places, lenses of fine, possibly wind-blown brown soil covered the silver/grey clay in which a white quartz pebble was found. Within these deposits, patches of moist, humic soil were mixed with lenses of charcoal. These deposits were sporadic due to the fact that a significant part had been removed by later activity (Grave 9) and they were also present beneath the church walls. Soils of similar type, which may represent a continuation of the same deposits, were identified at the same level in Grave 6, outside the east wall of the church (Section 5.2.3, F55, F56). An extensive layer, orange and mottled orange streaked with black, was found and it continued on in all direc-

56 high island: excavation of an early medieval monastery

1 1

Burnt debris (F221) The later church

Later feature 1

Obscured Limit of excavation

1 RD

Rabbit disturbance

1 RD

Fig. 5.4 The pre-church deposit of burnt debris (F221).

tions under the walls of the church (Fig. 5.3 (b), no. 4; Fig. 5.4, no. 1; Fig. 5.5, no. 2, F221). It is estimated that these deposits had accumulated over a period of between 40 and 200 years and they probably represent successive dumping of debris from domestic hearths (Sections 9.3, 9.4, F221, ’97O1–3). The residues contained wood ash from the fuel, identified as birch, hazel, ash, oak and juniper. Cattle, sheep/goat and pig bones were also present, testifying to a varied diet in which fish also played an important part, particularly scad. Radiocarbon analysis of a sheep bone produced a date range of cal AD 770–980 (OxA8946, 1182±36 BP). A second determination obtained from charcoal produced a date range of cal AD 860–1020 (OxA-8918, 1110±40 BP). Two stone-lined slots, with their long axes running north–south, were cut parallel to one another into the burnt debris (Fig. 5.6, nos. 2, 3, F260, F263; Pl. 5.2). The more easterly slot (F263) was 0.1m in depth and had vertical sides and a flat bottom. Small, flat stones set on edge lined the southern end and parts of both sides. Some angular-shaped stones lay along the upper edge on the eastern side and may represent stones left in place after the stone for which the slot had been created was removed. The slot was filled with friable brown soil mixed with fragments of mica-schist and flecks of ash. The inner face of the east

wall of the church overlay the eastern side of the backfilled slot. The second slot, 0.6m to the west (F260), was slightly larger and deeper (0.24m). This slot was also partially stone-lined, had rounded ends and its east side had been cut close to vertical and the west side sloped at an angle of 45°. Two different fills were identified, predominantly schist clay and, at the west face, a fine brown soil with inclusions of schist fragments and flecks of charcoal, which included a grinder (Section 6.2, 265:1). To the north of the slots and cut into the same burnt debris, three stones, 0.22–0.26m in height, had been placed upright and secured in position by carefully packed small stones around their bases (Fig. 5.6, no. 1, F269; Pl. 5.2). The spatial relationship between the upright stones and the slots (which together covered an area c. 1m2), coupled with the fact that the base of each of them was at the same level, suggested that they may all be the same date. To the south of the slots, no features were evident at this stratigraphic level, but disturbance by rabbits in this area was significant. A layer of fine, friable, grey clay (depth: 20–40mm), containing a large quantity of schist fragments, sealed the burnt debris and was possibly contemporaneous with the slots and upright stones (Fig. 5.5 (b), (c), no. 1, F216, F262). This layer underlay the south-east corner of the church and occurred sporadically in the vicinity of the north, south

(c)

2 5 6

West D4

South D4

Limit of first phase of masonry

Limit of second phase of masonry

10 Aumbry

Limit of first phase of masonry

East D1

(d)

(b)

North D1

6 5

4 8

CS 16

2 6

Fig. 5.5â&#x20AC;&#x201A; Internal elevations of church walls and sections through underlying deposits: (a) north; (b) east; (c) south; (d) west. Location of elevations on Fig. 5.7.

East D3

(a)

West D

North D

South D3

Cross-slab

Limit of excavation

Tie stone

Paving (F29)

Plaster/mortar

Lintel

10 Aumbry (F73)

9 Socket (F263)

8 Boulder clay (F271)

7 Fragmented schist (F248)

6 Soil and schist (F247)

5 Charred surface (F232)

4 Grey clay (F205)

3 Silver/grey clay (F224)

2 Burnt debris (F221)

1 Clay and schist (F216/F262)

the excavations â&#x20AC;&#x192; 57

58 high island: excavation of an early medieval monastery

1 Stones (F269) 2 Slot (F260) 3 Slot (F263)

4 Clay (F216)

5 Clay & schist (F216/F262) 6 Kaolinite clay ball (F270) 3

The later church

Upright stones Later feature

Limit of excavation RD

Obscured

Rabbit disturbance

Fig. 5.6 Pre-church features associated with the slotted features (F260, F263, F269a–c). Pl. 5.2 The slotted features (F263, F260) and upright stones (F269a–c) in foreground, looking south. Headstone (Cross-slab 8) within the later church wall visible on left (G. Scally).

and west walls. When the central section of the east wall was dismantled, the same layer was uncovered beneath where it had become mixed in with the underlying burnt debris of the hearth residues (Fig. 5.3, no. 3, F216, F262). Charcoal from the hearth residues in these deposits con-

tained oak and willow/poplar (Section 9.3, F216, F262). A ball of putty-like kaolinite clay was recovered within this mixed deposit (Fig. 5.3 (a), no. 10, F270). Mineralogical analysis of the kaolinite indicated that it is unlikely that it originated on High Island and that the clay was imported

the excavations 59

for use as a raw material, possibly for ceramic manufacture (Section 9.4, F270, ’00Q; Appendix B).

Phase 2: mid-11th to late 12th/early 13th century The church The unicameral, or single-cell, church, orientated east–west, is one of the best preserved structures at the monastery (Pl. 5.3). It is one of the smallest of the Atlantic island churches, almost square in plan (3.43m x 3.14m), resulting in a ratio of 1:1.1 (Fig. 5.7). Three distinct building phases can be identified within the masonry of the church, which in part are based upon earlier observations (Herity 1990a, 75–7; White Marshall and Rourke 2000, 73–6). Most of the masonry belongs to the earliest building phase, dating to the mid- to late 11th century. The church walls had no distinct foundations. The bases of the walls, exposed in the interior of the church, were built on the burnt debris (F221), with the exception of parts of the south wall where this layer was absent and the stones rested on the silver/grey clay or the clay and schist deposits (Fig. 5.5, nos. 1–3, F221, F224, F216/F262). The central section of the east wall, which was dismantled during conservation, enabled investigation of the deposits beneath. A layer of small, predominantly flat stones (F256) was revealed underneath, lying on the clay and schist (Fig. 5.3, nos. 2, 3, F216/F262, F256). A broken granite pounder, or grinding stone, which showed evidence for considerable use, was found amongst these stones (Section 6.2, 256:1). It was difficult to establish whether these stones formed part of the actual wall core, or if they were a bedding for the wall. The stones abutted the back of a headstone that formed part of the exterior façade of the wall (Sections 5.2.3, 7.1, Cross-slab 8). The headstone had subsided from its original position and may also have pushed the stones out of place (Fig. 5.3 (a), no. 2, F256; Pl. 5.2). Pl. 5.3 The church before excavation, looking southwest (G. Scally).

The first masonry phase of the church can be clearly distinguished on the inner faces of the west, north and associated with the restructuring of the graves beneath (Section 5.2.3). Excavation of the central section of the east wall revealed that it was mortar-bonded on both inner and outer faces, with a core of small stones, desiccated clay and lime mortar (Fig. 5.3 (a), (b)). A sample of mortar taken from the core determined that schist, mica, gneiss and quartz, all stone types available on the island, were crushed and that the lime binder contained mussel shell (Section 9.4, F10,’96 2; Appendix A). Patches of plaster had survived on the lower part of the east wall, behind the altar.

The aumbry A small aumbry, or cupboard (W: 0.3m x H: 0.17m), for storing sacred vessels was located at the northern end of the east wall (Figs 5.5 (b), 5.7, nos. 9, 10, F73; Pl. 5.5). Its roof was formed by three lintels and its base by a slender sillstone (Fig. 5.10, (a), (b); Pls 5.6, 5.7). The sillstone divided the upper, visible part of the aumbry from a concealed cavity (W: 0.35m x H: 0.55m) below, which was located behind the orthostat in the north-eastern corner (Figs 5.5 (a), 5.10 (a)–(c); Pl. 5.8). At sill level, the plan of the aumbry is roughly trapezoidal (L: 0.4–0.66m). This shape had been altered by subsidence of the east wall of the church, which was apparent in the upper parts of the east and south side walls of the aumbry, both of which had pulled away from the vertical. The north, south and east walls of the aumbry were formed by a single large stone at the base, with smaller-sized coursed masonry above (Fig. 5.10 (c); Pl. 5.8). The west wall was formed by the back of the orthostat in the north-eastern corner of the church. Despite the subsidence that had occurred, the north and west side walls of the aumbry had remained intact, and it was apparent that the north wall had a slight corbel (Fig.

60 high island: excavation of an early medieval monastery

Excavated area of wall

(CS 17) 10 D6

D5 D14

D13

9 10

D10

Altar (F209)

(F20)

(F60)

5 4

D11

D12

(F64)

(F29)

1 – 8 Graves outside east wall of church

Lintels

9 Aumbry (F73)

Upright stones

10 Slight 'edge' within paving (F29)

Obscured

11 Post-hole (F231)

Conjectural

12 Wor n area in paving – possibly from door upright

Limit of excavation

The church

Cross-slab

Fig. 5.7 The church and Graves 1–8 surrounded by the paved surface (F29) and church enclosure wall. Elevations D–D4, see Fig. 5.5 (a)–(d); D5–D8, see Fig. 5.8 (a)–(d); D9–D12, see Fig. 5.11 (a)–(c); D13–D14, see Fig. 5.10 (b).

Paving Lintel

3 Cross-slab 14

4 Covering slabs (F54, F30)

Bedrock

2 Aumbry (F73)

5 Cross-slab 16

Grave sidestones

1 Rubble

(d)

North D6

LIMIT OF SECOND PHASE OF MASONRY

(b)

South D7

LIMIT OF FIRST PHASE OF MASONRY

Limit of excavation

Conjectural

Obscured

Limit of first phase of masonry

East D7

West D6

Fig. 5.8 (a)–(d) External elevations of church walls: (a) north; (b) east – see detail Fig. 5.8 (e) (p.62); (c) south; (d) west. (Fig. 5.8 (d) excluding lower courses and paving, after White Marshall and Rourke 2000, Fig. 46a.)

(c)

West D8

(a)

Limit of first phase of masonry

East D5

South D8

North D5

the excavations

high island: excavation of an early medieval monastery

South D7

North D5

I D

F J

M 2

5 (e)

1 Collapse and rubble 2 Orange ash (F42a) 3 Black silt (F42b) 4 Clay and schist (F42) 5 Organic clay (F43a) 6 Black silt (F43) 7 Grey clay and schist (F50) 8 Organic soil and charcoal (F55/F56) 9 Brown mica-schist soil (F65) 10 Iron-flecked clay (F76) 11 Stone slabs (F54)

13 Recumbent slab (CS 14) over Grave 8 14 Boulder clay (F171) Paving Lintel Clay (F205) Grave sidestones Bedrock Conjectural Obscured Limit of excavation CS Cross-slab

12 Stone slabs (F30)

A Top of Grave 1

M Top of Grave 5

B Skull of burial (F41) in Grave 1

N Skull of burial (F37), Grave 5

C Recess (F51) for headstone (CS 2)

O Recess (F63) above Grave 6

D Top of Grave 2

P Top of recumbent slab (F25), Grave 6

E Headstone (CS 5), Grave 3

Q Aumbry (F73) visible through rubble

F Recumbent slab (CS 6), Grave 3

R Stones supporting upper sidestone between Graves 3 and 4

G Skull of burial (F70), Grave 3 H Headstone (CS 8), Grave 4 I Top of Grave 4 J Blocking stone (F62) in Grave 4 K Skull of burial (F40), Grave 4 L Headstone (CS 9), Grave 5

Fig. 5.8 (e) Detailed elevation of east wall of church.

Fig. 5.9 Sample illustration of masonry of the internal church walls (Patricia Johnson).

the excavations 63

Pl. 5.4 East wall of the church (in foreground) before excavation (G. Scally).

Pl. 5.5 The church interior showing the altar (F209) and the aumbry (G. Scally).

5.10 (d); Pl. 5.8). A noteworthy and unusual feature of the aumbry was that most of the northern side is set back behind the line of the north wall of the church (Figs 5.7, 5.10 (b); Pls 5.7, 5.8). The trabeate doorway in the west wall of the church was slightly off-centre to the north (Figs 5.7, 5.8 (d); Pl. 5.4). Almost the complete doorway (H: 1.5m) belongs to the first phase of construction (Fig. 5.5 (d)). The jambs were inclined so that the entrance was wider at the base (0.9m) than at the level of the top of the first masonry phase (0.8m), and they had a slight outwards splay of c.

Pl. 5.6 Lintels forming the roof of the aumbry (F73) (G. Scally).

64 high island: excavation of an early medieval monastery

Recess (F63) above Grave 6

Recess (F63) above Grave 6 D14

D14

Upper outline of aumbry

Base outline of aumbry A

A D13

D13 Altar

North wall of church

Altar

North wall of church

(a)

(b)

D14

D13 A

(d)

Stone concealed 60mm behind north wall of church

Stone continuing

Lintel

Line of wall unclear

Sillstone

Conjectural

Orthostat

Obscured

Church wall in section

Limit of excavation

Fig. 5.10 The aumbry (F73) in the north-east corner of the church: (a) plan with lintels in situ; (b) plan showing upper and base outline; (c) elevation of interior north wall; (d) vertical profile showing corbel in interior north wall. Location of elevations on Fig. 5.7.

60mm. Both internally and externally, the corner stones of the jambs were cut to 90° angles. The wooden door of the church would have opened inwards. A small depression (D: 6mm; depth: 10.5mm) on the surface of the paving north of the entrance represents wear from a door-post, while a stone-lined post-hole on the opposite (southern) side may also have been associated with the door attachment (Fig. 5.7, no. 11, F231). The second construction phase of the church walls was represented by two to three courses (H: 0.3–0.4m) laid on top of the first phase of masonry (Figs 5.5 (a), (c), (d) and 5.8 (a), (c), (d); Pl. 5.4). It was distinguished by the use of long, roughly shaped stones with small packing stones between. The top course of masonry in this phase denotes the level of the wall plate of the second building phase, when the walls were raised and the church re-roofed. The west doorway was increased to its present height of 1.67m. The head of the doorway was formed by two lintels with a packing of smaller stones between them. The inner lintel was a reused cross-slab, its decorated side

facing downwards (Section 7.1, Cross-slab 16). This stone was probably damaged beforehand as the broken top of the slab has flecks of mortar adhering to it, indicating that it was already in this condition when it was placed on the doorway. Remains of a lime plaster adhered to the inner surfaces of all four walls of the church, including the doorway. The masonry of the west wall changed dramatically above the second phase, indicating a third building phase. Large, squared blocks with small levelling stones beneath raised the wall for a second time. Two large stones of this masonry survived at the top of the inner face of the south wall and one on the inner north face (Fig. 5.5 (c), (d); Pl. 5.4). Above this level, the west gable (extant H: 3.32m) was constructed of stones of various sizes, randomly coursed (Figs 5.5 (d) and 5.8 (d); Pl. 5.4). Conservation work on the west gable required that a number of large stones, which had become displaced from its northern end, be removed and re-set. It was found that the large facing blocks were bonded together by a small amount of clay and mortar,

the excavations 65

EAST WALL OF CHURCH North

South

East D9

West D10

209:1

8 (a)

209:2

2 6

South D11

(b) North D10

West D11

East D12

9 9

10 RD

RD 3 4

(c)

1 Clay (F216)

RD 1

3 (d)

5 Charred surface (F232)

9 Cross-slab 15

Plaster/mortar Paving

2 Burnt debris (F221)

6 Soil and schist (F247)

10 Cross-slab 23

3 Silver/grey clay (F224)

7 Fragmented schist (F248)

11 Boulder clay (F271)

4 Paving (F245)

8 Stones (F269a–c) belonging to 'slotted feature'

Limit of excavation RD

Rabbit disturbance

Fig. 5.11 Plan and elevations of altar (F209) at east end of the church: (a) plan; (b) north face; (c) west face; (d) south face. Location of elevations on Fig. 5.7.

and that no core material was present. The width of the side walls suggests that they were of insufficient thickness to support the lateral thrust of a stone roof, an observation initially made by Macalister (Section 3.3). Different heights relate to the first, second and third building phases, but it is probable that the pitch and composition of the roof did not alter significantly from one building phase to the next (see the conjectural reconstruction in White Marshall and Rourke 2000, 74, fig. 44). The structure supporting the roof is thought to have been made from timber with a covering of sod, thatch or possibly shingles (wooden tiles). Remains of a paved surface survived within the church doorway and in a small area (1m2) just inside it (Fig. 5.7, F29; Pl. 5.1). It was composed of closely fitting slabs of

mica-schist between which small stones were placed to fill the joints. The deposits underlying this paving were not investigated. The jagged edges of the extant paving showed that it had originally covered a more extensive area, possibly the entire interior of the church. Paving was continuous around the exterior of the building, where it had covered the entire area between the church and the surrounding enclosure wall (Fig. 5.7, F29; Section 5.2.2).

The altar A small, low, roughly square altar (1.05m east–west x 1.1m north–south) had been built up against the east wall of the church (Fig. 5.7, F209; Pl. 5.5). It survives to a height of c. 0.9m. Its position, slightly north of centre, is in line with that of the west doorway directly opposite it. The lime

66 high island: excavation of an early medieval monastery

Pl. 5.8 North wall of the aumbry, showing corbel (John O’Brien).

Pl. 5.7 View of the aumbry interior with sillstone in situ separating the upper and lower cavities (G. Scally).

plaster adhering to the inner surface of the church wall indicated that the building had been in use for some time before the altar was constructed. The altar was built at the same foundation level as the church walls, i.e. on the layer of grey clay, and, where this was absent, on the underlying burnt debris (Fig. 5.11 (b)–(d), nos. 1, 2, F216, F221). The only place where the foundation level differed was on the northern side, where the upright stones (F269) that pre-dated the church had been retained and incorporated within the make-up of the altar, suggesting that the altar was built for use during the first phase of the church (Fig. 5.11 (b), no. 8, F269a–c). The front of the altar was composed of narrow, rectangular-shaped slabs laid on edge lengthways and three large, squared blocks which mainly formed the quoins (Fig. 5.11 (c)). The northern, and probably also the southern faces were built with narrow slabs (Fig. 5.11 (b)). A decorated cross-slab had been inserted onto the southern side (Fig. 5.11 (d), no. 9, Cross-slab 15). As part of the conservation requirements, the entire altar was dismantled and rebuilt. Examination of the core found that it was composed of rubble masonry bonded with desiccated clay and mortar. During the course of dis-

Pl. 5.9 Grave 9 inside the church showing the burial (F208) in situ, with hone stone and net sinker or plumb-line on the right and left shoulder, respectively (G. Scally).

mantling the altar, a decorated cross was recovered lying face-down in the lowermost course of masonry within the altar (Fig. 5.11 (c), no. 10, Cross-slab 23). Specks of plaster adhered to all exposed sides of the altar, including the cross-slab, which indicated that the altar was plastered after the cross-slab was put in place. Analysis of the mortar sample taken from the core of the altar determined that a single coarse aggregate, schist, was predominant, whereas the sample taken from the replastering of the east wall of

the excavations 67

5 4 1

Altar (F209)

1 Burial (F208)

3 Hone (208:1)

2 Grave cut (F229)

4 Line sinker or plumb-line (208:2)

5 Aumbry (F73) Obscured

Upright stones 0

Fig. 5.12 Grave 9 with burial (F208) inside the church.

the church and the altar contained gneiss in addition to schist (Section 9.4, F209, ’96 4, ’98 18; Appendix A).

Grave 9 The burial of an adult male (F208) between twenty and 30 years of age was revealed in an unmarked grave, orientated east–west, in the north-eastern corner of the church (Fig. 5.12, no. 1, Grave 9; Pl. 5.9). It was possible to establish, despite disturbance caused by burrowing rabbits, that the grave had been cut into the silver/grey clay and, where this did not exist, into the burnt debris deposit, its base truncating the surface of the natural boulder clay (F216, F221, F271). The grave cut (L: 1.7m) narrowed in width from head to foot (0.6–0.4m) and it decreased in depth (by 0.2m) from west to east (Fig. 5.12, no. 2, F229). The sides of the grave, in particular at the western end around the head, were marked by small, flat stones sloping downwards into the grave. The remains of the extended

inhumation were in a poor state of preservation and had been partly displaced (Section 8.1, Grave 9). It appeared that the right arm had been slightly flexed and lay over the pelvic area, while the left arm had been bent at the elbow and the hand rested beside the pelvis. A bone pin, probably used to secure the burial shroud, was recovered from just above the pelvis on the right-hand side (Section 6.8, 208:3). The burial has been dated to cal. AD 1163–1230 (UB-4000, 849±16 BP), (Table 5.1). Stone grave goods had been placed on the shoulders: a hone on the right shoulder and an object that is probably a line sinker for fishing on the left shoulder (Fig. 5.12, 208:1, 208:2; Pl. 5.9). The presence of grave goods, unusual in graves of the Christian tradition, was the impetus for the stable isotopic analysis of the skeletal remains, in order to establish the cultural background of this individual (Section 8.2). The results indicated that he was unlikely to have originated from Scandinavia or the west

68 high island: excavation of an early medieval monastery Table 5.1 Radiocarbon dates for High Island, Co. Galway.1

Phase

Area

F. no.

Lab. no.

Material dated

Radiocarbon years

95.4% cal. BC/cal. AD (two standard deviations)

Iron Age

F93L

UB-4986

Charcoal: Pomoideae

2201±38

cal. BC 387–170

Iron Age

F93L

UB-6452 (AMS)

Charcoal: Corylus

2161±33

cal. BC 262–102 (58%) 360– 273 (42%)

F37

UB-3999

Human remains

1126±22

cal. AD 881–977

F40

UB-4155

Human bone

1027±19

cal. AD 980–1023

F70

UB-4266

Human bone

1023±23

cal. AD 980–1025

F93a–g

UB-4522

Charcoal: Corylus/ Juniperus/Erica/ Quercus

1171±41

cal. AD 728–971

F221

OxA-8918

Charcoal: Corylus

1110±40

cal. AD 860–1020 (91.6%) 820–850 (2.4%) 780–800 (1.5%)

F221

OxA-8946

Sheep bone

1182±36

cal. AD 770–980 (93.1%) 720–750 (2.3%)

F337

OxA-8917

Charcoal: Corylus

1100±40

cal. AD 870–1030

F4037

UB-4987

Charcoal: Corylus

1138±48

cal. AD 778–1000

F41

UB-4156

Human bone

913±19

cal. AD 1033–1169

F208

UB-4000

Human bone

849±16

cal. AD 1163–1230

F457

UB-4521

Charcoal: Corylus/ Fraxinus/Juniperus/ Pomoideae/Betula/ Quercus

888±41

cal. AD 1027–1238

F515

UB-6453

Charcoal: Corylus

801±34

cal. AD 1176–1276

F803n

OxA-14042

Charcoal: Corylus/ Betula

968±27

cal. AD 1017–1070 (39.2%) 1080–1128 (34.5%) 1136–1159 (20.6%) 1003–1008 (1.1%)

F4079

UB-6454 (AMS)

Charcoal: Salicaceae/ Corylus/Fraxinus/ Alnus/Ericaceae

888±30

cal. AD 1117–1217 (65%) 1042–1107 (35%)

F8026

OxA-13665

Human bone

956±22

cal. AD 1080–1128 (38.8%) 1021–1070 (35.2%) 1136–1159 (21.4%)

F509

UB-4988

Barley grains

580±40

cal. AD 1287–1424

1 UB (Queen’s University, Belfast, Laboratory); Radiocarbon Calibration program after Stuiver M. and Pearson, G.W. 1986; Stuiver, M. and Reimer, P.J. 1993. OxA (Oxford Accelerator Laboratory) Oxcal computer program (v. 3.5) after Bronk Ramsey, C. 1995.

the excavations 69

4 1

Altar (F209)

1 Pseudo Grave 10 (F214)

4 Fill (F206) of Grave 9

Quartz pebbles

2 Mortar, clay & shells (F214a/b)

5 Aumbry (F73)

Obscured

3 Grave-marker catalogue no. 53

Upright stones

Fig. 5.13 Pseudo Grave 10 (F214) inside the church.

of Ireland, but from another part of Ireland or from Britain. The body had been covered with sticky, dark grey clay containing small stones. Small, rounded quartz pebbles (F206) along with a small fragment of mortar were found in the grave fill.

Phase 3: late 12th/early 13th to mid-15th century A stone-lined pseudo grave (i.e. a grave not used for burial), measuring 1.2m east–west x 0.6m north–south, was uncovered in the north-western corner of the church (Fig. 5.13, no. 1, Grave 10). The grave was set on a bed of small stones pressed into the underlying burnt debris (F221). The western end was formed by a slab propped against the church wall and to each side there was a similar, but smaller, slab set on edge at right-angles. The remainder of the northern side was formed by the church wall. All three slabs stood proud of the level of the earlier paving by 0.2m, indicating that this surface, if it had still existed

in this area, would have been removed to insert the grave. At the southern side of the grave there was a shallow slot and another slot was at the eastern end (Fig. 5.13). Small stones outside the slots probably functioned as wedges, to hold in place the larger stones that may have been set into the slots. Within the area defined by the stones there were two distinct deposits. The lower fill was a gritty, brown clay (depth: 30mm) containing charcoal flecks and mortar fragments, above which was a similar depth of crushed shells with attached mortar fragments (Fig. 5.13, no. 2, F214a, F214b). Birch, hazel, oak and willow/poplar were identified in the charcoal, and fish bones were dominant in the faunal remains of the lower fill (Sections 9.1, 9.3, F215). Forty-five white quartz pebbles were scattered throughout both deposits, and nine quartz pebbles were found in a pile in the north-western corner between the sidestone of

70 high island: excavation of an early medieval monastery

5 2

Pl. 5.10 Leacht (F212) in south-western corner of the church, looking west (G. Scally). 3

1m 1 Grave marker (catalogue no. 53)

5 Altar (F209)

2 Mortar floor (F202)

Mortar

3 Stones (F203)

Upright stones

4 Rabbit disturbance (F243)

Conjectural

Fig. 5.14 Later medieval features inside the church (Phase 3).

the grave and the church wall. The grave was too short to have accommodated an extended inhumation of average length and no human skeletal remains were present. The stone surround coupled with the concentration of quartz pebbles, which occur with other burials at the site, suggest this feature represents a grave. An undecorated, triangular-shaped stone set upright to the south-west of the grave probably marked its location (Fig. 5.13, no. 3; Section 7.1, catalogue no. 53). The stone had been set in a pit that had cut though the paved surface and it was secured in its upright position by stones wedged into the pit and its clay fill. A layer of mid-brown clay (F215), 90mm in maximum depth, covered the church interior and Grave 10. This layer contained a moderate quantity of sheep/goat, bird and fish bones (Section 9.1, F215). This soil was in turn sealed by a mortar floor (F202), 30mm in depth, consisting of grey schist clay mixed with cream-coloured mortar and small stones (Fig. 5.14, no. 2, F202). The floor was best preserved in the north-west and along the east wall. Elsewhere, fragments were identified sporadically where, particularly in the centre of the church, small flat stones appeared to have been laid down to form a patchy surface (Fig. 5.14, no. 3, F203). A decorated copper alloy strap,

possibly a piece of a book clasp, was found in the mortar floor (Section 6.4, 202:3).

Phase 4: mid-15th to late 20th century The base of a leacht (1.2m2) was uncovered in the southwest corner of the church, on top of the mortar floor and the stone surface (Pl. 5.10). Its north and east sides were each formed by a long, slender stone set on edge. The core of the leacht consisted of small, flat stones laid on a shallow bed of grey, stony clay in which there was a small amount of faunal remains (Section 9.1, F212). In the north-west corner of the church an undecorated stone had been set in an upright position, secured in place by the surrounding stones. The church interior, including the leacht, was covered by masonry rubble mixed with clay, overlying which were patches of fragmented schist sealed by the grass-covered sod. Finds recovered from these levels include another fragment of the decorated copper alloy strap found in the mortar floor (Section 6.4, 201:8). Modern items consist of an iron blade, a brass button/stud and a fragment of a green glass bottle (Section 6.4, 201:2, 201:7, 6–10, 201:19).

5.2.2 The paved surface Phase 2: mid-11th to late 12th/early 13th century The entire area of the church enclosure was excavated to the level of the paved surface (F29) (Fig. 5.7). The deposits on which the paving was laid were not excavated. In two places where the paving slabs were lifted, samples were taken to examine the environmental remains. One of the samples, the burnt debris, was obtained in the south-west of the enclosure and the other, the silty clay containing charcoal residues, was taken outside the north-east corner of the church. The burnt debris was identified as domestic hearth ash, which included heather, oak and peat (Section 9.3, F75). The small quantity of faunal remains contained one bird bone and the other unidentifiable fragments (Section 9.1, F75). The silty clay was identified as disturbed subsoil containing intrusive material that most likely had been in situ (Section 9.4, F105). Oak and hawthorn were

the excavations 71

Excavated area of wall

(CS 17)

Altar (F209) 3

1 Aumbry (F73)

Lintels

2 Findspot stones 4:1, 4:5 and Cross no. 24 (4:4)

Obscured Conjectural

3 Silty clay and stone (F4) The church Upright stones

Limit of excavation CS

Cross-slab

Fig. 5.15 The church enclosure during Phase 3.

identified in the charcoal, and other plant remains were bladder campion and two cereal grains; nine fish bones were also in the sample (Sections 9.1–9.3, F105, ’01B). The paved surface was laid after the extant church enclosure wall was constructed as the slabs largely respected it and did not extend underneath (Fig. 5.7). This was also the case in the north-west entrance and, where the paving survived, in the north-east entrance. The paving covered the north, south and west surfaces of the enclosure. At the eastern side, it survived adjacent to the enclosure wall but had been interrupted where the graves were situated. The misalignment of the church and the enclosure wall meant

that the area of the paving varied widely in extent. In the south-west of the enclosure, the space between the two structures was widest (2m), while it was narrowest (0.7m) at the south-east corner of the church, where it was just sufficient to allow unimpeded passage of a person between the two structures. The paving in the enclosure was at the same level as that preserved in the doorway and within the church (Fig. 5.7, F29; Section 5.2.1). The paving was comprised of slabs of mica-schist, carefully laid to fit alongside one another, and in places small stones had been used to fill the gaps. The paved surface was relatively level with a slope downwards (of 0.18m)

72 high island: excavation of an early medieval monastery

from north-west to south-west, reflecting the natural fall of the ground together with some subsidence. At the north, and to a lesser extent at the west, a shallow step was noted where slabs that partly underlay the church walls were set at a slightly lower level (Fig. 5.7, no. 10). Two exceptionally large slabs lay outside the church doorway (F20). Two other slabs, one (F64) partly underlying the south wall of the church and the other (F60) close to the north-eastern corner of the enclosure, had notches cut in their sides, suggesting attempts to form a cross shape (Fig. 5.7). This paving continued in use as the surface of the enclosure for the remainder of the monastic occupation, until the late 12th or early 13th century.

Phase 3: late 12th/early 13th to mid-15th century The paving laid down in Phase 2 continued as the surface of the church enclosure into the later medieval period (Fig. 5.7, F29). Silty clay and stone (depth: 50mm) accumulated over this paved surface in Phase 3 (Fig. 5.15, no. 3, F4). Small, flat stones were scattered on the soil all around the enclosure, suggesting that these were the remains of a rough, makeshift surface put in place after the monastery was no longer occupied.

black silt extended along the eastern and central area of the floor of the grave (Fig. 5.8 (e), no. 10, F76). A more extensive deposit of fine, yet compact, light grey/brown soil, containing fragments of mica-schist, partly overlay F76 and covered the western end of the grave (Fig. 5.8 (e), no. 9, F65). This deposit extended to the south into Grave 4, and to the north and to the east into unexcavated areas. It was also identified sporadically to the west overlying the iron-stained deposit (F76), where it was sealed by the church wall.

Grave 5 Grave 5, the northernmost of the three, was a simple earth-cut grave, 1.6m in length. No grave cut was identified for the fully extended skeletal remains, which were lying 10° south of east–west (Fig. 5.16, no. 1, F37; Pl. 5.11). The remains were those of an adult, possibly a male, aged over 55 years (Section 8.1, Grave 5). The skull and shoulder bones were lying in the grey/brown soil (F65), while the pelvis and upper leg bones lay at a slightly lower level on the silt- and iron-flecked grey clay (F76). The skull was tilted forwards to an almost upright position, so that the

Phase 4: mid-15th to late 20th century Layers of soil with low stone content (F47), 0.14m in depth, accumulated within the church enclosure on top of the Phase 3 soil and stone surface (F4) from the 15th century onwards. Small pieces of a whale skull, probably off-cuts from bone-working, were retrieved from this deposit (Section 9.1, F47). These soil layers were covered by dryer soil mixed with collapsed masonry (F2). These deposits in turn were sealed by a deep layer of rubble (F1), 1.1m in depth. The stones within this layer originated from the collapse of the sides of the enclosure wall and, at the eastern side, from the collapse of the church wall. Several lumps of mortar or plaster were recovered from the rubble. The depth of rubble was such that it almost entirely obscured the church enclosure from view.

(CS 9)

5.2.3 The graves Phase 1: 8th to mid-11th century Three adjoining graves were associated with the earliest Phase 1 features uncovered beneath the church (Figs 5.16, 5.17, 5.18; Section 5.2.1). The graves were aligned with a structure to the west that pre-dated the church, which was on a different axis. Consequently, the graves lay partially beneath and to the east of the wall of the church (Fig. 5.22, Graves 3–5). The boulder clay beneath Grave 5 was covered by two distinct deposits (Fig. 5.8 (e), no. 14, F171). A layer of moist and sticky, iron-stained grey clay and patches of

(CS 11)

1m 1 Burial (F37) in Grave 5 The later church

CS Cross-slab

Fig. 5.16 Grave 5 with burial (F37) in situ, Phase 1.

the excavations 73

Pl. 5.11 Grave 5 showing burial (F37) in situ, Phase 1 (G. Scally). (Note: the sidestones were added later when the grave was restructured in Phase 2.)

Pl. 5.12 Grave 5 showing stones at head of burial (F37) (G. Scally).

lower jaw was on the chest (Pl. 5.12). A stone had been placed on each side of the head. The flat stone (L: 0.3m) had fallen over beside the right shoulder, while a stone of similar size covered the upper left side of the skull. The lower arms were flexed slightly and overlay the pelvis (Fig. 5.16). A date of cal AD 881–977 (UB-3999, 1126±22 BP) was obtained for this burial, the earliest from the site. The remains were covered by 0.2m of dark, grey/brown, silty clay (F57), with inclusions of mica-schist fragments, small stones, occasional larger stones, a scattering of charcoal, one small fragment of mortar and nine fish bones (Section 9.1, F57). Patches of greasy silt and flecks of iron staining, similar to those in the clay (F76) on which the burial was laid, suggest that the fill is most likely backfilled up-cast from the grave. It is probable that this Phase 1 grave was marked by the in situ headstone, Cross-slab 9 (Fig. 5.16; Pl. 5.11). The slab was carved on the main face with an ornate cross in relief, and on the reverse face and on the top edge with incised linear crosses. It is not possible to determine if the footstone belongs to this Phase 1 grave or if it was inserted during its restructuring in Phase 2. The side of the footstone that faces the grave has a cross in

relief, resembling an anthropomorphic figure (Section 7.1, Cross-slab 11).

Grave 4 Grave 4, lying immediately south of Grave 5, was a wellbuilt, stone-lined grave with a paved floor (Fig. 5.17; Pl. 5.13). The grave (1.7m x 0.4m) was orientated approximately east–west (5° south of east), with each side formed by two long slabs and a shorter upright stone at both the western and eastern ends. At the head, or western end, a smaller stone had been placed inside the upright stone. No headstone or footstone was found. The slabs forming the northern side of the grave and one of the slabs on the south side were upright (max. H: 0.41m), the other stone, at the southern side, was lying flat and partially covering the grave (Fig. 5.17). Paving, consisting of four large, irregularly shaped slabs of mica-schist, covered the floor of the grave (Pl. 5.13). Small, flat packing stones secured the large side slabs in position and the floor sloped slightly (c. 50mm) west–east. Blocking stones were identified at both ends of the grave. At the west end, one large stone (0.4m x 0.3m; H: 0.2m), surrounded by several smaller stones, was set within a partially stone-lined cut that had been dug into the

74 high island: excavation of an early medieval monastery

Burial (F40)

Cut (F273)

Support (F61) for footstone The later church Quartz pebbles Upright stones

3 0

Obscured

Fig. 5.17 Grave 4 with burial (F40) in situ, Phase 1.

Pl. 5.13 Primary (pre-church) phase of Grave 4, showing sidestones and paved floor (G. Scally). (Note: the upper sidestones were added later when the grave was restructured.)

Pl. 5.14 Burial (F40) in situ within primary (pre-church) phase of Grave 4 (G. Scally). (Note: the upper sidestones, decorated headstone and blocking stone in front were all added later when the grave was restructured in Phase 2.)

natural clay (Fig. 5.17, no. 2, F273). The exposed eastern face of the larger stone lay flush between the sidestones defining the western end of the grave. At the foot of the grave, a flat slab extended the width of the grave at c. 0.1m above floor level and projected into the grave (Fig. 5.17, no. 3, F61). The slab was partially obscured by an upright stone (H: 0.3m), which formed the back of a socket against which a footstone for the grave would probably have rested. Two stones, one lying on the grave fill, had probably functioned as support stones for the missing footstone. A deposit of sand (F39a), 10–15mm in depth, covered the floor of Grave 4. Above the sand, the almost complete remains of an extended adult male were exposed (Fig. 5.17, no. 1, F40; Section 8.1, Grave 4). The burial has been dated to cal. AD 980–1023 (UB-4155, 1027±19 BP). The body had been carefully laid out, with the left forearm placed over the top of the pelvis and the right forearm rested on top of the left hand. Due to features associated with the later restructuring of the grave, visibility around the head area was restricted and it was unclear which direction the head was facing. Quartz pebbles had been placed beneath the pelvis, between the legs and knees and around the

the excavations 75

2 1

(CS 4)

(CS 7)

1 Burial (F70)

Obscured

2 Headstone (CS 5)

The later church

Upright stones

CS Cross-slab

Quartz pebbles

Fig. 5.18 Grave 3 with burial (F70) in situ, Phase 1.

Pl. 5.15 Primary (pre-church) phase of Grave 3 with burial (F70) in situ (G. Scally).

ankles (Fig. 5.17; Pl. 5.14). A fragment of porphyry was recovered from around the skull area (Sections 6.1, 6.2, 38:1). This was the only excavated grave in this area of the site to contain grave goods. Traces of sand, containing a distinctive black fleck and a significant quantity of shell fragments, was noted most especially over the ribs and lower leg bones. A fill of moist, brown silty clay (depth: 0.4m), which was full of pieces of schist, covered the burial. Inclusions of charcoal and flecks of mortar were present and cattle and bird bones were identified (Section 9.1, F38).

Grave 3 Grave 3, which adjoined Grave 4 to the south, was an exceptionally well-built, stone-lined grave with a paved floor, constructed in a similar manner to Grave 4 (Fig. 5.18; Pl. 5.15). Grave 3 (L: 1.64m x W: 0.44–0.32m), orientated 20° south of east–west and tapered progressively inwards as it extended to the foot, or eastern end. The sides were formed by stones set on edge. The southern side had a single stone (H: 0.34m x T: 80mm) running the full length of

the grave. One end of the stone underlay the church wall, while the other end had been fractured (Fig. 5.18). The northern side was formed by two stones of similar height and thickness. Like that on the southern side, the end of the more westerly of the two stones was sealed by the masonry of the church wall. The western end had an upright slab set between the sidestones (Fig. 5.8 (e)). Behind this slab, the headstone, a decorated cross, was slightly off-centre towards the north (Fig. 5.18, no. 2, Cross-slab 5). It was re-set in this position when the grave was realigned (Phase 2). At the other end of the grave, no end or footstone was found centered between the sidestones. Instead, two decorated slabs had been re-set upright behind or close to the ends of the respective sidestones when the grave was later realigned (Fig. 5.18, Cross-slabs 4, 7). Cross-slab 4 lay to the south, partly within Grave 3, while Cross-slab 7 lay entirely beyond the northern side of the grave and beyond the burial. It is suggested that Cross-slab 4 was the footstone for this earlier grave and was moved to the south when the graves were restructured (Phase 2). The floor had five flat slabs of mica-schist that covered the full length and breadth of the grave, with the exception of the western end. All of the paving slabs continued beneath the sidestones on both sides. They had been carefully laid and the surface had a slight gradient from west to east. At the west end of the grave, two irregularly shaped pieces of stone with jagged, broken edges protruded into the grave. The larger stone extended from beneath the southern side, while the smaller stone was partly sealed beneath the slen-

76 high island: excavation of an early medieval monastery

der upright end-stone. Both pieces may have belonged to a sixth paving slab, which was possibly broken during the later restructuring and realigning of the grave (Phase 2). A thin layer of sand (F72a) covered the paved floor. Above this, the skeletal remains were lying in an extended position (Fig. 5.18, no. 1, F70; Pl. 5.15). The burial was of an older male aged over 55 years (Section 8.1, Grave 3). The date of cal. AD 980–1025 (UB-4266, 1023±23 BP) was almost identical to that obtained for the individual in the adjoining Grave 4. In places, most particularly around the skull area, a few millimetres depth of sand (F72) was identified overlying the remains. The sand contained fragments of shell and it had the identical distinctive black fleck as the sand in Grave 4 (F38a). The skeletal remains in Grave 3 were well preserved. The skull, shoulders and upper arms were framed on the southern side by a long, narrow stone, while a smaller, but similarly shaped stone lay adjacent to the skull on the northern side. The head was turned so that it faced to the north. The lower arms and hands were placed over the pelvis but did not touch one another, and the outstretched legs lay slightly apart. One small quartz pebble lay adjacent to the right elbow, while several more lay around and between the ankles (Fig. 5.18). The body was covered by moist, brown, stony soil (F69) that extended the full length of the grave and was deepest (0.32m) at the foot end, where it reached the top of the sidestones. Fragments of mortar were mixed with the fill, most of which were located close to the feet. Analysis of the mortar produced unusual results (Section 9.4, F69, ’97 7; Appendix A). It differed from all other mortars examined in that the lime was made with limestone from the mainland or from another island location with these deposits, and its mixed aggregate was most likely obtained from a stream. The grave fill also contained faunal remains, which included fish and bird bones (Section 9.1, F69). It is probable that the grave was originally covered by a recumbent slab. The area (1.4m north–south x 1.6m east–west) to the south of Grave 3 as far as the south-eastern corner of the church was investigated to bedrock level (Figs 5.8 (e), 5.22). On the southern side of Grave 3, the bedrock stepped down in a shelf-like fashion and it appeared that it may have been artificially cut to create a hollow (depth: 0.25m). For safety reasons, it was only possible to explore a reduced area (0.54m north–south x 0.77m east–west) at the south-eastern limit of the cutting. Here, the bedrock was at a much deeper level where it was covered by patches of black silt (F43) 70mm in depth, with an organic clay (F43a) on top (Fig. 5.8 (e), nos. 5, 6). The overlying layer of moist, dark grey clay and mica-schist fragments (F42) was 0.3m deep and contained charcoal of hazel, heather and yew (Fig. 5.8 (e), no. 4; Section 9.3). Plant remains were those of goosefoot/orache, ribwort plantain and barley (Section 9.2). Fish scales and

3 1

0 1 Iron-stained clay (F76) 2 Clay (F65)

1m 6 Grave 8 The later church

3 Grey clay (F205)

Charcoal

4 Clay (F68)

Upright stones

5 Grave 5

Obscured

Fig. 5.19 The earliest deposits in area between Grave 5 (to the south) and Grave 8 (to the north), Phase 1.

bones were also present (Section 9.1). Analysis of this layer identified it as a mixture of organo-mineral soil and charred material that had been dumped or used as an infill (Section 9.4, F42, ’96 E). Small deposits of black silt and orange ash were found in the corner (Fig. 5.8 (e), nos. 2, 3, F42a, F42b). A deposit of moist, dark grey clay containing flakes of micaschist filled the hollow in the bedrock to the north (Fig. 5.8 (e), no. 7, F50). The dark grey clay extended beyond the area excavated towards the west, where it was sealed beneath the church wall, and to the east and south, where it was sealed beneath the paved surface (F29) covering the church enclosure and its underlying deposit. To the north, the clay extended to the sidestone of Grave 3. The area (1.2m north–south x 1.6m east–west) located between Grave 5 and the recumbent slab, Grave 8, at the north-eastern corner of the church, was excavated to undisturbed boulder clay (Fig. 5.8 (e), no. 14, F171). Finely textured, moist and sticky, iron-stained grey clay (F76), 50mm in depth, overlay the boulder clay at the central

the excavations 77

2 3

1 Recumbent slab (CS 14) 0

1 Stones (F54)

The later church

2 Slabs (F30)

Upright stones

3 Grave 8

Obscured

Charcoal

Fig. 5.20 Stone slabs (F54, F30) in area between Grave 5 (to the south) and Grave 8 (to the north).

and eastern end of the excavated area. This was the same deposit revealed at a comparable level in Grave 5 (Figs 5.8 (e), no. 10 and 5.19, no. 1, F76). The clay contained fragments of mica-schist and flecks of charcoal. It was sealed by a more extensive deposit of dry, friable, light grey/brown clay with mica-schist fragments (F65) that was also recognised under Grave 5 (Figs 5.8 (e), no. 9 and 5.19, no. 2). The clay was covered by isolated patches of a rich, humic brown soil mixed with lenses of charcoal (Fig. 5.8 (e), no. 8, F55, F56). Hazel, blackthorn and oak were identified in the charcoal (Section 9.3, F56). A single grain of barley and also fish and bird remains indicate a domestic origin for this soil (Sections 9.1, 9.2, F56). Notable inclusions in this deposit were small patches of redeposited grey clay (Fig. 5.19, no. 3, F205). These deposits were most in evidence at the western end and the grey clay continued beneath the church wall, where similar patches were found contemporaneous with the Phase 1 paved area (F245) that pre-dated the church (Figs 5.2, 5.3, no. 6, F205; Section 5.2.1). An

The later church Obscured

Fig. 5.21 Grave 8 with recumbent slab (CS 14) in situ partly sealed by the church wall.

area of dark brown, charcoal-flecked clay (depth: 10–20 mm) was revealed, extending beneath the recumbent slab of Grave 8 to the north (Fig. 5.19, no. 4, F68). The charcoal consisted of hazel and blackthorn (Section 9.3, F68). Archaeobotanical analysis identified goosefoot/orache and a single grain of both barley and oat (Section 9.2, F68). There were also fragments of faunal bones, the species of which could not be determined (Section 9.1, F68). Large slabs of mica-schist (up to L: 0.7m x W: 0.32m and T: 70mm) were laid diagonally on top (Figs 5.8 (e), no. 11 and 5.20, no. 1, F54). The slabs were in close proximity to one another and had the appearance of having been packed into the available space. Three slabs, orientated in an east–west line, were positioned to form an edge with the recumbent slab of Grave 8 (Figs 5.8 (e), no. 12 and 5.20, nos. 2, 3, F30, Grave 8). The slab at the western end lay partly beneath the church wall. The decorated recumbent slab covering Grave 8 was visible for approximately two-thirds of its length (1.18m), where it projected from the north-eastern corner of the church (Figs 5.7, 5.21, no. 1, Cross-slab 14). The remainder of the slab lay under the church wall and it was not possible to excavate underneath it to determine if it marked a burial. The orientation of the slab was 3° north of east– west, which was similar to that of Grave 4 but significantly different from the orientation of Graves 3 and 5. The carv-

78 high island: excavation of an early medieval monastery Pl. 5.16 The eight Graves (1–8) behind the east wall of the church, looking west (graves numbered 1 to 8, left to right). Note the window head overlying Graves 5 and 6 (G. Scally).

ing on the slab suggests that it may be lying inverted, with its top to the east (Section 7.1, Cross-slab 14).

Phase 2: mid-11th to late 12th/early 13th century Eight graves outside the east wall of the church were marked by fourteen associated funerary monuments with carved crosses on them (Fig. 5.22; Pl. 5.16). Significantly, the headstones from two of the three Phase 1 burials (in Graves 3 and 5) were incorporated into the external masonry of the east wall of the church wall when it was under construction in the mid- to late 11th century (Fig. 5.8 (e); Section 5.2.1, Phase 2). The church, the erection of the grave-markers and probably also the restructuring of Graves 3, 4 and 5 were carried out at the same time when the focal area of the ecclesiastical site was reorganised and rebuilt. In this restructuring phase, no human remains were interred in the graves with the exception of Grave 1 at the south-eastern corner, which post-dates the construction of the church. Grave 3 was realigned so that its orientation was changed to 3° south of east–west and differing by 17° from the orientation of the Phase 1 grave beneath. The restructured grave was the same length (1.6m) and was rectangular-shaped (W: 0.4m) and its position was realigned by 0.4m maximum to the north (Figs 5.18, 5.22–5.24). Two flat slabs rested on top of the sidestones and covered the grave for approximately two-thirds of its length (Fig. 5.23, no. 1, F67; Pl. 5.17). The slab at the western end (L: 0.46m) was partly sealed at its southern corner by the masonry of the church wall. The second, very large slab (L: 0.8m) covered approximately half of the grave and its northern side over-sailed the underlying (Phase 1) sidestones by 0.2m. No part of the underside of the slabs touched the earlier grave-fill (F69). At the eastern end, in that part not covered by the slabs and where the fill was

(CS 17)

(F29)

(F60)

7 CHURCH

(CS 13)

(CS 12)

6 (CS 10)

(CS 9) 5 (CS 8)

(CS 11) 4

(CS 5)

(CS 6)

(CS 7) 2

(CS 2)

(CS 4) 1

(CS 3)

CHURCH ENCLOSURE WALL

(F13) (CS 1)

1 – 8 Graves outside east wall of church Upright stones

CS Cross-slab

Fig. 5.22 Graves 1–8, with associated grave-markers, outside the east wall of the church, all surrounded by the paved surface (F29), Phase 2.

the excavations 79

Pl. 5.17 Grave 3 showing upper sidestones and slabs (F67) along the length of the restructured grave, Phase 2, looking west (G. Scally).

Pl. 5.18 Pseudo Grave 2 during excavation. Also visible are the sidestones of the primary phase of Grave 3 to the north (RHS). Recumbent slab (Cross-slab 6), headstone (Cross-slab 5) and later (Phase 2) sidestones for Grave 3 in situ (G. Scally).

(CS 5)

(CS 4)

(CS 7)

1m 0

1 Slabs (F67) covering Grave 3 Upright stones

Obscured CS

Cross-slab

Fig. 5.23 Realignment and restructuring of Grave 3, showing slabs (F67).

Upright stones CS

(CS 6)

1m Obscured

Cross-slab

Fig. 5.24 Grave 3 after realignment and restructuring, Phase 2.

80 high island: excavation of an early medieval monastery

Pl. 5.19 Upper sidestone between Grave 4 and Grave 5 supported by spalls laid on the primary fill (F57) of Grave 5 (G. Scally).

1 Headstone (CS 8) 2 Grave fill (F38)

Upright stones CS

Cross-slab

3 Support stones for footstone (missing)

Fig. 5.25 Grave 4 after restructuring, Phase 2.

deepest, small stones were carefully placed on top of the fill (Fig. 5.23; Pl. 5.17). Small flat stones were laid on top of the slabs (Fig. 5.23, no. 1, F67). These stones formed a bed upon which the slender side slabs of the restructured grave were set on edge (Fig. 5.24). Three slabs formed the southern side of the grave, while the northern side was formed by just two slabs. All of these sidestones were shared by the adjoining Grave 2 to the south and by Grave 4 to the north (Fig. 5.22). The three slabs that formed the southern side of the grave were the same height (0.43m) and all of them were leaning inwards (Fig. 5.24; Pls 5.17, 5.18). The base of the sidestone at the eastern end of the grave was set into the fill of the grave beneath. The northern side of the grave was formed by a long slab (L: 1.22m) and a smaller slab (L:

0.42m) placed close to the church wall. These two slabs stood on edge and overlapped one another for a short length. The slabs were supported in this position by thin pieces of schist that had been wedged tightly in the space where the two sidestones overlapped (Pl. 5.13). These were also supported on their undersides by several flat pieces, which had been placed on top of the stone that was lying flat in the Phase 1 Grave 4 (Fig. 5.8(e), R). The eastern half of the long sidestone of Grave 3 was set into unexcavated soil (Fig. 5.24). The fill of dry brown soil (depth: 0.3m) overlay the flat slabs (F67) and extended the full length of the restructured grave. In its upper part, the fill became increasingly stony and approximately 40 quartz pebbles were scattered throughout. Charcoal flecks and mortar fragments were in the soil. The mortar sample examined indicated that it contained aggregates of stone, including quartzite from High Island, and that the lime was produced from shells (Section 9.4, F66, ’97 6; Appendix A). The faunal remains were those of sheep/goat, fish and birds (Section 9.1, F66). The grave was covered by a decorated recumbent slab of limestone, which was resting on the fill (Fig. 5.24, Crossslab 6; Pl. 5.16). A cross was re-set to a central position marking the head of Grave 3 and was also incorporated into the masonry of the church wall (Fig. 5.8 (e), E, Cross-slab 5). Both faces of the cross were decorated with roundels and on the concealed face, the roundel contained a low relief cross (Section 7.1, Cross-slab 5). Prior to excavation, only the upper one-third of the cross was visible (Pl. 5.18). The shaft was set in a pit (F275), the base of which truncated the boulder clay (F171). The full extent of the pit was not exposed as it extended beneath the church wall. From what could be established, it seemed that the sides of the pit were lined with small, thin stones and it was filled with brown, sandy silt and packing stones. A stone set deep in the ground behind the cross shaft appeared to function as a support. The foot of the grave was marked by a cross, which was centrally placed between the upper sidestones (Figs 5.18, 5.23, 5.24, Cross-slab 7). This cross was decorated on its western face only and was positioned facing into the grave. The images depicted on the slab were in two sections. A human figure stands in an orans position on the shaft, and on the upper cruciform part of the stone a cross within a double roundel is framed by a border (Section 7.1, Cross-slab 7). The upper part of the stone was broken into two pieces. This limestone cross was one of only two funerary monuments recovered of non-local stone, the other being the recumbent slab covering this grave. The gently tapering shaft of the cross was set deep into the ground so that only the upper one-third was visible. It was supported from behind by a slender upright stone that was sealed by a slab of the surrounding paved surface (F29).

the excavations 81

Pl. 5.20 Headstone (Crossslab 8) of Grave 4 as found during excavation, leaning away from the church wall (G. Scally).

Pl. 5.21 Sidestones between Graves 5 and 6 showing reused cross (Cross-slab 10) in situ (G. Scally).

82 high island: excavation of an early medieval monastery

(CS 13)

(CS 12)

0 1

Stones (F52)

Slabs (F30)

CS Cross-slab

Fig. 5.26 Stone packing (F52) in Grave 6, Phase 2.

Pl. 5.22 Graves 1–8 outside the east wall of the church, with soil layer (F4) in situ. Finds spot for stones (4:1, 4:4, 4:5) as marked (G. Scally).

Grave 4 (1.7m x 0.7m), a stone-lined grave, was restructured on the same axis, 5° south of east–west, as the Phase 1 stone-lined grave beneath it (Pl. 5.14). The upper grave was defined by slender stones set on edge and a decorated headstone, which was at the base of the church wall (Figs 5.8 (e), H, 5.22; Pl. 5.14). The side slabs, two on each side, were shared by the adjoining Grave 3 to the south and by Grave 5 to the north (Figs 5.22, 5.25; Pls 5.14, 5.16). They were laid outside the sidestones of the earlier grave, with the result that this later grave was wider by 0.3m (Fig. 5.8 (e) – upright stones each side of J). On the southern side, the shorter of the two stones and part of the longer stone were on top of the earlier sidestone that had been laid flat (Fig. 5.17). On the northern side, both slabs had been set into the Phase 1 grave-fill (F57) of the adjoining Grave 5. The longer of the two sidestones had been secured in place by flat slabs wedged beneath (Pl. 5.19). The fill of dry, stony clay (F21a), approximately 0.2m in depth, covered the length of the grave and contained flecks of charcoal, a scattering of mortar fragments and faunal remains (Section 9.1, F21a). The uppermost level

of the fill was lower than the top of the sidestones by 80–100mm, suggesting that a recumbent slab had marked the surface of the grave. At the western end, close to the headstone, a large stone set within the fill extended the width of the grave (Fig. 5.8 (e), J; Pl. 5.14). This stone, which was cracked into two pieces, appeared to have functioned not only to prevent the inward collapse of the sidestones but also to prevent the outward collapse of the headstone. When the grave was excavated, this substantial headstone had fallen away from the church wall and was at an angle of 45° (Fig. 5.8 (e), H, Cross-slab 8; Pl. 5.20). Its base had remained in situ and rested on a slender, flat slab that was placed on top of the end-stone of the earlier phase of the grave. Part of a rotary quern had been placed on top of the headstone within the masonry of the church wall. The footstone of Grave 4 had been removed, but its socket was found and the support stones uncovered (Fig. 5.25, no. 3; Pl. 5.14). Grave 5 (1.6m x 0.5m) was restructured as a stone-lined grave on the same orientation (10° south of east–west), above the primary earth-cut grave (Fig. 5.22, no. 5). The restructured grave was defined by sidestones set on edge, with an undecorated recumbent slab marking its surface (Pl. 5.16). Its western end was marked by a decorated headstone that also formed part of the church wall foun-

the excavations 83

Q Q Q

1 2 Q

(CS 13) (CS 12)

1 Recumbent slab (F25) 2 Stony clay (F27)

CS Cross-slab Q

Upright stones

Quartz

Fig. 5.27 Graves 6, 7 and 8 with recumbent slab (F25) and associated deposits (F25a), (F27) in situ, Phase 2.

dations (Fig. 5.8 (e), L; Pl. 5.11). Its eastern end was marked by a decorated footstone (Fig. 5.22, Cross-slab 11). The side slabs were embedded into the surface of the Phase 1 gravefill (F57) and were mainly above the level of the earlier in situ burial, though the left arm bones were crushed (Pl. 5.11). No part of the skeletal remains was contained between the sidestones (Fig. 5.8 (e), N). The slabs to the south were shared by the adjoining Grave 4. On the other (northern) side, the slabs stood upright to a maximum height of 0.54m. One of the sidestones rested on the slabs uncovered at Phase 1 level to the north (Fig. 5.20, no. 1, F54; Pl. 5.11). The easternmost side slab was a reused crossslab, its decorated side facing away from Grave 5 (Fig. 5.22, Cross-slab 10; Pl. 5.21, Cross-slab 10). The fill of sticky, light grey, silty clay (F32), 0.3m in depth, covered the length and breadth of the grave. The surface of the clay was partly covered by a large, undecorated and irregularly shaped slab (L: 1.08m x W: 0.44m and T: 40mm) that covered approximately two-thirds of the grave (Figs 5.8 (e), M, and 5.22, no. 5; Pl. 5.16). Two smaller slabs lay at a slightly lower level and filled the gap between the recumbent slab and the church wall. There was no evidence to suggest that the headstone erected for the Phase 1 grave was altered in any way when the grave was restructured (Pl. 5.11, Cross-slab 9). It was the lower-

most stone of the church wall, and the masonry to the north of the headstone appeared to have been shaped to fit around it (Fig. 5.8 (e), L). It was not possible to determine with which phase the footstone, also a cross-slab, was contemporaneous (Fig. 5.16, Cross-slab 11; Pl. 5.11; Section 7.1, Cross-slab 11). Pseudo Grave 6 (L: 1.6m), aligned with the church wall, was defined at surface level by a large, undecorated recumbent slab and at its eastern end by a decorated footstone (Fig. 5.22, no. 6, Cross-slab 12; Pls 5.16, 5.22). Immediately above the grave and within the masonry of the church wall, an empty niche or recess would originally have held the headstone (Fig. 5.8 (e), O). On the southern side, Grave 6 was demarcated by the side slabs of Grave 5. The basal layer of grave-fill was grey schist clay (depth: 0.2m), which covered the Phase 1 slabs (Fig. 5.20, no. 1, F54). Small, rounded quartz pebbles were recovered within this deposit, and a concentration of up to twenty pebbles lay at the eastern end, close to the footstone. A layer of randomly placed stones covered the schist clay (Fig. 5.26, no. 1, F52). Four stones in the central area stood proud of the rest. Dry, stony, grey clay (depth: 0.25m) overlay the stones. Several angular and rounded quartz pebbles were recovered from within and on the surface of the clay. Lumps of mortar were present and the sample analysed contained local aggregates of schist and quartzite and the lime binder was produced by burning white shell (Section 9.4, F25a, ’96 4; Appendix A). Plant remains of common orache and campion were identified from this clay (Section 9.2, F25a). There were significant quantities of bird bones, predominantly storm petrel, but also herring gull, Manx shearwater and puffin, and also small, white land snails and fish bones (Section 9.1, F25a). The recumbent slab marking the grave overlay the clay and was supported on the four stones that stood proud (Fig. 5.27, no. 1, F25; Pl. 5.16). Several smaller pieces, possibly broken fragments of the slab, were scattered close by. The footstone was decorated on its east-facing side only (Fig. 5.27, Cross-slab 12). The entire area of carving was visible above the Phase 2 paving (F29) behind the footstone, which continued in use as the surface of the church enclosure (Pl. 5.22). Pseudo Grave 7 was defined by a footstone (Fig. 5.27, Cross-slab 13: Pls 5.16, 5.22). The slab with an incised cross on the side facing the grave stood close to vertical above the paved surface (F29) of the church enclosure. Excavation revealed that the stony clay layer covering the Phase 1 slabs (F30) was the same deposit (F25a) as that identified in Grave 6, and this soil continued on northwards over the recumbent slab of Grave 8 (Fig. 5.27, no. 2, F27). In the location of Graves 7 and 8, patches of dark silt were evident in the clay together with large, irregularly shaped lumps of quartz (Fig. 5.27, Q). This stony clay formed the surface of Graves 7 and 8, extending to the slender slab set

84 high island: excavation of an early medieval monastery

(CS 1)

(F13)

Pl. 5.23 Grave 1 with burial (F41) in situ (G. Scally). 0

Burial (F41)

50 cm

CS Cross-slab

Upright stones

Fig. 5.28 Grave 1 with burial (F41) in situ, Phase 2.

on edge (L: 0.8m) that delimited the area of the graves at the north-eastern corner of the church. The western end of this slab lay flush against the masonry of the north wall and it overlay the paved surface (F29) of the church enclosure. Pseudo Grave 2 (1.6m x 0.6m) was orientated 3° south of east–west (Fig. 5.22, no. 2). Its surface was partly covered by a substantial decorated recumbent slab (Cross-slab 3) and its foot was marked by upright Cross-slab 4 (Fig. 5.22; Pl. 5.16). The west end was unmarked at surface level and instead a decorated cross had been placed in a niche in the church wall immediately above the grave (Figs 5.8 (e), C, and 5.22, Cross-slab 2; Section 7.1, Cross-slab 2). Grave 2 was separated to the north by the sidestones of Grave 3 and there was no demarcation on the southern side (Fig. 5.22). The stratigraphy in Grave 2 was found to consist of similar layers of soil and stones to those revealed in Graves 6 and 7. A layer of dark grey/brown clay fill (depth: 0.1–0.15m) covered the length of Grave 2 above

the Phase 1 soil (F50), and at its lowermost level it was above that of the base of the church wall. Inclusions of quartz pebbles, charcoal flecks and fragments of mortar were found throughout. In the fill, a barley grain was identified and also fish and bird bones (Sections 9.1, 9.2, F46). Stones, one to two courses deep, were set into the clay and were packed tightly along the length and breadth of the grave. Three stones, all of which were large, were set at a slightly higher level than the other stones and formed a line demarcating the southern extent of the grave (Pl. 5.18, A–C). Two other stones of similar dimensions that were placed at the same level lay at the western end, one of which abutted the church wall, while the other overlay the sidestones of Phase 1 of the adjoining Grave 3 (Pl. 5.18, D, E). Above, a layer of small stones mixed with dry, friable clay (depth: 0.1m) filled the area of the grave. Several quartz pebbles, a small number of mortar fragments and also faunal remains were recovered from within the deposit (Section 9.1, F45a). The large, elaborately decorated recumbent slab was broken into three pieces. It extended for two-thirds of the length of the grave, resting on the clay, and it was supported by the five large stones that had been strategically placed underneath it (Fig. 5.22, Cross-slab 3; Pl. 5.16).

the excavations 85

Although the footstone was centrally placed relative to the decorated recumbent slab, it lay immediately behind the sidestone for the earlier phase of the adjoining grave (Figs 5.18, 5.22, Cross-slab 4). Both faces of the footstone were carved, the more elaborate Latin cross in low-relief facing the grave (Section 7.1, Cross-slab 4). The slab displayed a well-formed and intact base tenon, but the stone with the socket to hold the tenon was not located. The lower one-third of the footstone was set into the ground thus obscuring the decoration on this part. The slab was secured in its upright position on the side facing the grave by the upper layers of fill (F45a, F46) and on the outer side by the paving slabs of the surrounding surface (Fig. 5.22, F29). A small undecorated stone, which had partly subsided out of position, was situated between the footstone of this grave and that of Grave 1 to the south (Fig. 5.22, F13; Pl. 5.18). Grave 1, located at the south-eastern corner of the church, was a simple earth-cut grave, cut into the surface of the Phase 1 clay (Fig. 5.8 (e), no. 4, F42). The skeletal remains of a middle-aged or older male, orientated east– west, were lying in the shallow grave (Fig. 5.28; Section 8.1, F41). The burial was dated to cal. AD 1033–1169 (UB4156, 913±19 BP; Table 5.1). The poorly preserved skeletal remains were complete (Fig. 5.28; Pl. 5.23). A small pillow stone supported the right and left sides of the head, which was tilted downwards with the jaw touching the chest. The left arm was flexed and placed over the pelvis while the right arm was straight, apart from the hand that was turned in beneath the hip. The knees were flexed and the legs were crossed just above the ankles. A large stone weighed down the right femur. The head rested against the church wall and the left foot against the footstone. The flexed knees and the tight fit of the individual between the wall and the footstone indicated that he was too tall for the grave (L: 1.5m) that had been dug for him. A layer of stones (depth: 0.3m) had been placed on the burial. Quartz pebbles and lumps of mortar were scattered amongst the stones. A large quantity of bird remains were recovered, as well as a small number of cattle, sheep/ goat and fish (Section 9.1, F11a). The stones were covered by grey, friable, stony clay (depth: 0.1m). This layer was partly sealed by a slab that overlay the south-eastern corner. The footstone, a cross decorated on the upper part of both broad faces, was secured in place by the gravefill (Figs 5.22, 5.28, Cross-slab 1; Pl. 5.23). Grave 1 did not have a headstone. Two slabs set on edge at the northern side marked the boundary of the area of the graves at the south-eastern corner of the church (Figs 5.22, 5.28; Pls 5.16, 5.23).

Phase 3: late 12th/early 13th to mid-15th century The graves outside the eastern end of the church with its

façade of decorated headstones remained exposed for pilgrimage and veneration of the men buried there. Three small, water-rolled stones were placed on the soil close to the footstones of Graves 2 and 3 (Fig. 5.15, no. 2; Pl. 5.22). All were ironstone nodules, their source most likely being the Burren, Co. Clare, or the Aran Islands (Sections 6.2, 6.3). One of the stones (4:1) was decorated with incised concentric circles and another (4:5) had been shaped, but is undecorated (Section 6.2). The third (4:4) has a cross within concentric circles (Section 7.1, Cross 24). Two undecorated upright slabs were wedged between the footstones of Graves 6 and 7 in this period. Another similar upright slab was inserted to support the footstone at Grave 7 (Fig. 5.15; Pl. 5.22).

Phase 4: mid-15th to late 20th century The external façade of the east wall of the church and the adjacent graves were later obscured by soil and rubble (F2). The graves must have remained exposed into the 19th century, until the church wall above collapsed. When George Petrie visited High Island in 1820 they were visible, but by 1895 they were covered with rubble as R.A.S. Macalister did not mention them (Sections 3.2, 3.4). The arched head of the window was still in position when Petrie described it and it had collapsed less than twenty years later, according to John O’Donovan’s records. This architectural stone was found lying between Graves 5 and 6, probably in exactly the position it fell when the east wall collapsed, and the window-sill was recovered close by (Pls 5.11, 5.16; Section 6.2, 2:3, 199:1). Collapsed masonry (depth: 0.4–0.7m) covered the area (Fig. 5.8 (e), no. 1). The rubble comprised both small and large stones in which several lumps of mortar were recovered. The masonry derived largely from the collapse of the east wall of the church and some of it may also have accumulated from the collapse of the surrounding church enclosure walls.

5.2.4 The church enclosure wall The Iron Age An area (1m north–south x 5m east–west) delimited by the exterior face on the northern side of the enclosure wall and the remains of another wall (F91) further north was investigated to the depth of the undisturbed boulder clay (Fig. 5.1, Area 1 and Fig. 5.29, F91). The compact, stony, grey boulder clay (F99) was exposed throughout this cutting. At the eastern end of the excavated area, a deposit (1.2m x 0.7m and T: 50mm) of dark brown/black, peaty silt with charcoal fragments (F93L) overlay the boulder clay (Fig. 5.32, nos. 9, 10, F93L, F99). Radiocarbon dates obtained from charcoal within this deposit produced dates in the Iron Age of 360–102 cal. BC (UB-6452, 2161±33 BP) and 387–170 cal. BC (UB-4986, 2201±38 BP). These dates were

Buttress (F79)

North-west entrance (F36)

F10

F11

(F29)

Excavated area of wall

(F91)

Church

Retained burnt debris (F93a-i)

(F29)

F12

(CS 17)

Buttress (F817)

Cell B

*See detail Fig. 5.30

North-east passage (F414/F8040)

Fig. 5.29 The church enclosure with the church and Graves 1–8. Elevations F–F12, see Fig. 5.31 (a)–(h); section E–E1, see Fig. 5.32.

Cobbles (F877)

Limit of excavation

Conjectural

Obscured

Granite

Tie stones

Lintels

14 Trench (F8035)

13 Mixed clay (F8033)

12 Boulder clay (F8000)

11 Grave cut (F8034)

10 Grave 11 with burial (F8026) in situ

9 Masonry of blocked passage (F414/F8040)

1 – 8 Graves

*See detail Fig. 5.34

86 high island: excavation of an early medieval monastery

the excavations 87

2 1 Masonry (F882) 2

Cobbles (F887) Conjectural Limit of excavation

Fig. 5.30 Masonry (F832) and cobbles (F877) in situ beneath the south-east corner of the church enclosure.

Pl. 5.24 North wall of church enclosure (after conservation) showing kink at western end, looking east. Foundation of wall (F91) is visible to left of enclosure wall (G. Scally).

obtained from hazel and hawthorn charcoal and both are considered reliable indicators of an Iron Age presence at the site. The identification of hazel, hawthorn and juniper are indicative of the woodland species on High Island during this period (Section 9.3, F93L). Other deposits relating to activity that pre-dated the early medieval ecclesiastical site were not uncovered in the excavations.

Phase 1: 8th to mid-11th century At least two construction phases for the church enclosure wall were revealed. The wall face uncovered outside the upstanding northern side belongs to the earliest enclosure (Fig. 5.29, F91; Pls 5.24, 5.25). A layer of grey, gritty schist clay, a possible surface, sealed the peaty silt deposit of Iron Age date (Fig. 5.32, no. 8, F97). Heather was identified in the charcoal from the schist clay, and the lack of other species suggests that this layer relates to clearing of heather growth or the charred remains of thatch or bedding (Section 9.3, F97). It would seem that the wall was built on this clay. The remains of the early enclosure wall (L: 6m) were not complete (Fig. 5.29, F91; Pls 5.24, 5.25). Both ends had been disturbed and stones had been removed by later activity. The lowermost one to two courses were obscured

Pl. 5.25 Wall (F91) (on right) and north wall of church enclosure (on left) with burnt deposits (F93, F93a–i) between, looking west (G. Scally).

by paving at the eastern end and were not investigated (Fig. 5.32). The base of the wall was composed of substantial stones, many of which were set upright and with the exception of a few stones at the western end, which stood vertical, all the other stones were leaning inwards. Above this level, the wall which had survived (H: 0.4m) comprised two to four courses of mainly small, flat stones, all of which had collapsed inwards. The alignment of this wall

88 high island: excavation of an early medieval monastery

Pl. 5.26 East wall of church enclosure, looking north (G. Scally).

diverged by 13° from the east–west alignment of most of the later northern enclosure wall (Fig. 5.29; Pls 5.24, 5.25). The earlier wall (F91) is aligned with the southern enclosure wall, indicating that both probably belong to the same building phase. A cutting was excavated across the enclosure wall at the south-eastern corner (Fig. 5.1, Areas 1, 8). Masonry (L: 1.5m x W: 0.4m), possibly the outer face of an earlier wall, was uncovered. The drystone face was curved along its exposed length and it stood three courses (H: 0.3m) above the boulder clay upon which it was built (Fig. 5.30, no. 1). A small trench (L: 0.5m x W: 0.4m and depth: 0.25m) was excavated through this clay and a further two strata of compacted boulder clay were identified. The earlier masonry was composed mainly of small stones (0.4m x 0.2m x 0.12m max.). The original width of the wall could not be established as it was partly sealed by the later enclosure wall. At its southern end, the masonry petered out while at the other end, where it was more extensive, it stopped abruptly, suggesting that it had been cut away by later building activity, possibly by the construction of the later enclosure wall or when Cell B was built. At the base of the earlier wall face, a small area of mica-schist cobble-stones laid on the boulder clay was preserved (Fig. 5.30, no. 2, F887). The eastern side of the church enclosure wall (H:

1.4m) was the best preserved and also the most substantial. However, almost its entire outer face was obscured by the beehive hut, Cell B, which had been built up against it. Nonetheless, a line of coursed masonry (L: 1.7m) exposed beneath the rubble on top of the wall suggested that this outer face was original (Fig. 5.29; Pl. 5.26). It would appear from the evidence uncovered in the excavations that the first church enclosure was rectangular-shaped (approximately 8m north–south), with the corners rounded externally and angular internally. Remains of the original western wall were not found. The existing church enclosure is trapezoidal in plan, with no two sides parallel and where each of the four sides varies in length (Fig. 5.29). Its long axis lies east–west (9m) and it is considerably narrower from north to south at the eastern side (6.3m) than at the west (8m). The wall (H: 1m x W: 1–1.5m) had an entrance at the north-western corner and at the north-eastern corner (Fig. 5.29). Only the northern and western sides of the enclosure are aligned with those of the stone church and this misalignment has resulted in considerable variation in the space between the two structures, which is widest (2m) at the south-western corner and narrowest (0.7m) at the south-eastern corner, where it is just wide enough for a person to pass through (Fig. 5.29). Due to the varying length of each of the four sides of the enclosure, the angles of the internal corners vary significantly (Fig. 5.29). Externally, there is a combination of both angular and rounded corners. The drystone enclosure wall, where intact, was largely constructed with random courses of blocks of mica-schist and small packing stones with little quartz or granite used (Fig. 5.31). The southern side was, for the most part, well preserved, particularly in its centre section (Fig. 5.31 (d), (h)). The western side was the least substantial, especially along its inner face (Fig. 5.31 (e)). No part of its outer face was exposed as excavation did not extend to this area (Fig. 5.1). A trench, 1m in length, was opened across the northern wall where excavation revealed that the wall was extremely poorly built (Fig. 5.29). Boulder clay was exposed at the lowest excavated level. This was covered by a thin layer of sterile, silty clay and dry, fragmented schist on which two substantial, roughly hewn boulders had been set upright. The boulders were surrounded by clay and stone, which formed a compact fill. Large stones formed the inner face of the enclosure wall, a number of which overlay the fill, while on the outer face the roughly coursed masonry was also set on the clay (Fig. 5.31 (b), (g)). Whilst this was the only area where the wall was fully excavated, and thus the only place where the composition of the fill could be fully determined, it appeared to differ considerably from other parts of the enclosure wall, which were built with better quality masonry. This difference in the composition of the fill as well as its poor construction would suggest that the

the excavations 89

West F

East F2

Excavated area

Rubble

(a)

(b)

North F3

South F5

(CS 17) Rubble

(c)

East F5

West F6

Rubble

(d)

South F6

North F7

Rubble

East F7

West F8

(f)

(e)

East F9

West F10

Excavated area

(g) West F11

East F12

(h) 0

Paving

Obscured

Upright stones

Limit of excavation

Fig. 5.31 Elevations of church enclosure wall – internal: (a) north wall of passage (F36), (b) north wall, (c) east wall, (d) south wall, (e) west wall, (f) south wall of passage; and external: (g) north wall; (h) south wall. Location of elevations, see Fig. 5.29.

90 high island: excavation of an early medieval monastery

1 Peat and charcoal (F93/F93a) South E

2 Ash (F93b)

North E1

3 Charred peat and wood ash (F93c) 4 Ash (F93d) 5 Ash (F93g)

North wall of church enclosure

6 Charcoal ash and stone (F93e)

7 Peat, wood charcoal and ash (F93i)

(F422)

8 Grey schist clay, possible old ground surface (F97/F97a)

9 Peat ash and charcoal F93L (Iron Age C14 dates, [UB-4986, UB-6452] obtained from this deposit)

10 Boulder clay (F99) 10 0

Paving

Upper courses of wall (F91) retaining dumped hearth deposits Charcoal Conjectural Limit of excavation

Fig. 5.32 Section through north wall of church enclosure wall (F91) and deposits (F93a–i). Location of section on Fig. 5.29.

northern side had been rebuilt in this location. To the west of the excavated cutting there is a noticeable kink in the wall towards the north-west, at the entrance to the enclosure (Fig. 5.29; Pl. 5.24). The northwest entrance (F36) was a simple passage (W: 0.7m x L: 0.95m) with an upright stone marking each jamb (Fig. 5.31 (a), (f )). A threshold stone (W: 80mm) was set on edge between the outer jambs. The side walls were built with roughly square-shaped blocks of mica-schist. The northeast entrance was originally a simple passage (L: 1m) (Figs 5.1, 5.33a, F414). The space (7m x c. 1m) between the early enclosure wall at the north (F91) and the existing wall contained several layers of domestic refuse (Figs 5.29, 5.32, nos. 1–7; Pl. 5.25). These layers and lenses of redeposited ash and burnt debris (depth: 0.5m) originated from domestic hearths (Section 9.3, F93, F93a–j, F93m, F96). The layers of charcoal, charcoal ash and stones, mottled orange and black ash, grey ash, charred peat, orange ash, grey clay with stones, peat and charcoal were consistently deeper adjacent to the later enclosure wall. Away from the enclosure wall, the uppermost layers were sealed by the earlier wall (F91), which had partially collapsed onto them. Also, small water-rolled pebbles (D: <10mm), not found elsewhere in the excavation, were a notable feature in these layers. A combined charcoal sample from a number of these deposits (F93a–g, F96) produced a date of cal. AD 728–971 (UB-4522, 1171±41 BP). This was the earliest date range obtained from any sample belonging to the monastic period. The construction of the second wall at the northern side of the church enclosure pre-dates the dumping of the hearth debris. Soil analysis indicated that the

layers were composed predominantly of bands and fragments of charred fuel, which included peat and lacustrine mud (Section 9.4, F93a/’99C, F93c/’99F, F93i/’99Q). The firewood identified comprised mainly hazel, with lesser quantities of birch, heather, holly, hawthorn, willow/poplar and juniper (Section 9.3, F93a–j). Burnt and unburnt food waste was recovered from the layers, including one barley and two oat grains (Section 9.2, F93i). The largest collection of animal bone from Phase 1 was found in these deposits, the majority burnt and charred, rendering them unidentifiable (Section 9.1, F93a–j). Sheep/goat was the most frequent, followed by cattle and pig. There was also a small quantity of fish remains. Seal bones, of which there were three, were an unusual component and they represent the only evidence for the consumption of this mammal at the monastery.

Phase 2: mid-11th to late 12th/early 13th century A buttress (L: 2m, W: 1m, H: 1.2m) was built against the south-eastern corner of the church enclosure wall (Fig. 5.29, F817). It was composed of an outer face and a core of rubble, which were constructed on top of the original enclosure wall and its associated cobblestones (F877). This corner buttress dates to the same time as the construction of Cell B or soon afterwards (Section 5.3, Phase 2). One end of the buttress abutted the cell and the stone core of the enclosure wall, while at the other (southern) end an attempt was made to bond the masonry of the two walls together. The buttress and the enclosure wall behind were not in good condition, necessitating the dismantling of the buttress to carry out conservation work when rebuilding the south-eastern corner of the enclosure wall. Excavation

the excavations 91

in this location was confined to the area beneath the buttress corresponding with the central and outer face of the underlying enclosure wall that had collapsed. The Phase 1 north-east entrance to the church enclosure was rebuilt on a more elaborate scale as a longer, curved passageway roofed with lintels (L: 3.05m x W: 0.6–1.06m) that extended outside the original entrance (Fig. 5.1, Area 8, and Fig. 5.29, no. 9, F414/F8040). The realignment of the passageway indicates that it was designed to lead out to the beehive hut, Cell B, which was constructed around the same time (Section 5.3, Phase 2). The northern end of the east wall of the enclosure was reconstructed by widening it (to 2m max.) at the entrance and by realigning it towards the north-west (Figs 5.29, 5.34). The splayed and curved wall of the passage was built with coursed masonry, with a massive block forming the external jamb (Fig. 5.33 (b)). At the northern side of the entrance, a wall (L: 2m x W: 0.8m) was added on to the existing (Phase 1) masonry (Fig. 5.34). This wall was built to almost a semi-circle to match the curve on the other side of the entrance. The northern side of the passageway also consisted of coursed masonry (Fig. 5.33 (a)). The floor along the southern side consisted of the Phase 2 flagstones of the paved surface in the church enclosure (Fig. 5.34, no. 3, F29). Small, mostly rectangular-shaped slabs were laid on the northern side, seven of which were in situ (Fig. 5.34 (a), no. 2, F8050). The slabs partly overlapped one another, forming a rough and uneven surface that did not extend out to the side walls of the passage. Small patches of kaolinite clay (F8049) mixed with fine black silt lay between and, in places, partly covered the slabs. The clay, containing weathered granite, may have been imported to manufacture pottery items (Section 9.4, F8049/’02P). Three lintels covered the inner section of the passageway, two of which rested on the tops of the side walls and the third was displaced where the wall had collapsed (Figs 5.33, 5.34, no. 4, F414, F8040; Pl. 5.27). Grave 11, comprising a burial in a shallow clay-cut grave, was uncovered directly outside the north-east entrance to the church enclosure (Fig. 5.29, nos. 10, 11, F8026, F8034). The grave (L: 2m x W: 0.4–0.6m and depth: 0.14m) cut into the surface of the natural boulder clay and an area of mixed clay at the east (Fig. 5.34, nos. 8, 9, F8000, F8033). The mixed, dark brown, sandy clay, containing charcoal flecks, was redeposited earlier cultural material (Section 9.4, F8033/’02N). A small trench had been cut into these layers (Fig. 5.29, no. 14 and Fig. 5.33 (a), no. 10, F8035). It extended away from the grave towards the south, where it was sealed by the foundations of Cell B (Section 5.4, F8005). The severely decayed, damaged and incomplete skeletal remains of a middle-aged or older male were orientated 20° north of east–west (Fig. 5.29, no. 10 and Fig. 5.34 (a), no. 6 and Fig. 5.35 (a), no. 1, F8026; Pl. 5.28). The burial was dated to cal. AD 1021–1159 (OxA-

Pl. 5.27 Lintels above north-east passage (F414/F8040) through the church enclosure wall, looking south-west (G. Scally).

13665, 956±22 BP) and was covered by grey to brown, silty clay (depth: 0.2m), which was mixed with a dark brown, organic-based soil (Fig. 5.34 (b), no. 11 and Fig. 5.35 (b), no. 2, F8020). Oak was identified in the charcoal (Section 9.3, F8020). The fill, which covered the entire area of the grave, also filled the trench (F8035) to the south. On this side, a small hearth (D: 0.3m), comprising ash and burnt peat together with a cluster of small stones, overlaid the fill (Fig. 5.34 (b), no. 12 and Fig. 5.35 (b), no. 3, F8019). On the northern side, a more extensive spread of ash overlay the burial (Fig. 5.34 (b), no. 13 and Fig. 5.35 (b), no. 4, F8018). The charcoal sample contained oak (Section 9.3, F8018). This deposit was partly exposed as it was sealed beneath later, unexcavated features. A shallow lens of dry, schist clay covered the eastern end of the grave (Fig. 5.34 (b), no. 14, F8021). Compact, charcoal-flecked, silty clay (depth: 0.12m) covered the grave-fill, which in turn was sealed by a stonier, charcoal-flecked, grey clay (depth: 0.18m), the latter layer containing oak charcoal (Fig. 5.35 (c), no. 5; Section 9.3, F889). Both layers extended beyond the grave to the plinth-like foundations of Cell B to the south and the base of the enclosure wall to the west, while to the north they extended beyond the area of excavation and to the east they petered out close to the north-eastern corner of the cell (Fig. 5.29; Section 5.3, Phase 2). Grave 11, outside the north-east entrance to the enclosure, was subsequently restructured in a manner similar to the graves outside the east wall of the church (Section 5.2.3). The realigned Grave 11 was constructed on the surface of the grey clay (F889) and was stone-lined (Fig. 5.35 (c); Pl. 5.29). No human remains were contained within the restructured grave, which was constructed with sidestones set on edge. There were two slabs at the south, and one stone at the north was upright, the other two having fallen inwards. An upright, undecorated slab (H: 0.57m) marking the eastern end or foot of the grave had cut through the feet bones of the burial in the earth-cut grave beneath. This suggests that by the time the restructuring

92 high island: excavation of an early medieval monastery

South-west G

North G4

North-east G1

South G5

Faced masonry

2 1

(a)

(c)

North-east G2

1 Side wall of passage

South-west G3

2 Side wall of passage/church enclosure wall Lintels Upright blocking stone Paving Stone continuing Conjectural Limit of excavation 0

(F29)

(b)

Fig. 5.33 Entrance passage (F414/F8040) in north-east corner of church enclosure – internal elevations: (a) north wall; (b) south wall; (c) north–south elevation of blocking masonry and section through side walls. Location of elevations and section on Fig. 5.34 (b).

2 8

1 North-east passage (F414/F8040) through enclosure wall

2 Paving (F8050) visible beneath in situ masonry

3 Paving (F29) 4 Lintels over passage (F414/F8040)

5 North-east corner of Church 6 Burial (F8026) 5

7 Grave cut (F8034) 8 Boulder clay (F8000)

(a)

9 Mixed clay (F8033) 10 Trench (F8035) 13

11 Fill (F8020) over burial (F8026)

12 Burnt debris (F8019)

G1 11

15 4

14 Dry schist clay (F8021)

15 Lintels of blocked passage

G3 3

13 Burnt debris (F8018)

Lintels G5

Limit of excavation

(b)

Fig. 5.34 Entrance passage (F414/F8040) in north-east corner of church enclosure: (a) Grave 11 with burial (F8026) in situ; (b) blocking masonry.

the excavations â&#x20AC;&#x192; 93

took place, the location of the burial had been forgotten. Several layers of fill were identified within the restructured grave. The lowermost was grey, silty clay (F874) with a scatter of quartz pebbles (Fig. 5.35 (d), no. 6). The clay contained decayed stone, charred oak and burnt faunal bones (Sections 9.1, 9.3, F874). It was covered by a thin lens (depth: 20mm) of dark silt and clay (F876). A flagstone of mica-schist (0.5m x 04m x T: 4mm) covered the eastern portion of the grave, while stones and quartz pebbles set within a grey/brown silt and clay (F872) lay along the remaining part (Fig. 5.35 (e), no. 7, F872; Pl. 5.29). Above, three separate deposits were identified. A mid-grey, silty clay (F870) covered the central and western portion of the grave, while the eastern end was filled with schist soil and ash (F869) and dry, friable clay (F868) containing twenty quartz pebbles (Fig. 5.35 (f), no. 8). The redeposited material had been deliberately laid down as the quartz pebbles and clay at the eastern end were separated from the other fill layers by a slender, rectangular-shaped stone placed across the width of the grave. The surface of the grave was covered by a layer of light grey schist fragments (Fig. 5.35 (g), no. 9, F844). Oak charcoal and a variety of faunal remains, including those of cattle, sheep/goat and bird, were in the schist layer (Sections 9.1, 9.3, F844). The north-east entrance to the church enclosure was blocked up, a process that was possibly carried out around

Pl. 5.28â&#x20AC;&#x201A; Grave 11 with burial (F8026) in situ at entrance to northeast passage (F414/F8040) through the church enclosure wall, looking west (G. Scally).

the same time as the grave was restructured because damage was caused to the burial in Grave 11. Mid-way along the length of the entrance, a substantial triangular-shaped stone was set into the ground and wedged tightly between the side walls, thereby blocking access (Figs 5.33 (c), 5.34 (b)). Small stones were wedged between it and the flagstones of the paved surface (F29). Five lintels, two of which were broken, were fitted into the remaining area in the north-east of the entrance (Fig. 5.34 (b), no. 15; Pl. 5.30). The top of the lintels was at the same level as the top of the upright blocking stone. The lintels were supported on the northern side by large, roughly hewn boulders that appeared to have belonged to the core of the wall, while on the southern side they were inserted into the end of the enclosure wall. The outermost lintel rested on an upright stone at the opening to the entrance and was supported on smaller stones wedged at its base (Pl. 5.30). These stones crushed the skull of the burial in Grave 11 (Pl. 5.28).

Phase 3: late 12th/early 13th to mid-15th century The Phase 3 surface within the north-west entrance was a continuation of the soil and stone surface identified in the rest of the enclosure (Fig. 5.15, no. 3, F4). A low wall (F9) was built on the stone surface against the inner face of the western wall of the enclosure. This wall (L: 5.4m x W: 1.1m, H: 0.55m) was poorly constructed, with a crude

Pl. 5.29â&#x20AC;&#x201A; The restructured stone-lined Grave 11, looking west (G. Scally).

94 high island: excavation of an early medieval monastery

(a)

3 (b)

attempt at a face on its eastern side. It was filled with a core of soil and small stones, including a number of quartz pebbles. The wall may have been built as a buttress in an attempt to stabilise or prevent inward collapse of the west wall of the enclosure. In the north-east entrance, stone and clay (depth: 0.3m) sealed the paved surface (F8050). These deposits must have accumulated after the Phase 2 lintels had been placed in position above the outer, blocked part of the passage. Outside the entrance, a deposit of silty clay and fragmented schist containing a small quantity of faunal remains, mainly of sheep/goat, covered Grave 11 (Section 9.1, F847). Above, there was a shallow layer of fine, possibly wind-blown soil in which there was a significant number of cattle bones (Section 9.1, F837). Within the soil, a localised deposit of burnt debris, possibly a hearth, was exposed (0.6m x 0.4m and T: 20mm), which extended to the north beyond the area of the grave to overlie a low wall on this side (F848) (Section 5.4.1). Faunal remains of cattle, sheep/goat, fish and bird were recovered from the burnt debris (Section 9.1, F849).

Phase 4: mid-15th to late 20th century

(c)

(d)

(e)

A deposit of peat was uncovered on top of the Phase 3 levels outside the north-east entrance. The peat contained charcoal and a large quantity of sheep/goat and cattle bones (Sections 9.1, 9.3, F829). An amber-coloured glass bead of medieval date was found in this deposit (Section 6.9, 829:2). Rubble was interspersed with more peat, containing quantities of faunal remains (Section 9.1, F805, F806). Dry, schist clay and stone (F2) and masonry rubble (F1) had accumulated over the boundary wall and its entrances from the 15th century onwards, the bulk of which must have derived from the collapse of the enclosure wall.

(f)

Pl. 5.30 Lintels blocking north-east passage (F414/F8040) through the church enclosure wall, looking west (G. Scally).

(g)

1 Burial (F8026)

6 Fill (F874)

Upright stones

2 Fill (F8020)

7 Grave-fill (F872)

Charcoal

3 Burnt debris (F8019)

8 Grave-fills (F870, F869, F868) and quartz pebbles

Quartz

4 Burnt debris (F8018) 5 Deposit (F889)

9 Schist (F844)

Conjectural Limit of excavation

Fig. 5.35 Grave 11 beneath restructured stone-lined grave (F856): (a) burial (F8026) in situ; (b) fill layers over burial; (c) stone lining (F856); (d) grave-fill (F874); (e) grave-fill (F872); (f) upper layers of grave-fill (F870, F869, F868); (g) surface of Grave 11.

the excavations 95

5.3 Cell B

occasional outcrops of bedrock apparent within it (Section 9.4, F336/’00AM; Appendix B). The surface of the clay was truncated by a narrow gully (L: 3.6m x W: 0.2– 0.35m and depth: 0.1m), which was orientated for most of its length north-west–south-east (Fig. 5.36, no. 2 and Fig. 5.39 (b), no. 22, F335). It continued in an easterly direction under the cell wall. The gully was filled with moist, grey clay and dark silt (Fig. 5.39 (b), no. 23, F344). Two stake-holes cut into the clay beside the gully were filled with moist clay and silt similar to the gully fill and also charcoal fragments (Fig. 5.36, no. 6, F346). Areas of small, flat stones and, in places, compact, almost impenetrable deposits of splintered and fragmented schist covered the clay. These deposits were sealed by a shallow clay bedding beneath a surface consisting of compacted stones and mica-schist clay (Fig. 5.36, nos. 4, 5, and Fig. 5.39 (b), nos. 17–19, F339, F340, F353). A smithing hearth pit (depth: 0.3m) was dug adjacent to the gully, which cut into the disturbed and natural boulder clays (Fig. 5.36, no. 1, F352). The complete pit was not excavated because it continued to the east, beneath the wall of Cell B. The pit had sharply sloping sides and a flat base. In the centre, slender

The beehive hut, Cell B, is located in the eastern part of the monastic enclosure (Fig. 5.1, Area 3). With the exception of a small area (1m2) in the north-eastern corner, the entire interior of the cell was excavated to undisturbed boulder clay. The entrance was investigated to the level of the paved surface, as was the area outside the cell.

Phase 1: 8th to mid-11th century Metal-working activities Considerable evidence for metal-working, consisting of a smithing hearth, a gully and artefacts relating to its production, were uncovered beneath Cell B (Fig. 5.1, Area 3 and Fig. 5.36). Compact boulder clay was exposed within a small test trench explored in the centre of the cell (Fig. 5.39b, nos. 25, 26, F344). Within the trench and extending over the cell interior, a layer of clay mixed with silt and organo-mineral and charred matter, containing grit and small stones, was revealed (Fig. 5.36, no. 3 and Fig. 5.39 (b), no. 24, F336; Section 9.4, F336/’98AS). This layer represents the disturbed surface of the boulder clay, with

1 Smithing hearth pit (F352) 2 Gully (F355) 3 Clay (F336) 4 Stony surface (F353)

2 6

5 Stony surface (F339) 6 Stake-holes (F346)

4 East wall of church enclosure

Later cell 1

Tie stones bonding cell to enclosure Obscured

Conjectural 4

Fig. 5.36 Smithing hearth-pit (F352) and associated deposits beneath Cell B, Phase 1.

Limit of excavation

96 high island: excavation of an early medieval monastery

stones were set upright that may have divided the pit into compartments. To the south of these stones, a shallow, semi-circular depression (D: 0.12m x depth: 30mm) had been cut into the base of the pit. To the north, there was a stake-hole (D: 4mm x depth: 4mm). Neither the base nor the sides of the pit showed any evidence for burning or scorching. The pit was filled almost exclusively with wood charcoal, most of which was oak (Section 9.3, F347, F351). A date of cal. AD 870–1030 AD (OxA-8917, 1100±40 BP) was obtained from a sample of hazel charcoal. Analysis of the fill produced evidence to indicate that the most likely function of the pit was that of a smithing hearth for ironworking and probably also for non-ferrous metals (Section 6.5, F347, F351). A fragment of a clay crucible (347:1) and several droplets of iron and pieces of slag were recovered mixed with the charcoal. The presence of the crucible indicates that, most probably, precious metals were being worked, while the iron droplets are typical residue materials from welding pieces of iron together (J. Bailey, pers. comm.). A smithing hearth cake (8099:1) and the tip of a tuyère (852:1), which were associated with the smithing hearth, were recovered in the monastic enclosure in the silt fill of the rock-cut trough (Section 5.4.1, Phase 3).

Phase 2: mid-11th to late 12th/early 13th century Cell B, the beehive hut, or clochán, in the monastic enclosure is an imposing drystone structure that was built against the exterior face of the east wall of the church enclosure (Figs 5.37, 5.38 (a)–(c); Pl. 5.31). Internally, it is

Pl. 5.31 Cell B after excavation, looking west (G. Scally).

almost square (2.7–2.8m north–south x 2.8m east–west) with angular corners. From its (paved) floor to roof, it is 2.8m in height. The west or rear wall is the narrowest (0.6m), but it was built against the enclosure wall (combined W: 2m), and the other three range in width (W: 1.5–1.7m). The exterior of the cell is roughly rectangularshaped and its corners are curved almost semi-circular, giving the impression of a rounded structure rather than the straight-sided building that it is (Pl. 5.31). The cell is entered via a low, splayed doorway (W: 1.58m, H: 1.2m) in the centre of the east wall (Fig. 5.37; Pl. 5.31). The floor was paved with mica-schist flags, which was found to be a continuation of the surface outside the cell within the monastic enclosure (Fig. 5.37, no. 2; Section 5.4.1, F835). The threshold stone, wedged tight between the jambs, stood proud of the floor (Fig. 5.38 (a), (d), (e), no. 1). The doorway jambs were substantial (Fig. 5.38 (a), (d), (e), nos. 5, 6). The northern jamb has been tentatively identified as a possible reused cross-slab (Section 7.1, Cross-slab 65). A small, rectangular-shaped slot on the southern jamb was probably a bolt-hole for a door fitting (Fig. 5.38 (e), no. 6). The doorway was roofed with lintels, three of which had survived in situ (Fig. 5.37; Pl. 5.31). The innermost lintel was a substantial slab that formed part of the masonry of the cell wall on both sides (Fig. 5.38 (a), no. 2). Its underside was carved with a cross (Section 7.1, Cross-slab 18). The particular motif, coupled with its large size, suggests that originally it may have been an upright altar stone (Section 7).

the excavations 97

N 6 5

1 Cell B

2 Paved surface (F835) surrounding Cell B

H4 1

3 Cobbled surface (F877)

Cell entrance

4 Kerb (F859)

5 Plinth-like foundations (F8005) at base of Cell B

6 Silty clay (F899)

7 Buttress (F817) 8 East wall of church enclosure H1

Granite Tie stones

Lintel Obscured Conjectural Limit of excavation

4 2

Fig. 5.37 Plan of Cell B. Elevations, see Fig. 5.38 (a)–(e).

The internal walls of the cell were corbelled above 0.8m (Fig. 5.40). The basal courses (H: 1.2m) were composed predominantly of large, rectangular-shaped stones. Above this height, the masonry consisted of courses of smaller and thinner stones (Fig. 5.38 (a)–(c)). Spalls were frequently used and with considerable care, particularly in the upper courses of the walls. The overall result was that of smooth walls with gently curving faces on all sides. Masons’ tool marks were preserved on the wall above the door lintel. Recesses, or cupboards, were tentatively identified in the west and north walls. The recess in the west wall was the more convincing of the two as it was formed by shaping one of the internal facing stones (Fig. 5.38 (c), no. 4). A rare silver coin of mid-11th-century type was recovered from rubble behind the recess, suggesting that it may have been hidden there (Sections 6.6, 6.7, 320:1). However, the masonry behind the recess had partly collapsed, and it is possible that the coin may have fallen in from outside the cell. The other recess in the north-eastern corner of the cell had a gable-shaped opening (W: 0.36m x H: 0.42m, depth: 0.36m). Externally, the lower part of the walls was intact and, like the internal masonry, was corbelled above 0.8m (Fig. 5.40; Pl. 5.31). Five projecting corbels were regularly spaced

around the exterior. Their most likely use was for securing ropes which would have secured sods or other materials to the roof. They may also have functioned as a rudimentary scaffold, which, after the construction of the cell had been completed, would have provided access to the roof and upper part of the cell walls for repairs or other work. Above the well-preserved lower part of the walls (H: 1.2m), only the stones on the inner face of the cell survived. This masonry was in a fragile condition and small voids were apparent at each side and at the apex. It is likely that the outer walls were deliberately stripped and as a result the soil and rubble stone core eroded away. This destruction can be associated with the miners in the 1820s. Prior to their arrival, the cell was almost intact and 30 years later, it was recorded as partially destroyed (Sections 3.2–3.4; Pls 3.1, 3.2). The external foundations on the northern and southern sides of Cell B were unusual. Five boulders formed the lowermost course at the south, which were completely different in scale and form from the masonry above (Fig. 5.37; Pl. 5.31). The cell wall above rested on the two boulders nearest the church enclosure wall, and the remaining boulders projected from beneath the wall. Two of the boulders were on the cobbled surface that had been laid

(d)

West H4

East H5

(e)

South H1

East H6

West H7

(b)

East H1

(c)

South H2

Granite

Limit of excavation

Conjectural

Obscured

Quartz

Paving

Upright stones

Lintel stone

Cell walls

7 Paving (F835)

6 Jamb stone with bolt-hole

5 Jamb stone; potential cross-slab (see CS 65)

4 Recess

3 Kerb around subsidiary hearth (F322)

2 Cross-slab reused as lintel (CS 18)

1 Threshold stone at cell entrance

West H2

North H3

Fig. 5.38â&#x20AC;&#x201A; Cell B internal elevations: (a) east; (b) south; (c) west; (d) south side of entrance; (e) north side of entrance. In elevation (a), courses above threshold stone and section through cell are after White Marshall and Rourke 2000, Fig. 84c, with amendments. Location of elevations on Fig. 5.37.

(a)

North H

98â&#x20AC;&#x192; high island: excavation of an early medieval monastery

the excavations 99

on the boulder clay (Fig. 5.37, no. 3, F877; Section 5.2.4). The masonry above the centre boulder was balanced on top. The adjacent boulder (to the east) was located beyond the cell wall and the gap was filled by three small stones, which appeared to support the entire weight of the overlying masonry. Despite the seemingly precarious way in which the foundation was designed, no sign of collapse or subsidence was evident in the masonry above. The foundations of the northern side of the cell were of a different form. A single course of substantial boulders projected from underneath the base of the wall (Fig. 5.37, no. 5, F8005; Pl. 5.32). These boulders formed an irregular plinth, the upper surface of which was level at a height of 0.3m above the boulder clay (Fig. 5.40, nos. 12, 16, F8000, F8005). Two deposits of soil were laid down at the same time as the cell was constructed (Section 5.2.4, F889, F899). The compact, charcoal-flecked silty clay was the lowermost deposit (Fig. 5.37, no. 6, F899). The other deposit, also of silty clay containing small stones and charcoal flecks, was banked up against the cell where it was deepest (0.7m) and it sloped away from the wall and then levelled out. This clay was sealed by a layer of fragmented stone (depth: 0.2m) that spilled out from the cell wall and formed a distinct interface between the plinth-like foundations below and the coursed masonry above (Fig. 5.40, no. 11, F885). Its function appeared two-fold. Within the cell wall, it allowed movement and settlement of the overlying masonry to occur, whilst outside it acted as a grout upon which clay, used as a weatherproof cladding around the cell, was laid. Clay packing, which was almost concrete-like in consistency, surrounded the foundations of the cell on the northern and southern sides (Fig. 5.40, nos. 4, 10, F803n, F803s). On the northern side, as well as covering the fragmented stone and gravel, the clay also sealed the stones of the plinth-like foundations. On the southern side, the clay, which also sloped away from the cell wall and covered the large boulders of the foundations, extended as far as the paved surface around the cell and it abutted the buttress at the church enclosure wall (Fig. 5.29, F817, and Fig. 5.40, no. 3, F835; Pl. 5.33; Section 5.2.4). The clay packing contained small stones and charcoal of hazel, oak, birch, willow/poplar and heather, and faunal remains of cattle, sheep/goat, fish and bird (Sections 9.1, 9.3, F803). This clay has been interpreted as redeposited, organically rich subsoil that was probably placed on the foundations and basal wall courses of the cell to act as a weatherproof layer (Section 9.4, F803/’00AE; Appendix B). Charcoal extracted from the clay was dated to 1017–1159 AD (OxA14042, 968±27 BP). On the northern side, a fine-grained, organic-based soil with an orange hue, which had slipped out of its original position, covered the redeposited subsoil farthest away from the cell wall. Analysis has shown this deposit was a turf- or topsoil-derived clay (Section 9.4,

F850/’011). It may represent the decayed remains of a sod cover laid around the base of the cell as an outer layer of wind and waterproof protection. Inside the cell, a mixed layer of clay, ash, silt and charcoal (depth: 0.18m) covered the central and southern parts above the Phase 1 metal-working activities (Fig. 5.39 (b), no. 16b, F338). Darker, gritty clay with less ash and silt covered the western portion (Fig. 5.39 (a), no. 2, Fig. 5.39 (b), no. 16a, F334). A rubbing stone was found in this gritty clay (Section 6.2, 334:1). Twelve flat, mica-schist slabs overlay these deposits and formed the floor of a substantial hearth in the centre and east of the cell (Fig. 5.39 (a), no. 1, Fig. 5.39 (b), no. 14, Fig. 5.40, no. 9, F326). A slender stone set on edge on the western side may be the remains of a kerbed edge to the hearth. A broken cross-slab was reused in the floor (Fig. 5.39 (a), no. 3; Section 7.1, Cross-slab 30). Most of the slabs were laid directly on the clay layers, but a large slab at the entrance was resting on a bedding of small stones (Fig. 5.39 (b), no. 15, F350). Scorch marks were evident on the slabs in the centre of the cell. Lenses of clay, each no more than a few millimetres deep, overlay several of the slabs (Fig. 5.39 (a), Fig. 5.39 (b), no. 13, F333). These lenses, which had a laminated appearance, had bands of

Pl. 5.32 The plinth-like foundations (F8005) at the northern side of Cell B (G. Scally).

Pl. 5.33 The southern foundations of Cell B covered with the clay packing (F803 S) (G. Scally).

100 high island: excavation of an early medieval monastery

stone-lined pit (D: 1m, depth: 80mm) was filled with ash and other burnt debris (Fig. 5.39 (a), nos. 4, 5, F335, F341). The charcoal in this hearth was mostly of heather and oak, with small quantities of willow/poplar, birch and hazel (Section 9.3, F335). Burnt cereal remains were of barley, and fish were dominant in the faunal analysis (Sections 9.1, 9.2, F335). The ash spilled over the sides of the pit and covered the contemporary ground surface (F338). Analysis of the ash indicated that it probably represented domestic waste resulting from one of the last clearances from the hearth (Section 9.3). In the north-eastern corner of the cell, a smaller hearth consisting of one large slab was at a slightly higher level than the main hearth in the centre of the cell (Fig. 5.39 (a), no. 6, F322). Six slender stones set on edge surrounded the slab in the corner. Scorch marks were visible on the surface of the slab and a small patch of burnt debris lay in the centre.

cleaner, lighter-coloured clay and streaks of ash running through them and this was most evident on slabs where no scorch marks were seen. This would suggest that the clay had been deliberately laid over the hearth floor slabs to protect them from the heat. A mound of burnt debris (D: 2–2.5m, H: 0.2m) covered the hearth floor (Fig. 5.39 (b), no. 12, F319). The debris consisted of alternating layers of ash and peat charcoal with fragments of shell and bone scattered throughout, suggesting that food was prepared and probably consumed at the hearth (Section 9.1, F319). A single cereal grain fragment was identified (Section 9.2, F319). The fuel remains were almost entirely the residue from burning heather (Section 9.3, F319). A concentration of food waste, mostly sheep/goat bones and limpet shells, was mixed with the debris on the eastern side of the paved hearth (Fig. 5.39 (b), no. 11; Section 9.1, F321). In the south-east of the cell and adjacent to the hearth, a

Fig. 5.39 Cell B: (a) plan of interior showing hearth-floors (F326, F322); (b) section through in situ deposits.

N 6 1 Hearth floor (F326) 2 Gritty clay (F334) 3 Cross-slab (CS 30) reused as paving stone in hearth floor

4 Pit (F341)

5 Ash pit fill (F335)

6 Subsidiary hearth (F322) 3

Scorch marks Clay (F333)

Upright stones

(a)

West J

East J1

12 3

3 6

4 8

2 8

14 16b

16a 24

(b)

21 25

15 13 18

Threshold stone 0

1 Peaty loam (F310)

13 Clay (F333)

24 Mixed clay (F336)

2 Grey/brown soil (F311)

14 Hearth floor (F326)

25 Test trench

15 Stone bedding (F350)

26 Boulder clay (F344)

3 Stone and soil (F312) 4 Stone (F315)

16a Gritty clay (F334)

5 Silt and clay (F314)

16b Clay, ash & silt (F338)

27 Paving (F835) Paving

6 Clay and ash (F313)

17 Stone and schist surface (F339)

Hearth debris

7 Stone (F317)

18 Stone and schist surface (F353)

Charcoal

8 Silty clay (F318)

19 Clay bedding (F340)

Bedrock

9 Stone (F325)

20 Compact schist (F343)

Conjectural

10 Silty clay (F324)

21 Stones (F356)

11 Shell and bone (F321)

22 Gully (F355)

12 Hearth debris (F319)

23 Gully fill (F344)

Limit of excavation RD

Rabbit disturbance

the excavations 101

Cell B

South K

North K1

10 11 2 1

6 3

14 15

4 5

12 7

1 Paving (F833a)

6 Kerb (F858)

11 Small stones (F885)

2 Wall (F813)

7 Boulder clay (F8002)

12 Plinth-like foundations (F8005)

Paving

3 Paving (F835)

8 Cell B

13 Wall (F448)

Conjectural

4 Clay packing (F803 S)

9 Hearth floor (F326)

14 Paving (F467)

Obscured

5 Cobbles (F877)

10 Clay packing (F803 N)

15 Paving (F422)

Limit of excavation

16 Boulder clay (F8000)

Fig. 5.40 North–south section through Cell B and surrounding features. Location of section on Fig. 5.41.

Phase 3: late 12th/early 13th to mid-15th century Silty clay deposits mixed with layers of stone (depth: 0.4m) covered the hearth debris and surrounding features (Fig. 5.39 (b), nos. 7–10, F317, F318, F324, F325). The stones were of varying sizes and were quite concentrated, leaving no doubt that they had been introduced to the cell as their quantity far exceeded that which could have resulted from the collapse of the cell walls. The cell was probably a place where visitors or other short-term residents of the island sheltered. It seems likely that it was these people who brought the stones into the cell to provide temporary surfaces on which to make small open fires for cooking. Several lenses of ash, together with food debris, mostly animal bone, were found scattered through the deposits. Relatively large amounts of cattle and sheep/goat remains were recovered in the clay layers (Section 9.1, F318, F324). Phase 4: mid-15th to late 20th century Accumulations of clay and ash sealed by stone, silty clay and further stone, in places three courses deep, covered the cell interior (Fig. 5.39 (b), nos. 3–6, F312–F315). Some of the stones were substantial in size and may have originated from stockpiling the stone stripped from the cell exterior. Most of the food remains from these levels were in the lowermost layer of ash and clay (Section 9.1, F313). The overlying deposit of light grey/brown soil (depth: 0.3m) contained sheep/goat bones (Fig. 5.39 (b), no. 2; Section 9.1, F311). A fragment of a green glass bottle probably relates to the mining activity in the 1820s (Sections 3.3, 6.10, 311:1). Several large and some smaller stones lay on top, which could have fallen from the apex of the roof. The

stones were embedded into dark brown peaty loam, the uppermost layer within the cell (Fig. 5.39 (b), no. 1, F310).

5.4 The monastic enclosure The monastic enclosure incorporated an area approximately 27m north–south x 32m east–west, of which less than half was excavated (Fig. 5.1, Areas 1, 4, 6, 8). An L-shaped trench (20m x 2m) was opened in the south-west in an attempt to uncover the stratigraphic relationship between the church enclosure and the monastic enclosure (Fig. 5.1, Area 6). Once excavation began, however, it became clear that due to the restricted size of the trench, the wealth of features exposed within this area could not be interpreted satisfactorily. After recording the exposed features, the trench was back-filled. With the exception of this trench, which was opened primarily for research purposes, excavation of the other areas in the enclosure was undertaken largely to enable the structures that lay within to be stabilised and conserved.

5.4.1 Features within the monastic enclosure Phase 1: 8th to mid-11th century The rock-cut well The rock-cut well was located at the base of the sloping ground in the eastern part of the monastic enclosure (Fig. 5.41, no. 1, F8068; Pls 5.34, 5.35). The bedrock was cut to form a cavity (0.9m x 0.8m and depth: 0.68m), within which the well was built (Fig. 5.42, no. 1, F8068). The

102 high island: excavation of an early medieval monastery

entrance to the well (W: 0.42m x H: 0.33m) and the side walls were stone-lined, while the floor and back wall were formed by the bedrock. The entrance was roofed with lintels, the largest at the opening (L: 1.04m) (Fig. 5.42). Beyond the opening, a sharp slope downward in the rockcut floor created a relatively substantial holding area for the water (Fig. 5.42b). A threshold, or blocking stone, set deep in the ground at the mouth of the well stood 80mm above the adjacent floor level. This stone prevented water from flowing directly outwards, but it allowed excess water to trickle past on both sides. The remains of a retaining wall, probably constructed at the same time as the well, were exposed partly embedded into the steeply sloping ground rising towards the eastern boundary of the enclosure (Fig. 5.41, no. 3, F863; Pl. 5.35). The northern end of the wall was set within the cut in the bedrock and it overlay the roof of the well (Fig. 5.42 (a), no. 3, F863). The wall (L: 1.7m) extended in a southerly direction from the well. It was a single stone in width and was composed of up to four courses of roughly faced stone, 0.5m in height. At the southern end, the terminus of the wall was unclear as it merged into the stone rubble and compacted clay and scree that had accumulated along the natural schist and rock-face of the sloping ground. In places, the scree was of almost concrete-like consistency due to its accumulation over centuries of seepage and leaching from water running down the slope. A stone path, traceable for just over 3m, led southwestwards from the well (Fig. 5.41, no. 2, F8028). Almost the entire length of the southern side wall of the path had been removed by the later water management features (Phase 2), but five slabs of mica-schist remained. The two largest were lying outside the entrance to the well on a shallow accumulation of natural clay and gravel that covered the bedrock (Fig. 5.42 (b), no. 4, F8028; Pl. 5.36). Two stones set on edge on the southern side were probably the remains of a side wall, and further along the path the other three paving slabs had slender, upright slabs on the northern side that were set into the bank of natural clay and gravel into which the path had been cut. Where the path was ultimately leading was unclear. It was earlier than the other features in the area, yet it appeared to continue in use after these had been built. Its south-westerly direction suggests that it may have led to the Phase 1 metal-working area pre-dating Cell B (Section 5.3). If this was the case, the path may have been constructed to provide access to a convenient water supply, which was essential for the metal-working processes.

Phase 2: mid-11th to late 12th/early 13th century The water management system A rock-cut water trough with masonry side walls and

Pl. 5.34 Rock-cut well (F8068), looking east (G. Scally).

Pl. 5.35 Rock-cut well (F8068) and retaining wall (F863) cut into steeply sloping ground on the east side of monastery, looking east (G. Scally).

roofed with lintels lay at the base of the sloping ground in the east of the monastic enclosure (Fig. 5.41, no. 4, F845). The trough was just 2m from the entrance to Cell B and it probably dates to the same period of construction, in the 11th to mid-12th century. Approximately two-thirds of the trough (L: 3.9m x W: 1.05m) lay within the excavation area and the southernmost end lay outside. Furthermore, its northern end was partly obscured by the later drain and leacht, making its original extent difficult to establish (Fig. 5.43 (b), nos. 2, 4, 5, F845, F887, F8007). The cut for the trough at surface level was splayed and the bedrock was cut to make ledges on which the side walls were built (Fig. 5.44, no. 6, F853). The lower side walls of the trough were formed by the bedrock itself, which also delineated the

the excavations 103

Y G G

19 17

P 18

P1 L

S1 L2

L3 Y1

L1 G

G 9 Church

M1 5

*See detail Fig. 5.43 6

1 Rock-cut well (F8068)

12 Paving (F833a)

Later structure

2 Path (F8028)

13 Leacht (F887)

Lintels

3 Retaining wall (F863)

14 Leacht (F843)

4 Rock-cut water trough (F845)

15 Wall (F813)

5 Soakaway (F846)

16 Paved path (F875)

Upright stones

6 Wall-face (F832)

17 Paving (F422)

Annulus

7 Water collection & drainage channel (F8025)

18 Wall (F848)

Conjectural

8 Covered drain (F8007)

19 Cross base (F415)

Obscured

9 Paving (F835)

20 Paving (F615a)

Limit of excavation

10 Paving (F80)

21 Buttress (F79)

Path (F8028) G

Granite

11 Paving (F833)

Fig. 5.41 Plan of the monastery showing locations of features exposed within the monastic enclosure. Section K–K1, see Fig. 5.40; L–L3, see Fig. 5.42; M–M1, see Fig. 5.44; O–O1, P–P1, R–R1, S–S1, see Fig. 5.48; Y–Y1, see Fig. 5.64.

104 high island: excavation of an early medieval monastery

North L

South L1

East L2

West L3

(a)

(b)

1 Well (F8068)

Paving

2 Threshold stone

Bedrock

3 Retaining wall (F863)

Conjectural

4 Path (F8028)

Obscured

Lintels

Limit of excavation

Fig. 5.42 Rock-cut well (F8068) and retaining wall (F853): (a) external elevation; (b) elevation of south side of well interior. Location of elevation on Fig. 5.41.

largest portion of the holding area for the water (Fig. 5.44, no. 8; Pl. 5.37). The ledge on the western side was level and accommodated up to five courses of stone, which were best preserved towards the southern end. A kerb of slender slabs

Pl. 5.36 Stone path (F8028, under ranging rod) leading southwest from the rock-cut well (F8068), looking south (G. Scally).

on edge delineated this side of the trough as well as separating it from the paved surface outside Cell B (Fig. 5.44, no. 9 and Fig. 5.46, no. 9, F835). At the northern end, the kerb was obscured by the leacht built over it (Fig. 5.43 (b), nos. 2, 4, F845, F887). The ledge on the eastern side of the trough sloped at an angle and did not appear to have been a suitable support for the side wall. Towards the south, where the trough narrowed, the masonry was preserved (Fig. 5.44). Approximately mid-way along the length of the trough, a rectangular-shaped stone, extending the full width of the trough, was wedged in position, its western end fitting into a small but deliberately formed slot cut into the bedrock ledge. The other end was held tight by small stones placed between it and the bedrock. It is suggested that this stone functioned as a ‘sluice’ that would have been used to control the flow of water within the trough (Figs 5.43, 5.45, no. 3; Pl. 5.37). North of the ‘sluice’, the trough floor sloped gently downwards (depth: 0.4m). South of the ‘sluice’, a steep step cut in the bedrock floor increased the depth of the trough to 0.9m. At the bottom of the slope, the floor level rose again, creating a dip where silt and gravel could settle at the base of the trough. The two lintels which remained in situ were south of the ‘sluice’, indicating that the trough had, at least in part, been roofed (Fig. 5.45, no. 2, F845). At the west, they were supported on the side wall on which a small deposit of clay had been laid and into which they had been embedded. The fine-grained clay was a distinctive light grey colour, which apparently was used as a type of mortar bonding (F860). At the east, the lintels were wedged into the scree that had slumped over the trough edge (Fig. 5.44, no. 7, F8065). Further south, outside the excavation area, a void

the excavations 105

R 1

R1 4

2 3

N1 (a)

1 Channel (F8052)

3 Sluice stone

2 Trough (F845)

Leacht (F887)

5 Covered drain (F8007) Later leacht

(b)

Upright stones

Lintels

Conjectural

Fig. 5.43 Rock-cut trough (F845): (a) channel (F8052); (b) covered drain (F8007) and leacht (F887). Location of section N–N1 on Fig. 5.45; R–R1 on Fig. 5.49.

Pl. 5.37 Rock-cut water trough (F845) with sluice stone in situ, looking south (G. Scally).

in the section face created by the trough allowed it to be examined here (Pl. 5.37). The intact trough in this location had masonry side and end walls consisting of large blocks of stone and it was roofed with lintels. This means that the trough south of the ‘sluice’ was originally roofed. The function of the trough was to collect, hold and regulate the flow of water and therefore it would have had to have been accessible, at least in part, probably to the north of the ‘sluice’. Whether this section originally had some lintels covering it whilst also being accessible, or whether it was entirely unroofed, remains unknown. Due to the fact that the southern end of the trough lay outside the excavation area, it was not possible to establish how the water exited. In an attempt to identify how this occurred, it was emptied of water several times, yet each time it filled almost immediately. Subsequently, a probe was inserted through the opening of the trough (at the excavation section face) to try and identify the location of any openings in the masonry rear wall or western side walls. An opening was found in the south-western corner, leading to a subterranean channel. The channel (tracked for 2m in length by the probe) continued in this direction, but its entire route could not be traced. It was clear, however, that water from the trough would have exited via this opening and that it would have been taken away from

106 high island: excavation of an early medieval monastery

East M

West M1

1 2 3

4 5 6

9 8

1 Grass

7 Scree (F8065)

2 Sod (F800)

8 Opening of trough visible in excavation baulk

3 Peat and scree (F805)

9 Kerb defining trough edge

4 Rubble (F840)

10 Monastic enclosure wall

5 Shells

Bedrock Obscured Conjectural

Lintels

6 Cut (F853)

Paving (F835)

Trough (F845)

Limit of excavation

Fig. 5.44 East–west section through rock-cut trough (F845). Location of section M–M1 on Fig. 5.41.

North N

South N1

1 Lintel over drain (F8007)

Lintels

2 Lintels over trough (F845)

Bedrock

3 Sluice stone

Conjectural

4 Conjectural layout of unexcavated part of trough visible through opening in excavation baulk

Fig. 5.45 North–south section through rock-cut trough (F845). Location of section N–N1 on Fig. 5.43 (b).

the area in the subterranean channel underlying the paved surface (F833). The channel possibly continued around the structures at the main south-east entrance to the monastery to empty into the lake (Figs 5.1, 5.41). The original layout of the northern end of the trough was also difficult to establish as it was obscured by the

later leacht and covered drain (Fig. 5.43 (b), nos. 4, 5, F887, F8007). From what could be established, the western side wall of the trough was formed by coursed masonry in the same manner as that revealed towards the other end of the same wall. The kerbstone demarcating the trough edge was preserved in situ within the leacht (Fig. 5.46, nos. 7, 8, F845, F887; Pl. 5.38). To the north, beyond the kerbstone, a vertical cut (depth: 0.12m) was identified in the natural clay, suggesting that the end of the trough extended further north than its excavated extent. At the east, no comparable side wall to the trough was found. Instead, three slender stones on edge formed a narrow channel (L: 1.34m x W: 0.2m) immediately above the holding area of the rock-cut trough (Fig. 5.43 (a), no. 1, Fig. 5.46, no. 4, F8052). These stones were set into natural gravel, which was probably slippage from the steeply sloping ground to the east. Several lenses of various types of natural scree, gravel and small stones were identified in the vicinity (Fig. 5.46, no. 6, F8070, F8001). How water originally entered the channel (F8052) and into the rock-cut trough is a matter of conjecture as the entire area northwards from the channel had been truncated by later activity when the drainage and water-collection feature was constructed (see below, F8025).

the excavations 107

A stone-faced ovoid feature abutted the western side wall of the trough close to its southern end. Only half of the feature (1.2m x 1.3m) lay within the excavation area and its purpose remains unknown, but it is suggested that it functioned as a soakaway (Fig. 5.41, no. 5, F846). The feature was best preserved at its southern wall face, where it survived five courses in height (0.46m max.) (Pl. 5.39). The wall on the other side was reduced to just two courses in height (0.12m). The deposit of compact, mid-grey silty clay (F826), full of small mica-schist stones, situated inside the facing stones was not excavated to its full depth (0.4m). The surface of the deposit sloped from east to west to accommodate the drop in level of the stone-faced surround, suggesting that the five courses of stone on the western side and the two on the northern side were intentional features

to reflect this fall. The deposit extended below the surrounding paved surface (F835). A wall face abutted the southern side of the soakaway (Fig. 5.41, no. 6, F832). Unfortunately, due to the limited extent of the wall exposed within the excavation area, its relationship to the soakaway could not be established (Pl. 5.39). The wall face stood six courses in height (0.9m max.) and was built with tightly fitting, coursed, drystone masonry and there was a slight batter. It was orientated north-west–south-east, and its direction changed to north/south at the edge of the excavation area. The wall face continued on towards the south. The water management system was developed and extended by constructing a channel to the north that connected with the existing rock-cut water trough outside Cell B (Fig. 5.41, no. 7, F8025). This enlargement of the

Pl. 5.38 Leacht (F887) (centre foreground, under ranging rod) with kerb of trough (F845) visible within. Rear wall of leacht also formed the western side wall of the covered drain (F8007), looking east (G. Scally).

East O

West O1

Cell B

2 10 12 3

5 6 0

8 11

4 5m

1 Monastic enclosure wall

6 Scree (F8070/F8001)

11 Hearth floor (F326)

2 Scree (F8065)

7 Kerb for trough (F845)

12 Church enclosure wall

3 Lintelled drain (F8007)

8 Leacht (F887)

13 Paving (F29)

4 Channel (F8052)

9 Paving (F835)

Paving

5 Gravel (F8036)

10 Cross-slab 18 reused as lintel

Lintels

Fig. 5.46 East–west section through Cell B and surrounding features. Location of section O–O1 on Fig. 5.41.

Limit of excavation

108â&#x20AC;&#x192; high island: excavation of an early medieval monastery

water system was undertaken, probably in the mid- to late 12th century, at the same time as the beehive hut, Cell A, was built within the north side of the monastic enclosure wall (Section 5.4.3). The water-collection and drainage channel (total L: 20m) encircled almost the entire northeastern part of the monastery. Two sections of the channel were excavated, comprising its northern part (L: 10m) and its eastern part (L: 5m). From north to south a fall of 1.7m was noted over its entire length. The channel commenced as a shallow, rock-cut gully outside the monastic enclosure at the north-east (Fig. 5.41, no. 7, F8025). The gully (L: 5m x W: 0.2m, depth: 0.3m) had been cut into charcoal-flecked clay, which was possibly the old ground surface overlying the bedrock. Although the base of the channel was smooth, its sides were irregular and in places jagged edges of bedrock were clearly visible. As the channel extended in a south-easterly direction, it widened to 0.6m as it passed between the enclosure wall on its eastern side and the annulus around Cell A on the other side. Further south, where the channel ran past the

enclosure wall, it swung sharply east. In order to accommodate this change in direction, the annulus flared out. In this location, the side walls were built with stones set on edge (depth: 0.4m), which on the west were part of the masonry of the annulus (Fig. 5.41, no. 7, F8025; Pl. 5.40). The sidestones were neatly fitted together and the bases of many of them were set into slots cut into the bedrock floor and were secured in position with a compact packing of clay and small stones. When the water-collection and drainage channel reemerged from the unexcavated section into the east of the monastic enclosure, it was orientated in a south-easterly direction (Fig. 5.41, no. 7, F8025). The cut for the channel (W: 1.6m, depth: 0.55m) narrowed towards its base (Fig. 5.47). For the most part, the channel cut into the natural schist clay to the west and through various clays and natural gravel deposits to the east (Fig. 5.48, nos. 4, 5, F830w, F8025). As the channel extended south, it narrowed as it approached the earlier stone channel (F8052). In general, the channel (F8025) in the east of the enclosure was simiPl. 5.39â&#x20AC;&#x201A; South wall-face of soakaway (F846) in centre, with masonry wall-face (F832) (behind ranging rod); leacht (F843) to right and wall (F813) to left, looking north (G. Scally).

Pl. 5.40â&#x20AC;&#x201A; Water-collection and drainage channel (F8025) showing annulus and Cell A (right) and monastic enclosure wall (left), looking south (G. Scally).

the excavations â&#x20AC;&#x192; 109

lar to the part uncovered at the north. The flat-bottomed floor was also formed by the truncated surface of the bedrock. On the eastern side, a number of stones set on edge (some of which slipped out of position) formed a line (L: 2m) and on the western side, smaller stones were set on edge, indicating that this end of the channel had probably been stone-lined. Slender, rectangular slabs that extended the width of the channel may represent displaced lintels. A patch of grey schist (F8015) found adhering to the underside of one of the slabs suggests it had originally been secured in position in a similar manner to the in situ lintel covering the earlier rock-cut trough. When the channel was built, it removed paving stones from the earlier path to the well (Fig. 5.41, nos. 1, 2, Fig. 5.48, nos. 1, 2, F8028, F8068). The remains of the path that lay to the sides of the channel were left in situ, suggesting that the path and the well continued in use (Pl. 5.36). A stone-lined drain (L: 1.6m x W: 0.3m), in part roofed with lintels, replaced the earlier channel (F8052) in more or less the same position at the northern end of the rockcut trough (Fig. 5.41, no. 8, Fig. 5.43 (b), no. 5, F8007). The east wall of the drain was built outside and parallel to the wall of the earlier channel, while the west wall was formed by masonry of the rear (east) wall of a small leacht

constructed at the same time (Fig. 5.43 (b), no. 4, F887; Pl. 5.38). The east wall of the drain was composed of one to two courses of bulky, rectangular-shaped stones that had fallen inwards. The lowermost course was set into gravel that had accumulated against the wall of the earlier channel (Fig. 5.49, no. 8, F8036). The east wall of the drain projected northwards (0.3m) beyond the opposing wall, suggesting that it had originally continued further in this direction and connected to the more extensive watercollection and drainage channel (F8025). The other (west) wall of the drain (L: 1.3m) was composed of five stones laid on a shallow layer of fragmented schist on top of the wall of the earlier channel (Fig. 5.49, no. 3, F888). The stones on this side of the drain, which were smaller and more regular in shape, also functioned as the wall of the leacht. Three lintels, which were tightly wedged between the side walls, covered the drain and they were also supported on the earlier wall of the channel (Fig. 5.49).

The paved surfaces The eastern and southern parts of the monastic enclosure had paved surfaces surviving over most of the excavated areas. One of these covered the ground in front of and around the southern side of Cell B and it abutted the ker-

West P

East P1 1 2 17 Q

1 3

2 5

8 15

9 10

11 12

1 Grass

7 Stones (F8022)

13 Organic soil (F8032)

Ash

2 Sod (F800)

8 Silts (F880/F881/F890)

14 Natural scree (F830)

Bedrock

3 Peat (F805)

9 Coarse grit (F895)

15 Clay (F468)

4 Peat (F822/F829)

10 Silt (F8014)

16 Natural gravel (F497)

Channel sidestones

5 Schist and clay (F851)

11 Silt (F8029)

17 Monastic enclosure wall

Conjectural

6 Clay (F867)

12 Gravel and grit (F8031)

18 Scree (F8066)

Limit of excavation

Quartz

Fig. 5.47â&#x20AC;&#x201A; Westâ&#x20AC;&#x201C;east section through water-collection and drainage channel (F8025). Location of section on Fig. 5.41.

110 high island: excavation of an early medieval monastery

West S

East S1

Cell B 5 3 2

4 0

1 Well (F8068)

Lintels

2 Path (F8028)

Paving

3 Path (F875)

Bedrock

4 Channel (F8025)

Conjectural

5 Natural ground (F830w)

Obscured

Fig. 5.48 West–east section through path (F875), water-collection and drainage channel (F8025) and rock-cut well (F8068). Location of section S–S1 on Fig. 5.41.

East R

West R1

2 4

5 0

3 7

9 6

1m 8 Gravel (F8036)

1 Stony clay (F8041)

9 Scree (F8070/F8001)

2 Brown clay (F8042)

10 Kerb of trough (F845)

3 Fragmented schist (F888)

Side walls of channel (F8052)

4 Brown silt (F8044)

Side wall and lintel of drain (F8007)

5 Grey silt (F8048)

Leacht (F887)

6 Stones (F8053)

Leacht (F887) and drain (F8007)

7 Gravel and stones (F8051)

Lintel

Fig. 5.49 East–west section through covered drain (F8007) and earlier walled channel (F8052). Location of section R–R1 on Figs 5.41, 5.43 (b).

bed edge of the rock-cut water trough and the soakaway (Fig. 5.41, no. 9, F835). The paving continued southwestwards to cover the entire area between the wall of the church enclosure and a boundary wall to the south (Fig. 5.41, nos. 10, 15, F80, F813). To the south of the boundary wall, another paved surface was recorded (Fig. 5.41, nos. 11, 12, F833, F833a). The paving outside Cell B and in the area where it abutted the water trough and the soakaway consisted of mica-schist slabs, many of which were shaped to accommodate the outline of the adjoining slab, which gave the surface a degree of smoothness and sophistication not seen in other areas (Pls 5.33, 5.38). In front of the entrance to Cell B, most of the larger slabs had their long axes aligned east–west. As the surface extended around the cell, the long axes of the slabs changed to north–south, which gave the surface a path-like appearance. The largest slabs were located outside the cell entrance, one of which (over 1m in length) was partly sealed by the leacht, but it respected the kerbed edge of the water trough. Smaller slabs were equally well laid close to the southern side of the church enclosure wall. Away from the wall, the slabs were even smaller and they were not as well laid. These two areas were separated by a short length (2m) of stones on edge that appeared to be a kerb. The paved surfaces were laid partly on top of and partly abutting the later boundary wall (Fig. 5.41, nos. 9, 10, 15, F80, F813, F835). The paving

the excavations 111

lay beneath the south face of this wall and here it was composed of mainly small, thin slabs of schist that were well preserved. South of the wall, the condition of the surface deteriorated and it was composed of a more diverse range of slab sizes, nearly all of which were loose and set apart from one another (Fig. 5.41, nos. 11, 12, F833, F833a). A path led around the north-eastern side of Cell B that connected the paved surface (F835) in front and to the south with another extensive area of paving (F422) around Cell A and the area within the northern part of the monastic enclosure (Fig. 5.41, no. 16, F875; Pl. 5.31). The path, running south-east–north-west, was traced for most of its length (5m). It was laid in a hollow that was cut into the natural boulder clay (Fig. 5.48, nos. 3, 5, F830w, F875). The slabs of the path were not particularly well laid. Several overlapped one another and in places an upper layer of smaller slabs was visible above a lens of silt that had accumulated along the sides of the hollow. The southern end of the path was laid on top of the paved surface (F835) and was separated from it by a thin layer of soil (depth: 30mm). The paved surface covering much of the northern part of the enclosure was identified to the north-western limit of the path across the full length (15m) of the excavation area (Fig. 5.41, no. 17, F422; Pl. 5.31). This surface was well laid with tightly fitting slabs that were secured in place by smaller slabs and, less frequently, by spalls. In front of Cell A, the slabs overlay a shallow deposit of dry, fragmented schist clay, containing hazel and oak charcoal, which lay on the earlier paving on which the cell had been built (Sections 5.4.3, 9.3, F468). The southern part of the paving, with the long axes of the slabs lying mainly north– south, survived intact. At the western end of the excavated area, the paved surface abutted the wall outside the church enclosure (Fig. 5.41, F91). Further areas of paving, a path and a linear feature were revealed in the trench opened in the south-west of the monastic enclosure (Fig. 5.1, Area 6). They were likely to have been associated with the remains of the beehive huts recorded by George Petrie in the early 19th century and which were still visible in the mid-20th century (Sections 3.2, 4.4). The paving (F80) outside the church enclosure wall continued around the south-west corner (Fig. 5.41, no. 20, F615a). Another part of this surface (L: 3.3m) was uncovered to the south, where it formed a path (Fig. 5.50, no. 3, F615a; Pl. 5.41). The slabs, most of which were slender and rectangular in shape, had their long axes lying east–west. A marked north–south curve was noted along the exposed length of the path (W: 1m). At the southern end, the path continued in a south-easterly direction beyond the excavation area, but a small stone on edge adjacent to the path suggests it may have terminated close to this point. In the centre, the path overlay a large, flat slab, part of a possible paved surface, with stones on

edge (H: 0.26m) flanking the northern side (Fig. 5.50, no. 2, F622; Pl. 5.41). These stones demarcated another area of paving, which was composed of small slabs of schist (Fig. 5.50, no. 4, F628a). The northern extent of the slabs was defined by a kerb formed by two stones on edge. A linear feature (W: 0.6m), running north–south, was partially exposed (L: 2.4m) and consisted of one to two courses of mica-schist (Fig. 5.50, no. 5, F621). Hazel, hawthorn and oak were identified in the charcoal from this feature, which abutted the path (F626a) to the south (Section 9.3, F621). The paved surface (F628a) was sealed by a shallow lens of silty clay that contained mainly hazel charcoal and a single fragment of birch (Section 9.3, F626). Above, two levels of later paving represent repair or replacement (Fig. 5.50, nos. 6, 7, F628b, F628c). These surfaces comprised stones that were more rectilinear in shape than those in the underlying surface. Both of the later surfaces were contained by the stones set on edge to the north and south.

Leachta The leacht, situated directly in front of the entrance to Cell B, was built over the northern end of the rock-cut water trough and the side wall of the drainage channel (Fig. 5.41, no. 13, Fig. 5.43 (b), no. 4, F887). It was lying on top of the paved surface. The east or rear wall of the leacht was also the side wall of the covered drain, indicating that both structures were built simultaneously (Pl. 5.38). The leacht

Pl. 5.41 Paved surface (F615a), looking north (G. Scally).

112 high island: excavation of an early medieval monastery

N 1

Covered by lintels

1 Schist (F607, F624, F629)

5 Linear feature (F621)

Upright stones

2 Paving (F622)

6 Paving (F628b)

Lintels

3 Paving (F615a)

7 Paving (F628c)

Obscured

4 Paving (F628a)

8 Wall chamber (F721); Area 7

Limit of excavation

Fig. 5.50 Excavation Area 6 showing path (F615a) and associated paved areas.

measured 1.3m north–south x 0.9m east–west. Its walls of slender, rectangular stones survived one to three courses in height (0.22m). At the east, where the leacht wall was also the side wall of the covered drain, the stones were square and more substantial in size, giving them a more robust form suitable to their dual function. The interior of the leacht was filled with three deposits. At the base, a shallow layer of mica-schist stones was set in flaky schist soil, which was sealed by dry, fragmented schist, and above there was more fragmented schist mixed with silty clay. These deposits were contained within the leacht, with the exception of the dry fragmented schist, which spilled over the rear (east) wall and formed a bedding on which the upper course of the wall had been set (Fig 5.49, no. 3, F888). Hazel and oak charcoal was identified from this fill layer, from which particles of iron slag were also retrieved

(Section 6.5, 888:2; Section 9.3, F888). Slender, upright stones were placed at the north-eastern and south-eastern corners of the leacht (Pl. 5.38). The stone at the north-eastern corner stood 0.38m above the contemporary ground surface and was set in a circular pit and secured in place by a fill of compact grey/brown clay. Both stones, though unadorned, clearly marked the location of the leacht. A decorated cross-slab, found in the overlying rubble, may originally have stood upright at the leacht (Section 7.1, Cross-slab 37). The remains of another leacht were partially exposed in the south-eastern corner of the excavation area (Fig. 5.41, no. 14, F843; Pl. 5.39). This leacht abutted the wall face (F832) and it was built on the paved surface (F833). The front and northern sides survived for one to two courses in height (0.2m), within which a fill of silty clay and frag-

the excavations â&#x20AC;&#x192; 113

mented schist and stones was evident. Its construction replicated that of the other leacht situated in front of Cell B.

Boundary walls A low wall lay to the south of Cell B and the church enclosure wall (Fig. 5.41, no. 15, F813). This wall, which was partially exposed within the excavation area, was built against the soakaway, and from here it ran in a south-westerly direction. At its north-eastern end, an attempt had been made to bond the wall masonry into the side wall of the soakaway and for a distance of 1.6m the wall (H: 0.5m, W: 0.8m) was faced on both sides with shaped blocks and had a core of soil and rubble. From this point the wall narrowed on its southern side, and 5m from its eastern end two upright jamb stones marked a blocked entrance 0.5m wide. Further along the wall (at 9m) the masonry changed to upright rectangular blocks of stone. The wall turned sharply towards the south-east, where it extended into the section face at the edge of the excavation area. It could be traced for another few metres beneath the grass-covered peat where it appeared to be heading towards the main south-east entrance to the monastery. Another low wall was uncovered that was built approximately mid-way between Cell B and Cell A in the north of the monastic enclosure (Fig. 5.41, no. 18, F848; Pl. 5.42). The incomplete wall (L: 5.2m x W: 0.6m), orientated eastâ&#x20AC;&#x201C; west, had one to two courses of the faces (H: 0.12m) and a core of soil and small stones surviving. Its eastern end overlay the stone path (F875) and the paved surface (F422) where the slabs formed an upper layer that abutted the wall. Its western end was built against the curved side wall of the north-east entrance to the church enclosure (Fig. 5.41). The southern side of the wall partly overlay one of the sidestones of Grave 11 outside the entrance, suggesting that the wall may have been erected specifically to separate the grave from the paved areas to the north and east (Pl. 5.42). The western end of the wall was removed when a cross base was inserted within the wall directly in front of the entrance to Cell A (Fig. 5.41, no. 19, F415; Pl. 5.42). The base, a shaped rectangular block (0.52m x 0.25m, H: 0.17m), had a socket (0.18m x 65mm, depth: 70mm) centrally placed on its upper side. The socket would have held an upright cross, which would have been fully visible from inside the cell. The base was secured in place by large boulders that were inserted to the east and west between the ends of the walls.

Phase 3: late 12th/early 13th to mid-15th century Extensive layers of clay, silt and stone had accumulated over much of the paved surface within the enclosure after the monastery was abandoned in the late 12th or early 13th century. A small number of features, including hearths,

deposits with food waste and rudimentary surfaces, were identified that testified to an intermittent human presence relating to temporary occupation and to pilgrimage. In the north-east of the enclosure, the disused watercollection and drainage channel of the monastic occupation (Fig. 5.41, no. 7, F8025) contained shallow layers of soil with lenses of fragmented schist clay, which were the result of subsidence from the high ground to the east (Section 9.3, F489, F484). The extensive deposit of clay and stone covering the sloping ground to the north-east spilled into the channel to form part of the fill (Fig. 5.51, no. 1, F459). This was sealed by a more extensive layer of silty clay that filled a large portion in the northern end of the channel (Fig. 5.51, no. 2, F458). Further south, where the channel passed between Cell A and the monastic enclosure wall, shallow layers of silt and clay, deposits that were likely to have formed naturally after the channel had fallen into disuse, covered its rock-cut floor. Above these, the channel was filled with stones and clay flecked with charcoal (Section 9.3, F483). Further south, there was a deposit of soft brown peat and soil (depth: 0.14m) with a scattering of charcoal and a small amount of faunal remains, mainly of fish (Section 9.1, F482). A silver halfpenny of King John, dating to the early 13th century, was recovered from the surface of this deposit (Section 6.6, 482:1). The upper fill in this stretch of the channel consisted of layers of silty clay containing fish and bird remains (Fig. 5.51, no. 3; Section 9.1, F463). These layers were partially sealed by charcoal and ash-stained clay. The charcoal sample contained mainly hazel and heather, with lesser amounts of ash and oak (Section 9.3, F464). Thirty-four grains of barley and considerable quantities of fish and bird remains were present (Sections 9.1, 9.2, F464). A hone for sharpening implements was also recovered (Section 6.2, 464:1). In the south of the channel, a concentration of ash and charcoal was possibly the remains of a hearth (Fig. 5.51, no. 4, F445). Hazel, heather and ash were identified in the charcoal (Section 9.3, F445). Oat and barley as well as a small quantity of sheep/goat, cattle and fish bones comprised the food waste (Sections 9.1, 9.2, F445). The basal fill layers in the drainage channel in the section in the east of the enclosure were predominantly silt-based, which were combined with washed-out natural gravels and stones. The composition of these deposits, which were markedly different from those to the north, suggests that they accumulated naturally as a result of the location of the channel at the base of steeply sloping ground. The deposits extended south as far as the Phase 2 covered drain (F8007). Compact organic soil, a lens of coarse gravel and grit and a more extensive layer of grey, gritty silt and stones covered the base of the channel within this stretch (Fig. 5.47, nos. 11â&#x20AC;&#x201C;13, F8029, F8031, F8032). Layers of silt containing oak charcoal were sealed by coarse grit and stones which lay

114 high island: excavation of an early medieval monastery

Pl. 5.42 Wall (F848) built partly over restructured Grave 11 outside the church enclosure wall, looking south-west. The cross-base (F415) is at centre right (G. Scally).

in the vicinity. These layers were surrounded by a thick layer of iron pan, a feature indicative of deposits accumulated in still, watery conditions (Fig. 5.47, nos. 9, 10, F895; Section 9.3, F8014). Above the grit, yet further layers of silt with charcoal and fish remains filled the upper levels of the channel (Fig. 5.47, no. 8, Fig. 5.51, no. 5; Sections 9.1, 9.3, F890, F881, F880). Two small spreads of ash were lying on top (Fig. 5.51, no. 6, F879a, F879b). Slender, rectangular-shaped stones which were scattered throughout may represent collapse from the east wall of the channel. The sequence of fill deposits identified within the covered drain (F8007) and those within the earlier channel (F8052) largely reflected those found in the drainage channel (Figs 5.48, 5.51, nos. 1–6). This suggests that the silting up of the drain took place simultaneously with that of the channel. South of the drain, the rock-cut trough (F845) was filled at its base with fine, black silt (Section 9.3, F854). Coarsegrained, brown silt with small stones lay above (Fig. 5.51, no. 8; Section 9.1, F852). The tuyère fragment associated with Phase 1 smithing beneath Cell B was found in this silt fill (Sections 5.3, 6.5, 852:1). At the base of the steeply sloping ground to the east of the disused water management systems, deposits of coarse grits, gravels and soils had accumulated (Fig. 5.51, no. 13). Immediately adjacent and partly overlying these layers, a concentration of rubble that possibly collapsed from the enclosure wall overlay the lintelled roof of the drain (F8007). A small area of cobbles

(D: 0.6m) and a spread of silt lay in close proximity on the sloping bank of natural ground (Fig. 5.47, no. 14, F830, Fig. 5.51, no. 15, F894, F8013). A layer of silty clay and fragmented schist covered the ground between the north side of Cell B and the boundary wall to the north (Fig. 5.51, nos. 10, 11, F847, F848; Section 9.1, F847). Above, a lens of finely textured, possibly wind-blown soil covered the area and tapered away close to the north-east corner of the cell (Fig. 5.51, no. 10; Section 9.1, F837). It overlay the Phase 2 clay packing (F803) around the base of the cell wall. A small deposit of peat ash, possibly a hearth, lay on the surface of the wind-blown soil and partly overlay the boundary wall (Fig. 5.51, no. 12, F849). Around the north-eastern side of the cell, the Phase 2 path (F875) was covered by a deposit of mid-brown clay and fragments of mica-schist. A similar type of clay was identified further east covering the fill and ash spreads on the water-collection and drainage channel. This layer petered out at the north of the excavation area (Fig. 5.47, no. 6, F867). Close to the cell, it was sealed by a moderately compact and extensive lens of grey mica-schist clay. Adjacent to the north-east corner of the cell wall the clay was exceptionally compact and possibly represents a trampled surface. A deposit of ash, possibly the remains of a hearth, lay on the Phase 2 paved surface (F835) at the wall of Cell B to the south of the entrance (Fig. 5.51, no. 16, F836).

the excavations 115

24 G G

5 7

9 15

14 16 17

20 21

5m 23 Boundary wall (F813) 24 Silty clay (F479)

1 Clay and stones (F459)

12 Ash (F849)

2 Silty clay (F458)

13 Gravel (F8003, F8004, F8006, F8008, F8030)

25 Ash (F472)

3 Silty clay (F463)

14 Clay and stone (F891/F893)

26 Stony clay (F469)

4 Ash (F445)

15 Cobbles (F894) and silt (F8013)

5 Silt (F880/F881/F890)

16 Burnt debris (F836)

6 Ash (F879a/b)

17 Clay (F841)

7 Stones (F8022)

18 Natural gravel (F830 w)

8 Trough (F845) filled with silt (F852)

19 Clay and rubble (F834)

Ash

9 Clay (F803)

20 Possible surface (F838)

Obscured

10 Schist (F847) sealed in fine soil (F837)

21 Ash (F78a)

Conjectural

11 Wall (F848)

22 Rubble and clay (F78)

Limit of excavation

Lintels Upright stones G

Granite Annulus

Fig. 5.51 Plan of the monastery showing location of later medieval features within the monastic enclosure, Phase 3.

116â&#x20AC;&#x192; high island: excavation of an early medieval monastery

The ash and the area in the immediate vicinity was covered by a layer of clay containing charcoal of hazel, birch, oak and juniper (Fig. 5.51, no. 17; Section 9.3, F841). The clay may have been thrown down deliberately to extinguish the hearth fire. To the south, stone rubble, set within dark brown silty soil with frequent mica-schist fragments, covered the paved surface (Fig. 5.51, no. 19, F834). As the deposit extended to the west towards the church enclosure wall, the rubble was less randomly dispersed and it was difficult to consider it as merely collapse. It is possible that, in this location, it may represent the remains of a rudimentary stone surface (Fig. 5.51, no. 20, F838). Its western end was overlain by an area of burning, the remains of a small hearth, which was contained on its western side by a single edge-set stone (Fig. 5.51, no. 21, F78a). Continuing in a westerly direction, the deposit reverted to rubble, which was mixed with finely textured silty clay and fragmented schist (Fig. 5.51, no. 22, F78). This deposit covered the Phase 2 paved surface (F80) and abutted the boundary wall (Fig. 5.51, no. 23, F813). Deposits and features associated with the period after permanent occupation of the monastery ceased were uncovered in the trench in the south-west of the enclosure (Fig. 5.1, Area 6). Traces of a rudimentary surface, consisting of fragmented schist mixed with grey silty clay and a scattering of flat stones, covered the area adjoining the church enclosure wall (Fig. 5.52, no. 1, F613, F623). Charcoal from a variety of species and a small quantity of cattle and sheep/goat bones were identified (Sections 9.1, 9.3, F613, F623). A shallow lens of ash mixed with light grey schist clay and also with charcoal was evident within the clay and stones (Section 9.3, F616). To the south, the surface consisted of schist and clay (Fig. 5.52, no. 2, F609). A possible leacht (1.9m x 1.1m), with its eastern wall composed of four rectilinear stones and with well-defined corners, was revealed in the middle of the enclosure (Fig. 5.52, no. 3, F612). The stones were set into a lens of dark soil containing charcoal that had accumulated above the Phase 2 paved surface (Section 9.3, F619). Another area consisting of charred peat and charcoal lay in close proximity and a separate lens of the same type of soil covered the flat slab of the earlier surface (Section 9.3, F617, F625). The west wall of the possible leacht comprised three edge-set stones that were loosely held in place by the surrounding schist layer (F623) and may represent an area of repair or replacement, or it is possible that they had become dislodged from their original position. The core of the possible leacht was filled with fine brown soil covered by stones (Fig. 5.52, no. 4; Section 9.3, F614). To the south-west, a linear feature composed of small stones (W: 1m) was partially exposed (Fig. 5.52, no. 5, F608). The stones were defined on both sides by a slightly raised edge.

Phase 4: mid-15th to late 20th century The findings from the excavations indicated that the monastery on High Island continued as a place of intermittent activity from the mid-15th century to the 19th century, with the archaeological record suggesting that this was more intense in the latter part of the period. Pilgrims would have come to venerate the men buried in the graves outside the eastern end of the church, which remained on view. Temporary occupation of the monastic buildings occurred. Hearths and soil deposits contained considerable amounts of food waste consisting of the bones of domestic animals, with wild species of bird and fish also represented. Temporary residents were probably the tenants farming the island and fishermen, and the miners were permanent dwellers for a short period in the 1820s (Section 3.3). Between Cell A and the boundary wall (F848) in the north of the enclosure, stiff, grey clay mixed with stone accumulated above the Phase 2 paved surface (F422). Within the clay, a shallow spread of peat ash possibly represented the remains of a small hearth. Charcoal, the residues of hazel, ash, birch and heather, and a few bones of sheep/goat and cattle were in the clay (Sections 9.1, 9.3, F451). The area was covered by peaty soil (Sections 9.1, 9.3, F447). Layers of rubble had accumulated between the church enclosure wall and the monastic enclosure wall (Pl. 5.43). A moderate quantity of faunal remains was amongst the rubble (Section 9.1, F436, F441). Within the northwestern part of the excavation area, a layer of clay and stones (F443) covered the Phase 2 steps leading to the passage through the monastic enclosure wall, indicating that it too was out of use by this time. A decorated cross-slab was recovered from this layer (Section 7.1, Cross-slab 34). The area of the disused water-collection and drainage channel in the north-east of the monastic enclosure was sealed by a deposit of peaty soil (Sections 9.1, 9.3, F447). A deposit of shale soil was revealed that abutted the annulus of Cell A and sealed the earlier layers (F445, F464). A small, rectangular setting of stone, filled with burnt and unburnt limpet shells, formed a kerbed hearth on the surface of the soil (Section 9.1, F442). A spread of ash lay in the vicinity on the southern side. Stone rubble filled the surrounding area between Cell A and the monastic enclosure wall to the east. A grinding slab was retrieved in the rubble (Section 6.2, 450:1). Further south, the rubble was covered with fine, light brown, silty clay, containing a considerable quantity of food waste, mainly of fish (Section 9.1, F430). Layers of fragmented schist and stone lay on top (F435). Within these layers, there were spreads of ash with large amounts of faunal remains, especially those of fish and birds (Section 9.1, F434, F449) Limpet shells were also present (F416, F433). These fires and food waste concentrations were recovered around the east and north-

the excavations 117

N Church enclosure wall

Covered by lintels

2 6 5

1 Fragmented schist and rudimentary surfaces (F623, F613)

5 Linear feature (F608)

Obscured

2 Charcoal-flecked schist and clay (F609)

6 Wall chamber (F721); Area 7

Limit of excavation

3 Leacht? (F612)

7 Buttress (F79)

4 Stony fill (F614)

Lintels

Fig. 5.52 Excavation Area 6 showing later medieval features.

east of Cell A. Shelter afforded by the cell walls probably explains the proliferation of activity in this location. Peat growth, containing moderate quantities of cattle, sheep/ goat, rabbit and bird bones, but no fish remains, covered the entire area in the north and north-east of the monastic enclosure (Section 9.1, F425, F429). An iron donkey shoe was found, indicating that, most likely, these animals were transported to the island for use during the mining campaign in the 1820s (Sections 3.3, 6.4, 425:1). The peat was interspersed with rubble, within which there were pockets of fragmented weathered schist. A decorative piece of iron, possibly a furniture fitting or a handle, and a penny of George V, minted in 1922, were found (Section 6.4, 401:1, 402:2). The rubble obscured all but the uppermost part of the annulus around Cell A (Pl. 5.44). A cache of limpet shells and some peat ash were found in front of the cell entrance, suggesting that this was the remains of a meal.

Large rectangular stones (0.7–0.8m), lying amongst the rubble in front of the cell and to both sides, were probably the collapsed lintels that had covered the entrance. Layers of peat (up to 0.3m in depth) covered the area surrounding Cell B (Section 9.1, F822, F829). In the southeast of this area, a spread of ash consisting of burnt peat may represent the remains of a small hearth. Hazel charcoal was identified in another deposit of ash (Section 9.3, F820). The peat was shallower close to the cell walls where it overlay the Phase 2 clay packing (F803), and to the south-west it overlay a deposit of clay and stone that had accumulated over the Phase 3 rough stone surface (F838). In the east of the monastic enclosure, stone rubble, probably collapse from the enclosure wall, lay on top of the peat. The Phase 2 paved surfaces (F833, F833a) to the south of the boundary wall (F813) were covered by a build-up of soil and rubble (Section 9.1, F831). Stone mixed with peat

118 high island: excavation of an early medieval monastery Pl. 5.43 Rubble (F436) surrounding Cell A, looking south (G. Scally).

Pl. 5.44 Cell A before excavation showing rubble almost completely obscuring annulus (F428), looking north-west (G. Scally).

Pl. 5.45 Uppermost level of rubble (F801) surrounding Cell B, looking west (G. Scally).

the excavations 119

lay on top of the earlier layers (Section 9.1, F816). North of the boundary wall a similar deposit of stone and peat, containing the most extensive quantity of faunal remains, covered the area and continued as far as Cell B (Section 9.1, F811). A blue glass bead of medieval date was found in this deposit (Section 6.9, 811:2). Close to the south-east corner of the cell, a flat slab had been embedded into the rubble to form a small hearth on which peat ash, with burnt and unburnt fragments of bone and limpet shells, had survived (Section 9.1, F814). Four flat stones had been carefully placed on top, probably to extinguish the fire. A green glass bottle fragment, probably relating to the miners’ occupation in the early 19th century, was amongst the hearth residue (Section 6.10, 814:1). Pockets of ash were in close proximity to the hearth (Section 9.1, F818). These may have originated as wind-blown deposits, or perhaps they were dumped residue from earlier fires made within the hearth. The area extending south and west from Cell B was covered by rubble that probably collapsed from the church enclosure wall (Section 9.1, F82). Extensive layers of rubble covered the entire area, in which a modern brass button was found (Section 6.4, 802:7, Section 9.1, F806). Peat growth pushed through the stones. Three small pockets of limpet shells and other fish remains lay amidst the rubble, as well as an extensive spread of ash (Section 9.1, F807, F809). Another dense layer of rubble lay on top around the area of Cell B (Pl. 5.45; Section 9.1, F801). The rubble was most concentrated closest to the cell walls and its depth decreased as it extended away on both the northern and southern sides. In front of the cell, the collapsed rubble was composed of square and rectangular stones that may have originated from the monastic enclosure wall (Section 9.1, F840). This rubble filled the silted-up Phase 2 water trough (F845) and completely concealed its presence. The Phase 1 well (F8068) was filled with fine silt (depth: 0.3m). Considering that the well was the main source of freshwater within the enclosure, it is likely to have remained in use until relatively recently. The entire area was covered by a thick layer of grass-covered sod (0.25m), which enveloped all but the upper parts of the walls of Cell B (Section 9.3, F800). Stone rubble, probably collapse from the adjacent monastic enclosure wall, was exposed in the trench in the south-west of the site (Fig. 5.1, Area 6; Pl. 5.46; Section 6.5, 618:2). A shallow layer of sod covered the rubble. A scant cover of stone was found embedded on the surface of the sod to the east, while to the north the stone was more concentrated and formed a relatively dense layer. Above the Phase 3 leacht (F612) in the middle of the enclosure, stones were heaped up in a mound. A modern brass button was found (Section 6.4, 600:3) in the sod layer of the trench in addition to a cache of 35 quartz pebbles.

5.4.2 The monastic enclosure wall Phase 1: 8th to mid-11th century The northern side was the shortest stretch of the monastic enclosure wall, measuring 9m in length, and that part of it to the west of Cell A (L: 5m) was excavated (Fig. 5.1, Area 4). This was the least well-preserved stretch, and for this reason it was decided to investigate the wall where it was almost level with the ground (Fig. 5.53). Despite the poor preservation, two construction phases in the monastic enclosure wall were identified. To the north-east of the enclosure wall there was a marked downward slope. In the elevated areas, bedrock and fragmented weathered schist covered by shallow lenses of charcoal-flecked peaty clay, which appeared to be the remains of an old ground surface, were exposed at the lowest excavated level (Fig. 5.53, no. 1, F493, F497). The enclosure wall was built in a hollow, on a natural deposit of peat that covered the bedrock (Fig. 5.54, no. 42, Fig. 5.55, no. 38, Fig. 5.62, no. 12, F4076, F4089). When this level was reached, rising ground water completely submerged the area. Layers of clay (up to 0.4m in depth) were laid down, probably to consolidate the underlying peat and possibly also to raise the level of the ground in advance of

Pl. 5.46 Rubble (F627/F618) adjacent to western side of the monastic enclosure wall, looking west (G. Scally).

120 high island: excavation of an early medieval monastery

N 3

1 6

7 8

11a

11 9

11b

10 2

5m 1 Peaty clay over fragmented schist and bedrock (F493/F497)

10 Clay and stones (F4082)

2 Fragmented schist and bedrock (F497)

11 Unexcavated stretch of monastic enclosure wall (F4040)

3 Flaky schist (F4063)

11a North wall-face (F4040 N)

4 Schist and clay (F4060)

11b South wall-face (F4040 S)

5 Stony clay (F4066)

Later structures

6 Cut (F4059) for wall (F4040)

Obscured

7 Fill (F4058) of cut (F4059)

Conjectural

8 Fill (F4050) of cut (F5049)

Limit of excavation

9 Stony schist and clay (F4064)

Fig. 5.53 Excavation Area 4 showing foundation deposits for the primary monastic wall (F4040).

the construction of the wall (F4040). These shallow layers consisted of clay and schist (Fig. 5.54, nos. 31–34, Fig. 5.55, nos. 17–20). Oak charcoal was identified in one of these deposits and ten fragments of animal bones of indeterminate species were in another (Section 9.1, F4072, Section 9.3, F4073). Above, patches of peaty soil, containing a few fragments of faunal remains (Section 9.1, F4070), were interspersed with more extensive layers of small stones, flaky schist and silty clay (Fig. 5.54, no. 18, F4060, Fig. 5.55, nos. 30, 31, F4063, F4067). A shallow cut (depth: 0.15m) was dug parallel with and north of the wall (Fig. 5.53, no. 6, F4059). The fills of the cut, comprising schist and clay, a lens of black silt and a lens of brown, sandy loam continued south to form a bedding deposit on which the wall face and core of the monastic enclosure wall were built (Fig. 5.53, nos. 7–9, Fig. 5.54, nos. 20–22, 30, Fig. 5.55, nos. 16, 33, F4050, F4058, F4064). These deposits contained the bones of sheep/goat, and oak charcoal was identified from the

black silt (Section 9.1, F4050, F4051, F4064, Section 9.3, F4051). Excavation to this depth south of the wall inside the enclosure was limited to a small area (2.5m x 1.5m) beside Cell A, as the remainder of the area was not explored beneath the Phase 2 paved surface (F422) (Fig. 5.53). In this location, clay, stone and flaky schist interspersed with patches of grey clay covered the natural peat (Fig. 5.54, no. 41, F4084). A layer of dispersed stone and silty clay covered the area and formed the bedding for the wall on this side (Fig. 5.53, no. 10, Fig. 5.54, no. 40, F4082).

The primary monastic enclosure wall The primary wall (W: 2.4m) was essentially an earthen bank of soil and small stones revetted on each side with a stone facing (Fig. 5.53, nos. 9, 11a, 11b; Pl. 5.47). Only the lowermost part of the walls remained in situ. One to three courses of stone, 0.4m in height, survived on the northern (outer) side (Fig. 5.54, no. 8, Fig. 5.55, no. 7, F4040 N; Pl. 5.48). The poorly preserved masonry consisted of

the excavations 121

rectangular stones of small size. The wall face continued on underneath the wall of Cell A and its annulus (Fig. 5.54, nos. 2, 8, F428, F4040 N; Pl. 5.48). On the southern (inner) side, the wall face (H: 0.24m) survived for one to two courses (Fig. 5.54, no. 27, F4040 S). The stones on this side were better preserved and slightly more substantial in size. The core of the wall consisted of redeposited material thrown down to fill the space between the faces. The earliest fill layer comprised small, angular schist stones within a light grey-/fawn-coloured clay and gravel (Fig. 5.54, no. 29, Fig. 5.55, no. 15, Fig. 5.56, no. 4, F4033, F4035). It was deepest where it abutted the northern wall face and from here it sloped downwards, suggesting that it may have been deliberately laid down behind the face in order to stabilise and consolidate this stretch of masonry. A shallow layer of dark brown silt covered the remaining area of the wall core (Fig. 5.54, no. 28, Fig. 5.55, no. 14, Fig. 5.56, no. 5, F4052). The silt was fine-textured, smooth and greasy, similar to deposits formed in clayey areas after heavy rain. Charcoal, mainly of hazel and heather, and a considerable number of faunal fragments of indeterminate species were recovered (Sections 9.1, 9.3, F4052). It was covered by stony, brown, silty clay and by another layer of darker brown, peaty clay (Fig. 5.54, nos. 24, 25, Fig. 5.55, nos. 12, 13, F4030, F4034). Patches of ash, identified as oak, and a large amount of burnt and unburnt animal bone were in both of these layers, indicating that they originated from domestic refuse (Sections 9.1, 9.3, F4030, F4034). Two porphyry fragments and a dished grinding slab were recovered (Section 6.2, 4034:1–4034:3). Excavation beyond the monastic enclosure wall was limited to those areas not obscured by the later structures outside and the earlier paving inside. A layer of silty clay (F4043), full of flakes and fragments of mica-schist with occasional patches of charcoal, covered the pre-construction layers of schist and clay to the north of the enclosure wall (Fig. 5.54, no. 16, Fig. 5.56, nos. 6, 9, F4043, F4046, F4047). It was deepest close to the wall, and oak was identified in a charcoal sample and animal bone fragments were recovered (Section 9.1, F4043, F4046, F4047, Section 9.3, F4055). The preponderance of fragmented stone adjacent to the wall suggested that the deposit originated from activities related to its construction. A dispersed spread of small, flat stones with charcoal appeared to form a localised surface (Fig. 5.54, no. 15, Fig. 5.56, no. 10, F4038, F4044; Section 9.3, F4038). Further spreads of flat stones evident across the area suggest that this surface may originally have been more extensive. Accumulations of organic soils and a scatter of stones covered the area (Fig. 5.54, no. 14, Fig. 5.55, no. 26, F4039, F4042, F4045). Occupation debris consisted of hazel and oak charcoal and a few animal bone fragments (Section 9.1, F4039, F4045, Section 9.3, F4039). These layers were sealed by a shallow spread of charcoal-

Pl. 5.47 Primary monastic enclosure wall (F4040) showing north and south wall-faces, looking east. Large stone on left is part of the later wall-face (F4025 N) (G. Scally).

Pl. 5.48 Masonry from the northern wall-face (F4040 N) of the primary monastic enclosure wall continuing behind masonry of the later annulus/extended monastic wall, looking south (G. Scally).

flecked, silty clay that, in places, abutted the primary enclosure wall (Fig. 5.54, no. 13, F4037). Charcoal extracted from this layer, which post-dates the enclosure wall, produced a radiocarbon date of cal. AD 778–1000 (UB-4987, 1138±48 BP). Mid-brown clays with a distinctive iron stain throughout covered the area north of the wall (Fig. 5.54, no. 12, F4032, F4036, F4049). Birch, heather and juniper were identified in the charcoal (Section 9.3, F4036). Inside, or south of, the enclosure wall excavation was only possible in the small (unpaved) area close to Cell A (Fig. 5.53). Mid-brown, silty clay (depth: 0.16m) with patches of decayed grey schist clay and a scattering of stones was overlying the consolidation layers pre-dating the primary wall (Fig. 5.54, no. 39, F4057). This silty clay abutted the primary wall face (F4040 S) and it was also the layer on which Cell A was constructed. The clay was deepest at the wall, suggesting that it may have been banked up deliberately to level the ground in front. It is also possible that the clay was banked up in this fashion at a later period, to level the ground in preparation for construction

18 Silty clay (F4060) 19 Cut (F4059) 20 Schist and clay (F4058) 21 Sandy loam (F4053) 22 Silt (F4051) 23 Clay and schist (F4073) 24 Peaty clay (F4030) 25 Stony clay (F4034) 26 Wall-face (F4025 S)

6 Stones and clay (F4019)

7 Wall-face (F4025 N)

8 Wall-face (F4040 N)

9 Structure (F427); east wall

10 Clay and stones (F479)

11 Silty clay (F496)

12 Clay layers (F4032/F4036/F4049)

13 Charcoal and clay (F4037)

30 32 33

42 34

28 24

39 Silty clay (F4057)

38 Wall-face (F4025 SS)

37 Organic soil (F4027)

36 Ash and clay (F478)

35 Paving (F422)

34 Clay and schist (F4081)

33 Clay and schist (F4074)

32 Clay and schist (F4080)

31 Clay and schist (F4072)

30 Schist and clay (F4064)

29 Schist and gravel (F4033/F4035)

28 Silt (F4052)

27 Wall-face (F4040 S)

26 Q

South T1

Limit of excavation

Conjectural

Quartz

Bedrock

Paving

Masonry of extended monastic enclosure wall and annulus

Extended monastic enclosure wall (F4025)

Primary monastic enclosure wall (F4040)

42 Peat (F4076/F4089)

41 Stone and schist (F4084)

40 Silty clay (F4082)

Fig. 5.54 Elevation of west wall of Cell A and section through primary and extended phases of the monastic enclosure wall, post-monastic building (F427) and underlying deposits. Location of elevation T–T1 on Fig. 5.61.

17 Schist and clay (F4063)

5 Peaty soil with charcoal (F4017)

20 21 22

4 Fragmented schist (F4010)

16 Silty clay and schist (F4043)

3 Soil and stone (F4002)

15 Stones (F4038/F4044)

14 Organic soils (F4039/F4042/F4045)

1 Cell A

Offset 0.6m behind cell

2 Annulus (F428)/wall-face (F4025 N)

North T

122 high island: excavation of an early medieval monastery

32 Cut (F4059) 33 Clay (F4050)

18 Clay and schist (F4081) 19 Clay and schist (F4088) 20 Clay and schist (F4087) 21 Structure (F427); west-wall 22 Structure (F427); north-wall 23 Silt, clay and stones (F479) 24 Silty clay (F496)

6 Wall-face (F4025 N)

7 Wall-face (F4040 N)

8 Stony clay (F4018/F4022)

9 Peat ash (F4016)

10 Fragmented schist (F4010)

11 Peaty clay and stones (F4017/F4019)

12 Peaty clay (F4030)

36 Clay and schist (F4085)

35 Clay and schist (F4072)

34 Clay and schist F(4073)

31 Schist and clay (F4067)

30 Schist and clay (F4063)

29 Silty clay (F4047)

28 Silty clay (F4046)

27 Silt, clay and schist (F4043)

26 Organic soils (F4039/F4042/F4045)

17 Clay and schist (F4074)

25 Clays (F4032/F4036/F4049)

16 Clay and schist (F4064)

5 Wall-face (F4040 S)

4 Wall-face (F4025 S)

15 Schist and gravel (F4033/F4035)

3 Wall-face (F4025 SS)

14 Silt (F4052)

13 Stony clay (F4034)

1 Schist, clay and stones (F438)

Offset 0.4m behind monastic enclosure wall

2 Iron-stained stones and clay (F448/F486)

Limit of excavation

Obscured

Conjectural

Bedrock

Extended monastic enclosure wall (F4025)

Primary monastic enclosure wall (F4040)

38 Peat (F4076/F4089)

37 Clay and schist (F4080)

North T3

Fig. 5.55 North–south section through both phases of the monastic enclosure wall, post-monastic building (F427) and underlying deposits. Location of elevation T2–T3 on Fig. 5.61.

South T2

the excavations 123

124 high island: excavation of an early medieval monastery

of the extended wall face (F4025 SS). East of Cell A, a small length of the wall was exposed within the excavation area (Fig. 5.56, no. 3). The wall was slightly narrower here than it was west of the cell, and the masonry on the northern face overlay two large boulders. The eastern side of the enclosure wall was exposed on the high ground east of Cell B (Figs 5.41, 5.46, no. 1, Fig. 5.47, no. 17). The wall (max. W: 1m) was two courses in height (0.3m) and composed mostly of small stones, of which only short lengths on the western (inner) face survived (Fig. 5.41). It was poorly preserved and no part of this side of the monastic enclosure wall was excavated. Two mural chambers were contained within the western wall of the monastic enclosure. The larger chamber was situated near the south-western end and the adjacent entrance, and the other smaller, less well preserved cham-

ber was close to the northern end. A third chamber may be located at the main south-east entrance, but due to the amount of overgrowth in this location, it is unclear whether it lies within or abuts the wall (Fig. 5.1; Section 4.4). The chamber at the south-west of the enclosure was investigated and the results suggest that it was constructed as an integral part of the primary wall (Fig. 5.1, Area 7). The chamber was entered via a short, low passage from inside the monastic enclosure. The passage (L: 1.1m, W: 0.6m, H: 0.8m) was orientated east–west. Two lintels were in place on the roof and space for a third was evident on the inner side (Fig. 5.57, no. 7, Fig. 5.60, no. 1). The chamber (L: 6.62m x W: 0.9m), positioned to the north of the entrance passage, was centrally located and it followed the curve of the enclosure wall (Fig. 5.1; Pl. 5.49). Originally, the complete chamber (H: 1.35m) was roofed with lintels, four of

N 7

1 10

6 8 9

3a 4 3 5 3

11 2

1 Peaty clay over fragmented schist and bedrock (F493/F497)

8 Peaty soil (F4054)

2 Fragmented schist and bedrock (F497)

9 Peaty soil (F4046)

3 Unexcavated stretch of monastic enclosure wall (F4040)

10 Possible surface (F4038/F4044)

3a North wall-face (F4040 N)

11 Clay and stones (F4082)

3b South wall-face (F4040 S)

Later structures

4 Schist, clay and gravel (F4033/F4035)

Charcoal

5 Silt (F4052)

Conjectural

6 Fragmented schist and clay (F4043)

Obscured

7 Fragmented schist and clay (F4055)

Limit of excavation

Fig. 5.56 The primary monastic enclosure wall (F4040) showing in situ fill layers and excavated deposits north and south of the wall.

the excavations 125

which survived in situ at the northern end where the side walls also remained intact (Figs 5.57–5.60). The drystone walls were composed of large blocks in the lower courses, with smaller, more elongated stones above (Fig. 5.60). The bases of the walls were set into the chamber floor, and traces of the cut were apparent at the north and north-west (Fig. 5.57, no. 1, F705a). A slight but distinct corbelling was evident on the chamber walls above 0.8m. As the walls rose, the chamber widened to 1.15m just below the corbelling and narrowed again to 0.75m beneath the lintels (Figs 5.58, 5.59). The floor of the chamber, which showed a distinct slope to the south of 0.34m, was formed by compact schist, which in places was not unlike a metalled surface (Fig. 5.58, no. 3, Fig. 5.59, no. 2, F705). A lens of charcoal recovered in the floor, which contained willow/poplar and a fragment of heather, represented fuel debris (Section 9.3, F705).

Phase 2: mid-11th to late 12th/early 13th century Phase 2 levels were exposed in the small area explored at the inner face of the northern monastic enclosure wall, beside Cell A. Organic soil was sealed by a layer of burnt debris (Fig. 5.54, no. 37, Fig. 5.61, nos. 11, 12, F4027, F484/F4079). The ash contained inclusions of peaty sod

throughout and was deepest on the lower ground, away from the wall. A wide range of species, predominantly hazel, heather and oak, was identified in the charcoal (Section 9.3, F484/F4079). Four samples were examined for their archaeobotanical contents and the results showed that there was a considerable variety of plant remains (Section 9.2, F484/F4079). These consisted of barley and oat grains, and also sorrel, goosefoot/orache and bladder campion. The large quantity of faunal remains consisted of fish and bird bones and other indeterminate fragments (Section 9.1, F484/F4079). The burnt debris was partially covered by a deposit of ash, charcoal, burnt peat flecks and organic silty soil (Fig. 5.54, no. 36, F478). The charcoal sample analysed was mainly of hazel and heather, with small quantities of oak, juniper, birch and ash (Section 9.3, F478). Food remains included barley and oat grains and four animal bones (Sections 9.1, 9.2, F478). Heat-shattered stones, two hones and a stone disc were recovered (Section 6.2, 478:2–478:4). The concentration of small finds, indicative of working tools, suggested that some sort of localised industrial activity was carried on in this location that possibly took advantage of the shelter afforded in the lee of the enclosure wall and Cell A. Char-

Covered by lintels W

4 2

6 V 0

1 Cut (F705a)

6 Brown clay (F706)

2 Ash (F702)

7 Chamber entrance

3 Silt (F703)

Lintels

4 Stones (F704)

Conjectural

5 Ash (F707)

Limit of excavation

Fig. 5.57 Plan of features within the wall chamber (F721). Elevation V–V3, see Fig. 5.60. Section W–W1, see Fig. 5.58; W2– W3, see Fig. 5.59.

Pl. 5.49 Wall chamber (F721) showing fragmented schist and collapse (F701) in situ, looking north (G. Scally).

126 high island: excavation of an early medieval monastery

West W

East W1

North W2

South W3

1 2

2 0

1 Sod (F700)

Lintel

2 Sod (F600)

Conjectural

2m 1 Sod (F700)

Lintels

2 Schist (F705)

Conjectural

3 Schist (F705)

Fig. 5.58 West–east section through the wall chamber (F721). Location of section W–W1 on Fig. 5.57.

Fig. 5.59 North–south section through the wall chamber (F721). Location of section W2–W3 on Fig. 5.57.

South V

North V1

West V1

East V2

Fig. 5.62b-North wall Fig. 5.62a-West wall

North V2

South V3

East V4

West V

1 Fig. 5.62c-East wall Fig. 5.62d-South wall 0

2m 1 Chamber entrance Lintels

Conjectural Limit of excavation

Fig. 5.60 Wall chamber (F721), interior elevations. Location of elevations on Fig. 5.57.

coal extracted from the burnt debris (F4079) provided a date of cal. AD 1042–1217 (UB-6454 (AMS) 888±30 BP). The position from which the charcoal was extracted lay directly beneath and was sealed by the southern wall face (F4025 SS) of the secondary enclosure wall and it provided a terminus post quem for its construction and also for the extended wall to the north-west (see below).

The extended monastic enclosure wall The secondary monastic enclosure wall was constructed in the same manner as the primary wall, with an earthen bank faced on both sides with stone. It was exposed for a distance of 7m on its northern side (Fig. 5.61, no.

2, F4025). Three wall faces belonged to the extended wall, one at the north and the two at the south (Fig. 5.54, nos. 7, 26, 38, Fig. 5.55, nos. 3, 4, 6, Fig. 5.61, nos. 2a–2c; Pl. 5.50). The northern wall face (F4025 N) was built on top of the primary one. At the south, the inner wall face (F4025 S) was also constructed on top of the earlier wall face. A new wall face (F4025 SS) that extended the width by 1m was built on the Phase 1 deposit which had accumulated outside the primary wall. An inherent structural logic was apparent in the manner of wall construction. The fill layers in the core of the primary wall were retained between the northern (F4025 N) and southern (F4025 S) wall faces, while the

the excavations 127

T 1

15b

G 2a

15a

8 6

2 10 2c U

T1 0

1 Peaty clay over fragmented schist (F493/F497)

12 Burnt debris (F484/F4079)

2 Unexcavated stretch of extended monastic enclosure wall (F4025)

13 Grey silty clay (F496)

2a Wall-face (F4025 N)

14 Grey silty clay (F4014)

2b Wall-face (F4025 S)

15a Buttress (F491a)

2c Wall-face (F4025 SS)

15b Buttress (F491b) 16 Water-collection and drainage channel (F8025)

3 Cell A 4 Annulus (F428)

The three wall-faces of the extended monastic wall (F4025)

5 Buttress (F453)

Later structure Annulus (F428)

6 N-W entrance; stepped passage (F437) 7 Stones and peaty soil (F4017/F4019)

Granite

8 Fragmented schist (F4010)

Upright stones

9 Ash, charcoal and burnt peat (F4016)

Conjectural

10 Stony clay layers (F4018/F4022)

Obscured

11 Organic soil (F4027)

Limit of excavation

Fig. 5.61 The extended monastic enclosure wall (F4025) showing in situ fill layers and excavated deposits north and south of the wall. Elevation T–T1, see Fig. 5.54; Elevation T2–T3, see Fig. 5.55; Elevation U–U1, see Fig. 5.62.

more substantial upper fill layers were contained by the two outer wall faces (F4025 N, F4025 SS). Unfortunately, the northern wall face (F4025 N) only survived for three short lengths, two within the excavated stretch of the wall and the other just outside, but within the excavation area (Fig. 5.61, no. 2a). The best preserved sections were at the eastern end (H: 1.3m), where the wall and the annulus around Cell A formed a continuous feature, and at the western end (H: 1.4m), where the wall had survived within the section face of the unexcavated part of the wall (Fig. 5.54, no. 2, Fig. 5.55, no. 6, F4025

N). Where the wall and the annulus formed a continuous feature, the masonry could clearly be seen to abut and obscure the earlier wall (Fig. 5.54, no. 7). The middle section of the wall was represented by a single, long, rectangular stone (broken into three pieces) (Fig. 5.61). Where the masonry survived, it was composed of quite substantial stones, with the exception of the western end, where the lower part of the wall had much smaller stones than in its upper courses. The inner wall face at the south comprised between one to three courses (Fig. 5.54, no. 26, Fig. 5.55, no. 4, F4025 S). This was the best-preserved of

128 high island: excavation of an early medieval monastery Pl. 5.50 The extended monastic enclosure wall (F4025) showing remains of wall-faces, with Cell A at centre rear, looking east (G. Scally).

the three wall faces that made up the extended wall. Its masonry consisted of moderately sized, regular, cut blocks of stone. The new southern wall face (H: 0.6–1m) had one to three courses surviving (Fig. 5.54, no. 38, Fig. 5.55, no. 3, F4025 SS). Much of the central portion was missing and what survived in situ was in a poor state of preservation. At its western end it was relatively well-preserved where it had been reinforced with a stone buttress (Fig. 5.61, no. 5, F453). At the other end, close to Cell A, the masonry (L: 2.2m) was composed of rectangular stones with their long axes set upright. Away from the wall, the stones had collapsed and were leaning at an angle. This stretch of the wall face had part of its masonry removed, thus permitting investigation of the stratigraphy beneath it. The radiocarbon date of cal. AD 1042–1217 (UB-6454, 888±30 BP) obtained from the charcoal in the burnt layer (F4079) came from beneath this stretch of wall. The lowermost fill, forming the core between the wall face at the north (F4025 N) and that to the south (F4025 S), comprised angular mica-schist stones mixed with silty clay and patches of darker, organic-based soil (depth: 0.25m), which overlay the primary wall core (Fig. 5.54, no. 6, F4019). At the interface between the enclosure wall, Cell A and its annulus, this layer had been banked up to abut the cell wall. In this location, the remains of the primary wall (F4040) were completely sealed. The stones in this layer had a distinctive reddish/orange or black stain, similar to those found on stones submerged in water around the lake edge. Several of them also appeared heatshattered, suggesting that they were redeposited hearth debris. The stones were covered in places by a peaty soil full of charcoal fragments, which have been identified as hazel and oak (Fig. 5.54, no. 5, Fig. 5.55, no. 11, Fig. 5.61, no. 7; Section 9.3, F4017). A substantial quantity of faunal

remains, consisting of bird bones where identifiable, were present (Section 9.1, F4017). Fawn-coloured fragmented schist, which had decayed to almost pure clay in its lowest levels, overlay the peaty soil (Fig. 5.54, no. 4, Fig. 5.55, no. 10, Fig. 5.61, no. 8, F4010). The schist was not uniformly laid out between the wall faces. At the eastern end, it abutted Cell A and formed a bank (depth: 0.6m) at the masonry of the annulus and the enclosure wall face (F4025 N). As this fill extended westwards along the length of the enclosure wall, it spread out across the full width and, close to its westernmost excavated end, it had a concave surface within which there was a deposit of peat ash containing oak and hazel charcoal and a small number of animal bones (Fig. 5.55, no. 9, Fig. 5.61, no. 9; Sections 9.1, 9.3, F4016). The overlying fill layers were retained by the new southern face of the enclosure wall (F4025 SS). Stony clays occupied the space between the two wall faces at the south (Fig. 5.55, no. 8, Fig. 5.61, no. 10, F4018, F4022). Stone artefacts were recovered, one of which was a hone or rubbing stone, while the others were two perforated stone objects (Section 6.2, 4022:3–4022:5). Close to Cell A, small stones and schist covered the stony clays and spilled over the inner wall face (F4025 S) to extend northwards to abut the bank of fragmented schist (F4010). The stones and schist contained charcoal of oak and hazel and faunal remains, predominantly of sheep/goat (Sections 9.1, 9.3, F4011, F4012). Above, small stones, many of which were iron-stained a deep red or orange colour, were mixed with dry, silty clay (Fig. 5.55, no. 2, F448, F486). Scraps, or lumps, of iron, possibly related to smithing activities, were recovered (Section 6.4, 448:1, 448:2). The fill also contained oak charcoal and faunal remains (Sections 9.1, 9.3, F448, F486). This was the most substantial and uni-

the excavations 129

form fill layer of the core (depth: 0.5m) within the upper part of the extended wall. At its base, there was a deposit of burnt debris, representing a hearth (1.6m x 0.8m and depth: 80mm) that contained small iron finds of nails, and scraps or lumps (Section 6.4, 457:1, 457:2, 457:4–457:7). The recovery of so many finds suggested that rudimentary iron-working was carried on in this location. Charcoal from the burnt debris was mainly of hazel and oak, with lesser amounts of birch, heather, ash, hawthorn and willow/poplar (Section 9.3, F457). Small quantities of fish and bird remains were identified (Section 9.1, F457). The surface of the fill of iron-stained stones and clay (F448, F486) was cut by a shallow depression (D: 50mm, depth: 30mm), which was possibly the remains of a post-hole (F4020) that held a post to support a wind-break around the hearth. Charcoal extracted from the burnt debris produced a date range of cal. AD 1027–1238 (UB-4521, 888±41 BP). Fragmented schist, small stones and grey clay formed the uppermost fill of the wall (Fig. 5.55, no. 1, F438). This layer did not completely seal the burnt debris (F457) as later disturbance in the centre of the wall extended to this level. Excavation at the western end of the northern monastic enclosure wall uncovered the entrance, which can be dated to the latest remodelling of the wall in the mid- to late 12th century. It was revealed that this was a stepped passage built over the enclosure wall, unlike the other three entrances that ran through the wall (Fig. 5.61, no. 6, F437; Pl. 5.51). The north-west entrance was widest in the centre (0.9m) and it narrowed to 0.4m between the upright jambs at the northern (external) side. It was built with a pronounced skew along its length, which was most apparent at the northern side, a feature no doubt incorporated to deflect the effects of the on-shore winds. The passage, 4.4m in length, measured 1m longer at the north than the width of the monastic enclosure wall. This was achieved by the addition of a triangular area of stones (2m x 1m) that formed a buttress on the eastern side of the entrance and by a similar buttress built on the western side (Fig. 5.61, nos. 15a, 15b, Fig. 5.62, no. 6, F491a, F491b). The western buttress was only partially exposed as much of it lay outside the excavation area. Six steps had survived on entering the passage from the south. Beyond the sixth step, a thin slab set on edge (H: 0.22m) provided a level area in the centre, the highest part of the passage (Fig. 5.62; Pl. 5.51). This part of the passage was 0.9m higher than the paved surface inside the enclosure (Fig. 5.62, no. 5, F422). The level area was paved with four mica-schist slabs, which were demarcated to the north by a stone set on edge that extended across the width of the passage and formed a threshold. Beyond the threshold stone to the north, the passage continued between the buttresses where the floor had poorly laid slabs that sloped progressively downwards to the con-

temporary silty clay surface outside (Fig. 5.61, no. 14, Fig. 5.62, no. 8, F4014). The entrance on the northern side was formed by two pink-grained granite boulders (both H: 0.55m), which had been cut from the same glacial erratic, a stone commonly found on the island (Fig. 5.61, G). The wall of the passage on the eastern side was formed by the masonry of the enclosure wall and the buttress (Fig. 5.61, nos. 2, 15a, F491a, F4025; Pl. 5.51). On the western side, the best-preserved stretch was in the central, elevated part, where six upright slabs formed a well-defined face (Fig. 5.62; Pl. 5.51). Beyond to the north, the wall was formed by masonry of the buttress (F491b). To the south, no sidestones had survived, however it is probable that this stretch of the passage would originally have had upright stones like those in its central area. A grey, silty clay surface abutted the external wall face (F4025 N) and covered the area outside to the north of the enclosure wall (Fig. 5.61, nos. 13, 14, Fig. 5.62, no. 8, F496, F4014). The clay was moderately stony with inclusions of minute fragments or globules of iron throughout, the result of natural leaching from the soil, which gave it a distinctive appearance. Oak was identified in the charcoal sample analysed (Section 9.3, F496). With the exception of a shallow spread of ash lying on the surface, the dearth of occupation-related material suggested continuous negligible activity in this location. A buttress was built against the south-west corner of the monastic enclosure wall (Fig. 5.41, no. 21, F79). It was

Pl. 5.51 Stepped passage (F437) over the extended monastic enclosure wall (F4025), looking north (G. Scally).

130 high island: excavation of an early medieval monastery

South U

North U1 1 2 6 Q

7 8

10 11

1 Sod (F400)

6 Buttress (F491b)

11 Clay and schist (F4066)

2 Rubble and peat (F429/F441;F402/F411/F426)

7 Stony clay (F469)

12 Peat (F4089)

3 Rubble (F460)

8 Silty clay (F4014)

Steps

Upright stones

4 Clay and stone (F474)

9 Brown clay (F4049)

Stone-lined wall of passage (F437)

Limit of excavation

5 Paving (F422)

10 Silty clay (F4055)

Paving

Bedrock Q

Quartz

Fig. 5.62 Elevation of west side of stepped passage (F437) through the monastic enclosure wall. Location of elevation on Fig. 5.61.

composed of a substantial boulder that had been pushed up against the outside of the wall and small, roughly hewn blocks placed alongside it. These were set into a bedding of gravel and small stones within a depression in the paved surface (Section 5.4.1, F80). The ground in this area had clearly subsided, thus providing the reason for the buttress, a structure that bore all the hallmarks of being built in haste.

Deposits within the mural chamber (F721) Excavation uncovered little evidence for the function of the chamber in the south-west of the monastic enclosure (Fig. 5.1; Pl. 5.49). Post-excavation analysis included an extensive programme of sieving samples from the deposits, yet this resulted in only minimal evidence for human activity. A deposit of clay mixed with oxidised ash and oak charcoal overlay the Phase 1 schist floor (F705) and abutted the end wall of the chamber at the south (Fig. 5.57, no. 5; Section 9.3, F707). This was probably the remains of a small hearth that was set within a shallow depression (depth: 20–100 mm) in the floor. Gritty, brown clay sealed the ash, and its localised extent suggested that it was thrown down to extinguish the fire (Fig. 5.57, no. 6, F706). A layer (depth: 20–50 mm) of fine silt, in which heather charcoal and a single grain of barley were identified, covered the remaining area of the chamber floor, extending to the entrance passage (Fig. 5.57, no. 3; Sections 9.2, 9.3, F703). A scatter of stones, possibly a rudimentary surface, lay on top of the silt in the south of the chamber (Fig. 5.57, no. 4, F704). Two small spreads of ash containing fish bones were beside the chamber wall overlying the stones (Fig. 5.57, no. 2; Section 9.1, F702).

Phase 3: late 12th/early 13th to mid-15th century The Phase 1 ground surface (F493) outside the monastic enclosure at the north-east was covered by clay mixed with fragmented schist (Section 9.3, F488). This layer was shallowest on the high ground, but as it extended down-slope it deepened to 0.2m where it abutted the annulus around Cell A. It covered the sod fill (F489) of the gully (F487) that formed the northern part of the water-collection and drainage channel (F8025), indicating that by the time the stony clay was laid down, the channel was already out of use and the monastery abandoned (Fig. 5.41, no. 7). Layers of mid-brown clay, stones and silt covered the ground above and also to both sides of the disused channel (Fig. 5.51, nos. 1, 2, F458, F459). The silty clay (F458) abutted the annulus around Cell A and extended westwards, where it formed a uniform layer (depth: 50mm) with a scattering of small schist stones that abutted the enclosure wall (Fig. 5.51, no. 24, F479). Charcoal of hazel, oak and ash and a single animal bone fragment were in the clay (Sections 9.1, 9.3, F479). A spread of ash (2m in extent) overlay the surfaces of both deposits and was partially sealed by the wall of the sub-rectangular building (Fig. 5.51, no. 25, F472; Section 5.5). Charcoal, mainly of heather and hazel and lesser amounts of oak and ash, a grain of barley and faunal remains were in this burnt debris (Sections 9.1–9.3, F472). In the north-west of the excavated area, a scatter of clay and stones (F469) was identified that was contemporary with the silty clay levels. The other deposits in this area relate to the sub-rectangular structure (Section 5.5).

the excavations 131

Phase 4: mid-15th to late 20th century The monastic enclosure wall was obscured on both sides by areas of collapsed rubble interspersed with layers of peat and fragmented schist (F436; Pl. 5.43). Yet more rubble, schist and grass-covered sod lay on top and across the wall. An iron strap and a 1922 penny of George V were recovered from these layers (Section 6.4, 401:1, 401:2). A localised deposit outside the entrance to the chamber in the enclosure wall at the south-west contained charcoal (Section 9.3, F708). Inside, fragmented schist and stone (up to 0.6m in depth) covered the central and southern parts of the chamber (Pl. 5.51). The deposit was completely sterile and appeared to have been derived almost exclusively from the natural process of weathering and decay of the overlying rubble, which had collapsed from the walls and roof. The intact northern end of the chamber and the rubble were covered by sod (Fig. 5.59, no. 1; Section 9.1, F700).

5.4.3 Cell A Phase 2: mid-11th to late 12th/early 13th century Cell A, the beehive hut and its annulus, was built within the thickness of the northern side of the monastic enclosure wall (Fig. 5.1, Areas 4, 5). The entire area within the cell was excavated and features pre-dating the cell were uncovered. Naturally deposited boulder clays were exposed at the lowermost excavated level. These were covered in places by stones above which a shallow layer of brown silty clay had accumulated. Sporadic outcrops of shattered bedrock pierced the surface. A deposit of stiff, shale clay (depth: 0.2m) covered the entire area beneath the cell interior, which continued in all directions and was sealed by the wall faces (Fig. 5.63 (a), no. 1, F517, Fig. 5.64, no. 2). It contained frequent flakes or chippings of micaschist, charcoal fragments and a scatter of small quartz pebbles. This clay may have been introduced to provide a bedding for the cell walls. The cell was largely complete, but poorly preserved (Pl. 5.52). Internally, it was rectangular in plan (2.2m east–west x 1.85m north–south) with angular corners (Fig. 5.65, no. 1). Externally, the cell was close to circular in plan (4.06m east–west x 3.65m north–south), though subsidence of some parts of the external walls may have somewhat altered its original shape. It was surrounded on two sides by the annulus. On the western side, the annulus was a continuation of the monastic enclosure wall and on the eastern side, it formed part of the side wall of the watercollection and drainage channel that encircled the entire eastern side of the monastery (Fig. 5.65, nos. 2, 3; Pls 5.52, 5.53). The cell was entered via a splayed doorway, which was positioned slightly west of centre in the south wall (Fig. 5.65; Pl. 5.52). The doorway (H: 1.08m) was roofed with

lintels, of which only the innermost had survived in situ (Fig. 5.66c). The paved entrance was wider at the external jambs (0.8m) and it narrowed at the internal jambs (0.54m). A granite stone set on edge formed the threshold (H: 0.12m) at the inner jambs (Fig. 5.66 (c), (e), (f )). A shallow channel (depth: 15mm x W: 60mm) carved on the surface of a large slab within the passage floor probably functioned as a rudimentary drainage channel. The internal walls of the cell (H: 1.4–2.1m) were corbelled above 1m (Fig. 5.66). The south wall was the only one where it was possible to measure the width at paving level (1.25m), and it is estimated that the other walls were of similar width. Internally, the masonry of the lower courses was composed of large, predominantly square-shaped boulders, and above there were smaller, rectangular-shaped stones (Fig. 5.66). Most were of micaschist, but granite blocks were also included. Two stones, one of granite in the south-west corner and the other of quartz in the north-west corner, lay partially beneath the cell walls and may have been used to mark out the position of the cell when it was being built. Two niches survive within the cell, one in the east wall (W: 0.45m x H: 0.15m, depth: 0.75m) and the other in the west wall (W: 0.33m x H: 0.17m, depth: 0.55m) (Fig. 5.66 (b), (d), no. 1). The rear walls of both features were formed by the inner side of the exterior facing stones and the bases were the compacted clay in the core of the cell walls. Outside the cell, the bases of the walls on the northern and eastern sides were obscured by the annulus and, therefore, were only visible in limited areas to the south and west. On the southern side, the external wall face was built on paving (Fig. 5.64, no. 4, F490). At the west, the external wall face overlay the schist and gravel, one of the fill deposits of the primary monastic wall (Fig. 5.65, no. 4, F4033, F4035). The facing stones from the enclosure wall had also been left in situ and the cell wall was constructed on top of them (Fig. 5.54, nos. 1, 8, Fig. 5.65, nos. 1, 3.1, F4040 N). The construction of Cell A appeared to have required the dismantling and rebuilding of the monastic enclosure wall, and a terminus post quem of the mid-11th century for this work was obtained by radiocarbon analysis (Section 5.4.2). Externally, the masonry of the cell walls (approximate H: 2m), where visible at the west and south, was rough and irregular in outline instead of having smooth, curved faces (Fig. 5.64; Pl. 5.52). This poor workmanship suggested that the cell was built in haste or that the mason was of limited ability. The annulus, which provided additional structural support and acted as weatherproofing, was constructed on the surface of fragmented schist (Fig. 5.65, no. 6, F493). Fragments of burnt peat were scattered throughout, suggesting that this surface may have sustained a superficial peat growth that was cleared prior to construction. The

132 high island: excavation of an early medieval monastery

North wall of Cell A

7 4

6 8

1 G

(a)

Cell entrance

(b)

1 Clay (F517)

8 Paved floor (F510)

2 Post-hole (F522) with packing stones

9 Cross-slab 32

3 Post-hole (F523)

Clay (F512)

4 Burnt debris (F515)

Ash

5 Stake-hole (F525)

Charcoal

6 Stake-hole (F516)

7 Cut (F519)

Granite Obscured

Fig. 5.63 Cell A: (a) pre-floor deposits; (b) paved floor (F510).

Pl. 5.52 Cell A after excavation, looking north (G. Scally).

Cell entrance

the excavations 133

Pl. 5.53 Masonry side wall of the water-collection and drainage channel (F8025) and overlying annulus (F428) of Cell A (left) and the monastic enclosure wall (right), looking north (G. Scally).

North Y

annulus (approximate W: 0.6m and H: 0.9m) comprised one to four courses of roughly faced drystone masonry (Fig. 5.64, no. 1, F428, Fig. 5.65; Pls 5.52, 5.53). It had a core of small stones and dry, friable clay that was faced on its external side with stones, many of which were large, squared blocks. At the eastern end, it flared outwards (to W: 1.6m) where it overlay the side wall of the water-collection and drainage channel, which probably belonged to the same construction phase (Fig. 5.65, F8025). At the northwest, the annulus continued westwards away from the cell to form part of the extended monastic enclosure wall (Fig. 5.61, nos. 2a, 4, Fig. 5.63, nos. 2, 3.2, F428, F4025 N). The masonry of the annulus and the extended monastic enclosure wall was bonded together, suggesting that both features were contemporaneous with one another. Patches of redeposited boulder clay were revealed inside the cell, especially at the west and north walls, above the clay layer (F517) on which the cell had been built (Fig. 5.63a, F512). These clay patches, which were partially sealed beneath the walls, were light grey in colour and extremely pure and fine-grained in texture. This clay was used as bedding or as a rudimentary mortar in other structures at the monastery (Section 9.4, Appendix B). Burnt debris (depth: 0.12m) covered a large portion of the cell interior (Fig. 5.63 (a), no. 4, F515). The debris comprised stratified layers of charcoal (depth: 20mm), yellow ash (depth: 40mm) and oxidised red ash (depth: 50–60 mm) set into a depression within the clay (F517). The charcoal was of oak, birch and hazel, and barley and oat grains were identified (Sections 9.2, 9.3, F515). A deposit of charcoal-

South Y1

Cell A

1 4

8 9

G 2

5m 11 Paving (F29) 1 Annulus (F428)

6 Cross base (F415)

Paving

2 Clay (F517)

7 Side wall of passage (F8040)

Bedrock

3 Paved floor (F510)

8 Blocking stone (F8055) of passage

4 Paving (F490)

9 Passage (F414/F8040)

Upright stones

5 Paving (F422)

10 Church enclosure wall

Conjectural

Granite

Fig. 5.64 North–south section through Cell A and surrounding features. Location of section Y–Y1 on Fig. 5.41.

134 high island: excavation of an early medieval monastery

2 Annulus (F428)

3.1

1 Cell A

3.2

Water collection and drainage channel (F8025)

3 Monastic enclosure wall; see Fig. 5.56 3.1 Masonry (F4040 N) of primary wall; see Fig. 5.56 3.2 Wall-face (F4025 N) 3.3 Wall-face (F4025 S) 3.4 Wall-face (F4025 SS)

3.3

4 Schist and gravel (F4033/F4035)

5 Paving (F490)

6 Fragmented schist (F493)

3.4

Upright stones X4 5

X7 Cell entrance

Conjectural

Obscured Limit of excavation

Fig. 5.65 Cell A and annulus (F428) and surrounding features. Elevations X–X7, see Fig. 5.66 (a)–(f).

flecked clay (depth: 60mm) covered the burnt debris but did not extend beyond it, which would suggest that the clay was thrown down to extinguish the fire. Two stakeholes (depth: 70mm) were located to the west of the burnt debris (Fig. 5.63 (a), nos. 5, 6, F516, F525). Radiocarbon analysis of charcoal extracted from the burnt debris produced a date range of cal. AD 1176–1276 (UB-6453, 801±34 BP), indicating that the earliest occupation of the cell was in the late 12th century. A post-hole was uncovered on either side of the threshold of the cell doorway (Fig. 5.63 (a), nos. 2, 3, F522, F523). The post-hole on the eastern side (D: 0.15m, depth: 0.2m) was filled with clay, ash and small rectangular stones that may have been packing stones that had fallen out of position. The post-hole on the western side was more substantial (D: 0.36m x depth: 0.14m). It was filled with a similar deposit of clay and ash and by three slender stones on edge, which had been wedged tightly into it. It is probable that these stones had been used to secure a wooden doorpost in position. A shallow, ovoid cut (depth: 40mm) filled with a similar deposit of clay, ash and stone, the purpose of which is unknown, abutted the western side wall of the cell (Fig. 5.63 (a), no. 7, F519). The cell had a paved floor that was preserved over most of its interior (Fig. 5.63 (b), no. 8, Fig. 5.64, no. 3). Approximately 30 slabs were set onto a lens of grey clay and a further seven slabs lay on top (Fig. 5.67, no. 1, F510). The slabs formed a rough and uneven surface as they abutted one another and some overlapped. None of them showed any evidence of wear. Two decorated cross-slabs were reused in the paving in the south-east corner. One of the cross-slabs, set into a lens of sand, was lying with its carved

face upwards (Fig. 5.63 (b), no. 9, Fig. 5.67, no. 3; Crossslab 32). The other cross-slab was found in close proximity at a slightly lower level, with its decorated face downwards (Section 7.1, Cross-slab 33).

Phase 3: late 12th/early 13th to mid-15th century Burnt debris (1m2 and depth: 0.25m), the remains of a small hearth, lay on the surface of the paved floor in the cell (Fig. 5.67, no. 2, F509). The debris comprised brightly coloured ash in which there were stratified lenses of charcoal, mainly of hazel (Section 9.3, F509). Almost 600 charred cereal grains, most of which were of barley, were recovered (Section 9.2, F509). The seeds produced a date range of cal. AD 1287–1424 (UB-4988, 580±40 BP), indicating that Cell A was reoccupied in the later medieval period, after the religious community had left. An accumulation of clay sod covered the burnt debris and the paved floor. A small quantity of faunal remains, consisting of sheep/goat, cattle and bird, were in the sod layer (Section 9.1, F508). The deposit also contained occasional limpet shells and fish bones that were possibly food debris introduced by nesting birds. Approximately fourteen slabs of schist overlay the sod to form a rudimentary surface in the south-east corner of the cell. The slabs were not well laid and they overlapped one another, resulting in a rough and uneven surface. This surface had ash and stratified lenses of charred peat (depth: 50mm) on top of it. The burnt material covered the entire interior of the cell with the exception of a small area in the north-east corner, where there were further paving slabs.

South X4

(e)

East X1

(b)

North X5

(f)

North X6

South X2

South X7

(c)

West X3

Quartz

Limit of excavation

Conjectural

Obscured

Granite G

Paving

Lintel

1 Possible recess

(d)

South X3

Fig. 5.66 Cell A interior elevations: (a) north; (b) east; (c) south; (d) west; (e) west entrance passage; (f) east entrance passage. Location of elevations on Fig. 5.65.

(a)

West X

North X1

East X2

North X

the excavations 135

136â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Phase 4: mid-15th to late 20th century Layers of stone debris, separated by deposits of soil and fragmented schist, accumulated within the cell. The uppermost stone layer, which included very large blocks (L: over 1m) represented the collapsed superstructure of the cell.

5.5 The sub-rectangular building outside the monastic enclosure Phase 3: late 12th/early 13th to mid-15th century A sub-rectangular building was erected against the annulus of Cell A and the northern wall of the monastic enclosure, close to its western end (Fig. 5.1, Areas 4, 5, Fig. 5.50, Fig. 5.68, no. 11, F427; Pl. 5.54). The building was constructed after the water-collection and drainage channel had fallen out of use as its east wall was built on the deposits that had accumulated above the channel (Section 5.4.1, Phase 3, F458, F479). The foundation trench in which the east wall was constructed cut into these clay layers (Fig. 5.68, no. 15, F470). A post-hole (D: 0.1m) located close to the northern end of the trench was cut from the same level (Fig. 5.68, no. 14, F471). By the time this building was erected, the water-collection and drainage channel had ceased to function, suggesting that the monastery was in serious decline at this time. The building (enclosing 15m2) consisted of east and west side walls, both of which were slightly splayed, and a rear (north) wall (Fig. 5.68, no. 11, F427). Its front (south) wall was formed by the monastic enclosure wall. The walls (approximate W: 1m) survived from one to three courses in height (0.4m max.) and were of varying types of construction (Pls 5.54, 5.55). The east wall, which abutted the annulus of Cell A, was faced on both sides with large, rectangular stones and it had a core of clay with a small

quantity of stones. The north and west walls were built with smaller facing stones and the core consisted of rubble and a small quantity of soil. The end of the west wall abutted the buttress that had been built against the monastic enclosure (Fig. 5.68, no. 16, F491a). The external face of the west wall had been reinforced with large boulders, several of which had been pushed up against it (Fig. 5.68, no. 17, F427a; Pl. 5.55). The boulders were roughly hewn and the most southerly of them had been placed in front of the stepped passage (F437) over the monastic enclosure

1 Uppermost level of paving (F510) 2 Burnt debris (F509)

3 Cross-slab 32 G

Ash

Granite

Fig. 5.67â&#x20AC;&#x201A; Cell A showing uppermost level of paving (F510) with burnt debris (F509) in situ. Pl. 5.54â&#x20AC;&#x201A; Sub-rectangular building (F427) abutting the annulus (F428) of Cell A and the northern side of the monastic enclosure wall, looking south-east (G. Scally).

the excavations 137

wall and had partly blocked its entrance. The bases of the boulders were set into the stony clay on which the west and north walls of the building were erected (Fig. 5.68, no. 18, F469). The building was entered from the monastic enclosure to the south via a passage that was broken through the enclosure wall (Fig. 5.68, no. 19, F4028; Pls 5.56, 5.57). The passage (W: 0.5m) was composed of two parts, a paved and partly stepped stretch (L: 2.7m) on its southern side and a further unpaved stretch (L: 2.3m) to the north. Slabs from the Phase 2 paved surface in the enclosure (F422) were relaid with their long axes east–west to partly overlap one another, creating two or three steps leading up to the monastic enclosure wall (Fig. 5.68, Slabs A–C; Pl. 5.56). Two more steps lay within

the dismantled section of the wall (Fig. 5.68, Slabs D, E). The uppermost slab was secured in position by carefully placed stones set on edge. Space for a third step in between was visible where the masonry of the enclosure wall was removed. The height difference noted between the top of the steps to the paved surface (F422) below was 0.45m. Beyond the steps, the passage was recognised by a linear area of compacted and depressed ground, which was the fill in the core of the enclosure wall (Section 5.4.2, Phase 2, F448, F486). The passage emerged from the wall into the sub-rectangular building slightly east of centre. The passage did not have side walls. The only delimiting feature was a low bank of stone and soil with patches of possibly wind-blown ash, which was piled up on the eastern side and extending as far as Cell A (Fig.

13 14

12 15 7

10 2

G G 16

8 3

4 20 E D

9 X

C B

A 5

1 Cell A

10 Silty clay (F479)

20 Bank of stone and soil (F4002)

2 Annulus (F428)

11 Building (F427)

21 Ash (F4005)

3 Extended monastic enclosure wall (F4025) (deposit F448/F486)

12 Post-holes (F495a/b)

4 Stepped passage (F437) 5 Paving (F422) 6 Clay and stones (F459) 7 Silty clay (F458) 8 Silty clay (F463) 9 Ash (F445)

13 Ash (F472)

Route of Passage G

14 Post-hole (F471) 15 Construction trench (F470) 16 Buttress (F491a) 17 Boulders (F427a) 18 Stony clay (F469) 19 Passage (F4028)

Granite Ash

A-E Steps X

Sidestone from extended monastic wall-face (F4025/SS) Conjectural Limit of excavation

Fig. 5.68 Sub-rectangular building (F427) with access passage (F4028) through monastic enclosure wall, Phase 3.

138â&#x20AC;&#x192; high island: excavation of an early medieval monastery

5.68, nos. 20, 21, F4002, F4005). A large rectangular stone found lying under the south-west corner of the bank may originally have stood upright in the wall face of the monastic enclosure wall (Fig. 5.68, Slab X; Pl. 5.57). When the passage was created, the stone was laid flat partly on top of the slabs forming the passage steps. Inside the building, two post-holes had been cut into the clay on which it had been constructed (Fig. 5.68, no. 12, F495a, F495b). The post-holes were not equi-distant from the north wall. Due to their location and small size (D: less than 0.1m), it is unlikely that they held posts that were strong enough to support the superstructure for this building. No other features were identified within the building.

Phase 4: mid-15th to late 20th century The abandoned interior of the sub-rectangular building was covered by a deposit of grey silty clay mixed with rubble at its surface (Sections 9.1, 9.3, F452). The rubble probably originated from the collapse of the structure as it was concentrated close to the north and east walls. A shallow hearth of peat ash lay on the surface, near the east wall (Sections 9.1, 9.3, F461). Another, possibly wind-blown,

scatter of ash lay in the north-west of the building. The entrance passage was sealed with grey silty clay and small stones. As this deposit was indistinguishable from the upper fill (F438) within the core of the monastic enclosure wall, it is most likely that it represented slippage from the fill on both sides of the passage. Outside the building, a scatter of rubble lay to the north-west and north-east. On the north-eastern side, the rubble stopped short of the annulus around Cell A, suggesting that this area around the cell was cleared at some point. The area on the higher ground to the north-east was covered by almost stone-free clay, which had possibly been washed down. The overlying deposits inside the building and outside to the north and south were the dry, schist clay and the dense layers of stone rubble interspersed with peat that also covered the north and north-east of the monastic enclosure (Section 5.4.1, F425, F429, F436, F436a). Another deposit of schist and rubble (F402) completely concealed the presence of the sub-rectangular building. The entire area was covered with a layer of peaty soil.

the excavations 139

Pl. 5.55 Boulders (F427a) abutting external west wall of the sub-rectangular building (F427), looking south (G. Scally).

Pl. 5.56 The paved surface (F422) leading towards the passage (F4028) through the monastic enclosure to the sub-rectangular building (F427), looking west (G. Scally).

Pl. 5.57 Stepped paving slabs of passage (F4028), looking north. Stone X may originally have been a facing stone in the wall (F4025 SS) of the extended monastic enclosure wall (G. Scally).

140

6 the finds Claire Cotter

6.1 Introduction The small finds assemblage from High Island is made up of approximately 150 items. Almost a third are made from stone and comprise coarse stone tools, a small number of flint/chert pieces, and some items that can be categorised as grave goods and artefacts linked with pilgrimage (inscribed pebbles and porphyry fragments). The remaining objects are three glass beads, a bone pin, a clay crucible, two medieval coins, a possible book clasp, an iron assemblage (mainly nails) and some miscellaneous early modern items. With the exception of the inscribed pebbles and the coins, the artefacts are generally in very poor condition. Most were probably discarded as a consequence of being worn out or broken. Not surprisingly, given the coastal location, the bulk of the metal items are also extremely corroded. A number of architectural fragments were recovered during the excavations. Most were carved from the local coarse-grained mica schist. The pieces include the head and sill of the narrow, round-headed, east window of the church (the window was still in situ when Wakeman visited the site in the late 19th century), and a well-made holy water stoup. The majority of the small finds from High Island would have served the everyday needs of the monastic community. Most can be linked to domestic/craft activities, such as iron and wood-working, food-processing and fishing. Given the fact that the excavated areas included ecclesiastical, domestic and working spaces, the assemblage is quite restricted in range. Poor preservation conditions, especially for bone and metal, may partly account for the absence of some items; pins are particularly poorly represented, for example. The recovery of only one quernstone (compared to eight from Church Island, Co. Kerry (O’Kelly 1958, 105–6) and seven from the monastery at Reask, Co. Kerry (Fanning 1981, 133–5)) might be explained by the presence of a horizontal mill on the island. As only a single quernstone was also recovered at Illaunloughan (White Marshall and Walsh 2005, 189), however, other factors (such as recycling of stone, limits of the excavation) could account for the deficit.

The proportion of broken, worn and damaged items in the assemblage is striking. Even hard-wearing tools, such as those made on coarse cobbles, are almost invariably broken. Apart from the coins, and possibly the glass beads, there is little in the domestic assemblage that can be considered as ‘casual’ loss. Most items seem to have been discarded, and the overall impression is of very poor material circumstances, with goods being curated until they were no longer functional. This is well illustrated by the condition of the simple small bone pin (208:4), which continued to be used even after the head had splintered. The small size of the flint piece, 830:1, suggests that commodity too was carefully curated.

Sources of raw material The island lacks any significant stretches of accessible shoreline and the origin of the cobbles is likely to be the glacial till that occurs throughout the island, as a thin layer, or occasionally (as above the south landing) in much thicker deposits. Cobbles can be readily picked up where the clay cliffs have been eroded back either by weathering or by wave action. For the monks, however, the most convenient source of raw material was directly adjacent to the monastery, around the margins of the lake, where a variety of cobbles and water-rolled stones could be easily collected. The excavation assemblage is too small to carry out any meaningful analysis on the selection of particular lithologies for specific tool function. As with other assemblages, fine-grained stone, such as sandstone, seems to have been selected for light abrasive tasks (honing and sharpening), while harder or coarser grained rocks, such as granite and quartzite, were preferred for more heavy duty pounding and grinding. In contrast to the everyday functional items, the majority of the objects categorised as grave goods and/or items linked with pilgrimage were made from non-native materials. Some of the items, e.g. the hone (208:1) and weight (208:2) associated with burial (F208), could have been brought to the island in finished form. X-ray fluorescence analysis of the hone (208:1) and the three disc-shaped nodules (4:1, 4:4, 4:5) (see this volume, p.159) indicated a high (25–38%) iron content derived from precipitated iron car-

the finds 141

bonate. Spheroidal or discoidal nodules of carbonate or ironstone have been recorded in the Namurian mudstones and sandstones of the Slieve Elva area, near the southwestern edge of the Burren (Simms 2001). Given the lithology, and the distribution of other recorded examples of these nodules in the west of Ireland, the most likely origin for the High Island stones would therefore seem to be secondary weathered deposits either in the Burren or (less likely) on the Aran Islands. Two of the larger carved stones from High Island (Cross-slabs 6 and 7, see Section 7) are made from imported carboniferous limestone, probably brought from the same region. Whether the nodules were carved on the island or elsewhere is impossible to say with any degree of certainty, but the presence of an undecorated stone may make the former more likely. The most exotic items in the lithic collection are undoubtedly the three pieces of recycled porphyry (38:1, 4030:1 and 4030:3), the likely origin of which is the ancient quarries at Laconia, in the south-eastern part of the Peloponnese region in Greece.

Grave goods Burial F208 (in Grave 9) was accompanied by two stone objects: a hone (208:1) placed on the right shoulder, and a line sinker (208:2) placed on the left shoulder. Both show some degree of wear and very likely represent the personal possessions of the deceased. A small, extremely worn bone pin (208:3) found in the pelvic area suggests either that the body was wrapped in a shroud or winding sheet, or alternatively that an undergarment was present. The body was that of a male aged between twenty and 30 years (O’Donnabhain, this volume, p.232). The burial has been dated to the period cal. AD 1163–1230 (UB-4000, 849+16 BP). It is the latest of the six interments from the site and although it may have been inserted in the church while the building was still in use, this cannot be confirmed (G. Scally, pers. comm.). Stone weights have occasionally been found in grave soil (e.g. at Skeam West, Co. Cork; Cotter 1995), but there is no certain Irish instance of their being deliberately included with a burial. Indeed, the inclusion of grave goods of any type in medieval Christian burials would be most unusual (E. O’Brien, pers. comm.). Of the 1,500 or so skeletons uncovered at the 5th-/6th–12th-century cemetery excavated at Cabinteely, Co. Dublin (Conway 2003), for example, only a handful included any personal items. The placing of objects on the shoulders of the body also seems to be extremely rare and it has not proved possible to find a contemporary Irish parallel. A review of 8,000 medieval monastic burials in Britain indicates that the practice was virtually unknown there also (Museum of London 2005). A mature adult male from St James’s Priory, Bristol, who was buried with a jet pendant and had

two folded coins placed on his shoulders, is probably the closest contemporary parallel (Jackson 2006, 76, no. 64). This grave was dated to the 12th century. Grave goods, including hones, are a feature of burials of Viking type. Hones are fairly common finds in both male and female graves, although like most grave-goods they are rarer in ‘Viking’ graves in Britain and Ireland than in Scandinavia (S. Harrison, pers. comm.). Harrison’s study indicates that full size or miniature hones occur in c. 8% of the recorded British/Irish corpus of ‘Viking’ burials. Most of these burials are significantly earlier in date than the High Island example, however. Harrison (2001, 74) suggests that all the known Irish ‘Viking’ burials may have been deposited within a 50–100-year period between c. 840 and 950. The majority (80% of the c. 71–76 examples listed by Harrison) are located in the east of the country. A richly furnished burial at Eyrephort, Co. Galway (Raftery 1960, 3), is the only recorded ‘Viking’ burial west of the Shannon. Whether or not it was associated with a local settlement is debatable (Harrison 2001, 63; Sheehan 1988), but, even if this could be shown to be the case, the fact that it predates the High Island burial by about 200 years rules out making any direct connection between the two.

Items linked with pilgrimage and/or other ecclesiastical activity In addition to everyday goods, there is a significant group of items that can be linked to the spiritual activities of the monks and/or of visiting pilgrims. The porphyry pieces would have been treasured as relics. Piece 38:1, which was found with burial F40 (Grave 4), appears to be the only definite instance on record where a piece of porfido verde antico can be directly linked to an individual burial. This may be an early occurrence of the medieval practice of burials being accompanied by pilgrim badges and mementos, such as scallop shells. We can only speculate as to the route by which the pieces of porphyry came to High Island. As relics, their primary association would have been with Rome, and they may have been brought to the island directly by one or more monks returning from pilgrimage there. Other more indirect means are also possible: the relics could have been dispersed from one of the larger monasteries (e.g. Armagh or Clonmacnoise); acquired from Scandinavian/ Hiberno-Norse traders; or (perhaps a less likely scenario) gifted to the monastic community by pilgrims visiting the island. Based on the radiocarbon evidence, all three High Island pieces may have been brought to the island before the end of the first millennium AD. This date bracket is slightly earlier than the mid-11th century dating suggested by Lynn (1984) as the core period when porphyry was being imported into Ireland. If the late 7th/8th century dating for the Clonmacnoise piece (H. King, pers. comm.), and the mid- to late 12th-century dates suggested for the Water-

142 high island: excavation of an early medieval monastery Pl. 6.1 St John’s Well, Killone, Co. Clare. Drawing by T.J. Westropp, 1896 (Westropp 1911).

ford finds (McCutcheon 1997, 432) are taken into account, however, it would seem that porphyry was in circulation both before and after that core period. The inscribed pebbles (4:1, 4:4, 4:5) belong to a western tradition of ‘altar’ stones, many of which have unfortunately disappeared over the last 100 years. White Marshall and Rourke (2000) suggest that the High Island stones were made by pilgrims who left them on the graves as offerings, as they did white quartz stones, and the inscription on the Temple Brecan nodule would lend support to that interpretation. According to Westropp (1911), the Co. Clare altar stones were frequently used as a ‘rude rosary to keep count of prayers and rounds’ (Pl. 6.1). Similar practices may also have been fairly widespread in earlier times, although, very possibly, this was not the primary purpose of such carved pebbles. The find place of the nodules on High Island suggests an association with the lintel graves at the east gable of the church and thus, possibly, with the founding saint of the monastery. Whether the stones were kept there or elsewhere (e.g. on the altar at the east end of the church), they are likely to have been regarded as ‘holy objects’ and thus objects possessing certain supernatural powers (Pl. 6.2). The use of ‘cure stones’ is probably of very great antiquity. In some cases (e.g. quartz and crystal), it is the intrinsic property or properties of the stone itself that are considered to have magical powers. Like most of the recorded altar stones, the High Island examples have a smooth, rounded shape, making them very pleasing to hold in the hand. They are water-rolled and were probably

collected on the seashore or, in the case of the nodules, possibly in a riverbed. There is some evidence that the connection with water had special significance. In a study of Scottish folklore customs, Ross (2000) records that certain river stones were believed to have magical properties and were used both as protective amulets and as ‘charms to promote harm’. The association of the artefact with some revered person or place (usually a grave, church or well) added another dimension. Adomnán’s Vita Columbae includes an account of a white river stone (lapis Candidas) being blessed by Columba for the purpose of working cures among the pagan Picts (Forsyth 1995, 692). There is no record of any of the High Island stones being used as ‘cursing stones’ and it is not known how ancient the Irish tradition of using ‘altar’ stones for maledictory purposes might be. (For a comprehensive discussion of formal liturgical cursing, see O’Sullivan and Ó Carragáin 2008, 335–41.) There are no known documentary references to such a practice in the early medieval/ medieval sources (F. Kelly, pers. comm.), although the use of ‘charms’ may have been fairly common. The earliest known reference to any of the surviving cursing stones relates to Inishmurray (O’Sullivan and Ó Carragáin 2008, 22–3, pl. 13). An early 17th-century map of the northwest of the country includes the following legend regarding the island: ‘in this ilande there dwelleth a holye man named Scanlon of whom the countrie people holde a superstitious opinion that if he be angry with anyone he doe turne the speckled stones upon them, that he keepeth for that

the finds 143

Pl. 6.2 The altar at the early ecclesiastical site in Kinallia, Co. Clare, as drawn by T.J. Westropp in 1896 (Westropp 1911).

use, they shall die within that yeare’ (O’Sullivan and Ó Carragáin 2008, pl. 13 and p.359). The best-known ‘modern’ incidents involving cursing stones are the sinking of The WASP off Tory Island (McLaughlin 1990), and the threatened use of the Kilmoon, Co. Clare stones, which resulted in a court case (Westropp 1911). Both events took place in the late 19th century, but there are accounts of cursing stones being used in the west of Ireland as late as the 1950s (Evans 1957, 300). Finally, mention should also be made of the ‘globe’, the large round erratic located in the monastery on High Island (95E124:199:14; Fig. 6.11). It is not clear if the shape of the stone is entirely natural or if it has been modified. There seems to be no specific folklore attached to the erratic, but it may have had some special significance at some stage in the past.

Comparanda The individual finds from High Island can be paralleled at other early medieval monasteries and church sites, such as Iniskea North, Co. Mayo (Henry 1945), Skeam West, Co. Cork (Cotter 1995), and at Church Island (O’Kelly 1958), Illaunloughan (White Marshall and Walsh 2005) and Reask (Fanning 1981) in Co. Kerry, as well as at most secular sites of the period. When the different taphonomic conditions are taken into consideration, the number and range of material goods found at High Island and Church Island, Co. Kerry, are not too dissimilar. Both assemblages are quite limited, however, when compared with the far more diverse collection of objects from Illaunloughan, Co. Kerry (White Marshall and Walsh 2005). That site produced over 250 finds, including a bone trial piece, clay moulds, spindle

whorls and stone bracelet fragments – all items that were absent at High Island, and either absent or very poorly represented at Church Island. Indications are that Illaunloughan may be quite exceptional among eremitic island monasteries. Excavations carried out at the neighbouring, but much larger, monastery on Skellig Michael, Co. Kerry, for example, also produced relatively few finds (Bourke et al. 2011). Larger and more diverse assemblages have been found at comparable mainland monasteries, such as Caherlehillan (Sheehan 1999) on the Iveragh Peninsula and Reask on the Dingle Peninsula (Fanning 1981). Both of those sites also produced imported pottery, an indication perhaps of a relatively wealthier lifestyle. In addition to enjoying different levels of wealth and patronage, small ‘eremitic’ island monasteries may have fulfilled different roles in the community. Church Island seems to have been important as a burial ground, a role possibly fulfilled in the Streamstown/Cleggan Bay area by the ecclesiastical site on Omey Island (O’Keeffe 1990, no. 55; 1992, no. 92; 1993, no.111). The production of high status metalwork seems to have been an important aspect of the monastery on Illaunloughan. The rich legacy of grave-slabs, crosses and other inscribed stones at High Island suggests that the principal resources of the monastic community there were invested in stone carving. There is nothing in the finds assemblage that can be directly connected with that activity, although it is quite possible that some of the coarse stone tools were used for rudimentary stone-breaking, etc. The carving itself was probably carried out using iron chisels and punches, which, given the site conditions, would be unlikely to survive, even if they were discarded. Several aspects of the High Island assemblage suggest that the monastic population keyed into a network of cultural influences that can be very loosely classified as ‘Viking’ or ‘North Atlantic’. The recycled porphyry pieces and the silver coin, probably minted in Norway, are items that could have been acquired directly or indirectly from Viking traders, possibly sometime during the 10th or 11th centuries respectively. Similar evidence for Scandinavian contact came to light at Illaunloughan (Marshall and Walsh 2005, 175), where it was interpreted as likely to be associated with settlement of probable Hiberno-Norse origin in the area. A 10th/11th century date has been suggested (O’Kelly 1956; Sheehan et al. 2001) for what appears to have been a Scandinavian settlement on the neighbouring island of Beginish. There seems no reason why similar settlements could not have existed along parts of the Connemara coast – some of the offshore islands, for example, would have made ideal maritime bases for traders working the North Atlantic routes. To date, however, material evidence for the existence of Hiberno-Norse way-stations in the Connemara

144 high island: excavation of an early medieval monastery Table 6.1 Dated items from the finds assemblage.

Find/Feature no.

Object/context

Date/date range

95E124:38:1

Porphyry fragment found with burial F40 (Grave 4)

cal. AD 980–1023 (UB-4155, 1027±19 BP)

95E124:320:1

Silver penny from Cell B

1047–1066 (see Section 6.7)

F208

Burial F208 inside church (Grave 9)

cal. AD 1163–1230 (UB-4000, 849±16 BP)

95E124:482:1

Silver halfpenny from fill of disused water channel (F8025)

1204–1211

region remains rather meagre. The picture may, of course, be radically altered by future finds. The closest affinities of burial F208 (Grave 9) certainly appear to be with Scandinavian rather than with native practices, but mystery surrounds the individual and how he came to be buried on the island. The fact that he was interred in a (probably) disused church on a fairly inaccessible offshore island suggests that those who buried him were already familiar with the island. He may therefore have been associated, either through trade or kin, with a local settlement of Hiberno-Scandinavian origin. His wider connections may have been with other HibernoScandinavian settlements in Ireland, with the Norse kingdoms in Scotland or, even further afield, with the Viking homeland.

Dating Most of the assemblage is not amenable to close dating. The small number of items that are datable (Table 6.1), either in their own right (the coins), by comparative means (the porphyry) or by radiocarbon evidence, all seem to fall into the period around the end of the first/beginning of the second millennium AD. Based on stylistic elements, White Marshall and Rourke (2000, 168–9) also consider that the majority of the High Island crosses and carved stones were made in the late first /early second millennium AD. The inscribed nodules and pebbles could date to that period or later, with the person, or persons, who carved them possibly copying the cross-in-circle motif, which is a characteristic element of the ‘local’ style. A slightly broader date bracket of between the ‘late ninth and eleventh centuries’ is suggested by O’Sullivan and Ó Carragáin (2008, 335) for the carving of the more elaborate altar stones at Inishmurray, but this remains to be demonstrated. The date span of the chronologically undiagnostic items from the High Island excavations could be much broader. The absence of any recognisably ‘early’ (i.e. pre9th century) artefacts is noteworthy, but, as noted above, factors such as poor preservation, the poor material circumstances of the monks and, perhaps, the character of the areas excavated may partly account for it.

The post-monastic period With the exception of the coin of early 13th century date, there is nothing in the finds assemblage that can unequivocally be said to belong to the period from roughly the 13th to 17th centuries. That there was some activity on the site during this period is indicated by the radiocarbon-dated barley remains (cal. AD 1287–1424; UB 4988, 580+40BP) from Cell A. A building in Area 4 may also post-date the main period of monastic occupation on the island (see Section 5.5). The nature or duration of those activities remains a matter of speculation; one consideration is that the mill may have been kept in repair for some time after the monastery was abandoned. The fact that its existence was known of into the 19th century (OSL Galway 1839, Vol. 111, 52) would suggest this. A number of other excavated monastic sites (Illaunloughan (White Marshall and Walsh 2005, 202), Church Island (O’Kelly 1958, 112–13), Caherlehillan (Sheehan 1999) and Reask (Fanning 1981, 115)), have produced evidence for a medieval presence. The disused church site on the small island of Skeam West, Co. Cork, was also occupied at some stage or stages during the late medieval/early post-medieval period (Cotter 1995). While the difficulties involved in getting on and off High Island may have discouraged permanent settlement, it is quite possible that the island was occupied from time to time. This seems to have been the pattern on Inis Tuaisceart, one of the most remote of the Blasket Islands, where landing conditions were equally difficult. The island was unoccupied when Charles Smith visited it in the mid-18th century, but for the following 100 years or so, the ruins of the monastery on Inis Tuaisceart were intermittently inhabited by shepherd families (Cuppage 1986, 297). Seasonal occupation is also a possibility in the case of High Island and in the recent past, at any rate, there are plenty of instances of west coast islands being used during the summer months as fishing bases, or for animal grazing. The ‘later’ goods in the High Island assemblage can be bracketed into the period between the late 17th or 18th century and the early 20th century. The majority probably derive from the activities of the miners who resided for a

the finds 145

brief period at the monastery in the early 19th century. The finds are catalogued according to the material from which they are made. Geological identification of stone objects is by Brian McConnell of the Geological Society of Ireland (Section 6.3). The metal-working debris has been analysed by Tim Young (Section 6.5). Identification of the coins is by Michael Kenny (Section 6.7). The pottery report is by Clare McCutcheon (Section 6.11). The assemblage is now in the National Museum of Ireland (museum registration number, 952E124).

6.2 The stone objects Introduction The lithic assemblage includes 48 objects comprised of six architectural fragments, three pieces of porphyry, three ‘altar stones’, eight hones (of which five are the manufactured type and three are made on cobbles/pebbles and tabular pieces of stone), eight cobble tools (including rubbing stones, grinders, pounders, grinding slabs and burnishers), one rotary quern, six perforated and notched stones, eight pieces of flaked lithics (two of flint and six of chert/flint) and four other possible artefacts. Hone stones were the most numerically dominant tool type. At High Island, they appear to have been used primarily for grinding down and smoothening off; sharpening lines/grooves are relatively rare. The finer examples are made on thin tabular pieces of stone (430:2, 208:1) or thicker rectangular- or square-sectioned stones (487:1, 4002:1, 464:1). The two finely tapering forms, 4002:1 and 464:1, tend towards the type of hone that is generally regarded as ‘Viking influenced’. Both are more robust than the perforated examples found at Cahercommaun (Hencken 1938, fig. 35, no. 259), Illaunloughan (White Marshall and Walsh 1998, fig. 101, 401:7) and Waterford (McCutcheon 1997, figs 14.7 and 14.8). Of the perforated stones, 208:2 was probably a line weight for use in handline fishing from the shore, but the second weight, 438:4, seems too well-finished for that purpose. The probable sharpening wheel (478:2) is very well-made; 4022:4 is also likely to have had some industrial function. The wear pattern suggests it rotated against another surface, but the perforation is too small for the piece to have been a pivot. Most of the remaining lithic implements are made on cobbles and, apart from wear on the working edges or facets, the stones show very little modification. Hones made on natural cobbles are frequently difficult to distinguish from rubbing stones, and in practice many coarse stone tools were probably multi-functional. As well as smoothing off items made from softer stone, bone/horn, wood or possibly metal, rubbing stones could be used in the preparation of hides, in leather-working or for breaking down vegetable fibres.

None of the High Island rubbing stones displays the staining or high polish that comes from working with organic materials, however, and it is more likely that they were used on harder material. The grinder (265:1) is hand-sized and could have been used to break down organic (bone marrow, nuts, seeds, etc.) or inorganic (rocks, ores and minerals) materials. Depending on the rock types involved, it could also have been used at different stages during the manufacture of stone tools. Pounders tend to be more robust than grinders and, like the High Island example (256:1), are generally made from hard coarse-grained rock, such as granite. The piece could have been used for roughing-out blocks for stone carving. The three grinding slabs are also made on cobbles; they are similar to examples from Dún Aonghasa and Dún Eoghanachta (Clarke 2012a; 2012b), but smaller than the grinding slabs from Cahercommaun (Hencken 1938, 58). The function of these slabs is debatable. They do not have the dished profile of prehistoric saddle querns, which they resemble, but could nonetheless have been used for grinding or breaking down vegetable matter. The High Island examples once again lack the glassy shine seen on some of the Aran slabs, but this may be due to the properties of the rock type, rather than to any specific usage. Charcoal staining on 457:3 may derive from hearth debris. Alternatively, the slab could have been used for an industrial purpose, such as the breaking down or processing of iron ores or blooms. The association of grinding stones with metal-working can be traced back to the Bronze Age (e.g. Lough Eskragh, Co. Tyrone; Williams 1978). While there are occasional instances of prehistoric saddle querns being reused during the early medieval period (e.g at Rinnaraw cashel, Co. Donegal (Comber 2006, Pl. viii:E416:41), where the stone was probably used in conjunction with ironworking), there is no reason to presuppose a prehistoric origin for the type of grinding slab found at High Island. Only two, or at most three, of the eight items in the flint and flint/chert assemblage show any signs of having been worked or used. The shape of 830:1, a piece of darkgrey flint, is comparable to a flint reamer found at Reask, Co. Kerry (Fanning 1981, fig. 28, no. 315). Piece 211:1 is a gun flint of a type that was widely used in pistols and muskets until the close of the 19th century. The remainder may be waste chips or, in some cases, simply the by-product of stone breaking. There is possible retouch along one, concave edge of 889:2; the shape of the piece suggests it may have been used as a scraper, but this is very tentative.

146â&#x20AC;&#x192; high island: excavation of an early medieval monastery

95E124:2:3 Window head

See conjectural reconstruction of window, Fig. 6.2

95E124:199:1 Window-sill

95E124:209:1 Holy water stoup

200mm

95E124:833a:1 Possible support stone

100mm

95E124:499:2 Carved stone

Fig. 6.1â&#x20AC;&#x201A; Architectural fragments.

Fig. 6.1

the finds 147

The Architectural fragments 95E124:2:3 (Figs 6.1, 6.2)

Window head, complete. Carved from garnet mica-schist. Stone, L: 640mm, W: 210mm, T: 115mm; Ope, Depth: 160mm, W: splays from 140mm to 250mm, H: 190mm. Area 1; found in rubble outside east end of church, probably in the position it fell when the gable collapsed. 95E124:199:1 (Figs 6.1, 6.2)

Window-sill, complete. Made from a rectilinear piece of mica-schist of irregular thickness; the two shorter ends of the stone have been trimmed at an angle. A sloping, splayed depression (19mm deep) on the surface of the stone marks the location of the window opening. The narrowest end of the splay corresponds in width (14cm) with the span of the window head. Found in spoil removed from general area of church. Both stones (199:1 and 2:3) are part of the narrow vertical light that was still in situ in the east gable of the church in 1820 (White Marshall and Rourke 2000, 79). The window was described by Petrie as 0.35m high by 0.15m wide. O’Donovan’s sketch of the surviving section of the inner embrasure (Fig. 3.1) indicates that it was widely splayed, c. 1m in maximum height and possibly not a true round arch. It is unclear if the floor of the embrasure was stepped or sloped. Descending steps are the most common form in early stone churches (Leask 1955, 58–9) and the shape of the sillstone suggests that the High Island window may have conformed to that pattern. See Fig. 6.2 and conjectural reconstruction of the interior eastern window in the church (from White Marshall and Rourke 2000, fig. 48a and 48b).

95E124:209:1 (Fig. 6.1)

Holy water stoup, complete. Made from a roughly ovoid piece of garnet mica-schist. The stone has two straight edges, but the greater part of its circumference follows the elliptical outline of the basin, leaving a rim 75mm in width. The well-carved basin has sloping sides, more pronounced at one end than the other, and a flat base. The underside of the stone has a shallow, sloping facet with traces of polish, possibly from friction. The stone may originally have been set into a wall; the earliest antiquarian accounts show it placed on the altar inside the church, the same position it was found when excavation began in 1995. Stone, L: 450mm, W: 300mm, T: 94mm. Basin, L: 305mm, W: 185mm, Depth: 50mm. Area 2, found on the altar at east end of church. 95E124:499:2 (Fig. 6.1)

Carved stone, fragment, possibly a pivot stone or portion of a bullaun. Made on a rounded boulder of talcose schist, broken across one end and spalled on the other. The underside of the stone is also spalled or chipped away, but whether intentionally or not is impossible to say. Portion of a rounded perforation survives on the broken end of the stone. The lower sides of the perforation taper inwards, suggesting the piece had a base originally, and it is quite possible that this has either worn through or been broken off. The size of the carved depression is larger than is usually associated with spud or pivot stones and the possibility that the object might be a broken bullaun stone cannot be ruled out. Stone (max. dimensions), L: 300mm; W: 240mm; H: 132mm. Remains of depression: 165mm x 82mm x 100mm deep. From spoil removed from Area 4.

Fig. 6.2 Conjectural reconstruction of the interior of the eastern window in the church. Reconstruction by G.D. Rourke (White Marshall and Rourke 2000, Fig. 48a).

148 high island: excavation of an early medieval monastery

95E124:402:1

(Not illustrated) Possible support stone, incomplete. Irregularly shaped slab of mica-schist of wedge-shaped cross-section, probably broken off a larger block. The remains of a U-shaped notch survive on the thinner end of the stone. A concentric depression, 90mm wide, surrounds part of the circumference of the notch on one surface of the stone. This could have been made to accommodate the swivelling action of a timber upright, such as a door-post. Stone, L: 440mm; W: 320mm; T: 120mm. Area 4, from rubble. 95E124:833a:1

(Not illustrated) Possible support stone, more or less complete. Flat, rectilinear slab of mica-schist, regular in cross-section. A U-shaped notch has been carved on one of the shorter ends; one edge is broken off, but the missing portion is unlikely to have been longer than 50mm. There are no definite signs of wear and it is uncertain what the function of the stone might have been. It could have supported an upright, acting either as part of a cross-support or as a pivot stone. The slab was reused as a paving stone. Stone, L: 470mm; W: 330mm; T: 47mm. Notch: 71mm x 61mm. Area 8, reused within paved surface (F833a) south side of Cell B. Porphyry The three pieces of dark-green feldspar porphyry (38:1, 4030:1, 4030:3) are all parts of sawn and polished tile pieces; 38:1 and 4030:3 retain short sections of the original edges. Two of the pieces (38:1 and 4030:1) are very similar in appearance and thickness and may well have come from the same tile. The third piece (4030:3) is thinner and, unlike the other fragments, has a wedge-shaped cross-section. Piece 38:1 was found, probably in situ, around the skull region of Burial F40/Grave 4, located outside the east gable wall of the church. The burial has been dated

95E124:4030:1

95E124:38:1

Fig. 6.3 Porphyry pieces.

to the period cal. AD 980–1023 (UB-4155, 1027+19 BP). The other two pieces (4030:1, 4030:3) came from the fill of the primary monastic enclosure wall and are likely to have been dumped there with clays used to bulk up the core of the Phase 1 wall. Charcoal from soils (F4037) abutting that particular phase of the wall produced a radiocarbon date of cal. AD 778–1000 (UB-4987, 1138+48 BP). Charcoal from fill (F457) associated with the later refurbishment of the wall yielded a radiocarbon date of cal. AD 1027–1238 (UB-4521, 888+41BP). On the radiocarbon evidence, therefore, the two porphyry fragments seem likely to have been incorporated into the wall core prior to the end of the first millennium AD. Given that layers F4037 and F457 include secondary redeposited sediments, however, the date range for burial F40 may be a more reliable indicator of when the porphyry pieces were brought to the island. If piece 38:1 was buried with its ‘owner’, it may have been in circulation in the monastery for a half-century or more. Including the pieces from High Island, the number of known porphyry fragments from Irish contexts is 33, an increase of over 100% on the number recorded in 1984 (Lynn 1984 and pers. comm.). Twenty of the specimens came from the port towns of Dublin and Waterford. The remainder are associated with ecclesiastical sites, three of which are located on islands (High Island, Co. Galway, Skeam West, Co. Cork, and Church Island, Lough Carra, Co. Mayo). The Irish corpus is made up almost exclusively of green porphyry (porfido verde antico), exceptions being single pieces of red porphyry from 16 Scotch St., Armagh (Lynn 1976, no. 7; 1984) and Christchurch Place, Dublin (E172:12662). Outside Ireland, porphyry fragments have been recorded in Scotland (c. nine pieces), Denmark (c. 50 pieces), parts of Norway/the Baltic and, more recently, Iceland and the Faroes (Cormack 1989, 149–50; Cormack quoted in Chadburn and Hill 1997, 469). It is more difficult to identify recycled porphyry in areas where Roman Imperial buildings formerly existed, but a few porfido verde antico pieces have been recorded from medieval English and continental sites.

25mm

95E124:4030:3

the finds 149

In a detailed study of exotic porphyry, Lynn (1984) suggests that most of the pieces found in Ireland come from continental early medieval mosaics and pavements, having been cut down and reused from Roman imperial buildings. The origin of the dark-green feldspar porphyry is likely to be Laconia, Greece, where it was quarried up to the end of the 5th century. The fragments may have been imported as tiles/parts of tiles for use in portable altars and as relic covers in fixed altars, or they may be pilgrims’ keepsakes (Lynn 1984). Cormack (1989) suggests that some of the Scottish pieces and the fragment from Movilla Abbey, Co. Down, may have been dispersed from a dilapidated reliquary that was housed at the monastery of Whithorn in Galloway, Scotland. Although there is no archaeological evidence to support this (the single fragment of porphyry recorded from Whithorn came from a 14th-century context: Hill 1997, 258), that large or prestigious monasteries were themselves dispersing (or redispersing) ‘relics’ is certainly a consideration. While pilgrimage may have been the principal mechanism for the dissemination of recycled porphyry, the distribution and context of most of the known pieces strongly suggests that there was also a Scandinavian ‘trade’ in the commodity. Some of the larger pieces found in urban contexts (e.g. E122:1408 from Fishamble St., Dublin) may well have been destined for cutting up and redistribution. Most of the porphyry pieces recovered from ecclesiastical sites are not closely datable. As Lynn (1984) notes, however, at all of these sites there is evidence for activity going back to the 11th century and earlier. A fragment of porphyry (E558:960) found in 1992 during excavations associated with the New Graveyard at Clonmacnoise (King 1992 and pers. comm.) is the earliest known instance of the occurrence of recycled porphyry in Ireland. The piece was found in the same context (F87) as E-ware and lay underneath a horizon that has been radiocarbon dated to the period cal. AD 660–802 (2 sigma; UB-4337). The provenanced specimens from Dublin were deposited in the 11th century (Lynn 1984 and pers. comm.) and include a piece from the interior of a house at Christchurch Place. The house lay under one dated by dendrochronology to 1059 and which, in addition, contained a coin of Sitric (1035–55). The Waterford City pieces are assigned to the 12th/13th centuries; three were found in the backyard of a sill-beam house dated from the mid- to late 12th century (McCutcheon 1997, 432). 95E124:38:1 (Fig. 6.3)

Porphyry, fragment. Retains short (43mm) length of original straight edge. The edge and upper surface of the tile are polished and saw marks in the form of fine milling are clearly visible on the underside. The tile is rectangular in cross-section; a small area of wear occurs on the upper edge.

L: 43mm, W: 27mm, T: 13mm. Area 1, the piece was found in grave-fill near the skull of burial F40 in Grave 4. 95E124:4030:1 (Fig. 6.3)

Porphyry, fragment. The piece is of similar thickness and appearance to 38:1, but lacks any original edges; both pieces may be fragments from the same tile. L: 45mm, W: 35mm, T: 13mm. Area 4, from the lower fill of the primary monastic enclosure wall (F4040). 95E124:4030:3 (Fig. 6.3)

Porphyry, fragment. The piece is wedge-shaped in crosssection and retains two short, straight lengths (20mm each) of the original, opposing, slightly angled edges. Both edges, and the upper surface, are polished. The sawn underside is far more worn on this piece than on 4030:1 and 38:1. Found in the same context as 4030:1. L: 53mm; W: 56mm; T: 5–9mm. Area 4, from the lower fill of the primary monastic wall. The ‘altar’ stones; artefacts associated with graves, leachta/altars and pilgrimage This group comprises three distinctive decorated/worked stone nodules (4:1, 4:4, 4:5), two of which are decorated (see below). The undecorated nodule, 4:5, has been ground along the edges; the faces have also been flattened and there is a central depression on one face. These modifications may be associated with some secondary use. There is no evidence that the stone was ever decorated like its counterparts. The nodules were recovered in post-abandonment deposits in the vicinity of the graves at the east gable of the church, suggesting that they may have been physically associated with those graves at one time. X-ray fluorescence analysis of the objects (see this volume, p.159) suggested classification as ironstone. The provenance of the objects could well be the Burren or the Aran Islands, as almost all the comparable ‘natural’ nodules known from the archaeological record come from that region. A number of other ‘coalfield’ locations are also possible, however (see this volume, p.159). In their natural state, the nodules are usually discshaped. Depending on the degree to which an individual stone has been rolled and abraded, the central part of the face may be either slightly domed or flattened, but is generally quite smooth. The outer edge of the stone is usually rough owing to laminar scaling. The High Island nodules have been ground and smoothed all over. Two (4:1, 4:4) have incised concentric circles/pairs of circles on both faces. One of these (4:4) has, in addition, a central equal-armed ‘Maltese’ cross made by incising triangles at the cardinal points (Section 7.1, Cross 24, Fig. 7.5). The cross-in-circle motif has a long history in Christian art and the motif occurs frequently as a centrepiece on the High Island crosses and carved stones; the ‘circle within a circle’

150 high island: excavation of an early medieval monastery

95E124:4:1

95E124:4:5

50mm

Fig. 6.4 Stone nodules. (See also Fig. 7.5, Cross 24.)

is a particular characteristic shared with the carved stones of Clonmacnoise (White Marshall and Rourke 2000, 169). Parallels for these pieces are discussed in White Marshall and Rourke (2000, 161, fig. 127) and by Maddern (this volume, p.198). The authors place them in the tradition of altar stones, cursing/cure stones, decorated examples of which are known from other well-known island monasteries, such as Inishmurray, Co. Sligo, and Iona, Argyll, Scotland. Similar, but uninscribed stones were a common feature at pilgrimage sites in the west of Ireland up to recent times (Wakeman 1862, 572). In the early part of the 20th century, Westropp (1911) remarked on their frequent occurrence at penitential altars and holy wells in the Burren, Co. Clare. The now missing ‘cursing’ stones from the monastery on Tory Island, Co. Donegal (McLaughlin 1990), and the ‘lucky’ or ‘wishing’ stone from Mevagh Church, Co. Donegal (Welsh 1902, 231; Lacy 1983, 257), probably also belonged to the same tradition. (For other examples, see O’Sullivan and Ó’Carragáin 2008, 379–80, footnotes 113 and 114). The closest parallel for the inscribed nodules is a similar-sized black pebble from Temple Brecan, Inis Mór, Aran Islands (Stokes 1878, Pl. X11, no. 25 and 26), which bears a plain, sharply incised cross and the inscription ‘OR DO BRAN N-AILETHER’ (‘pray for Bran the pilgrim’). The pebble has been smoothed and flattened to such a degree that it is difficult to say from visual inspection alone if its lithology is the same as the High Island nodules. The stone was found at the beginning of the 19th century, when the reputed site of St Brecan’s grave was opened. According to Petrie’s account (Stokes 1878, 20), it lay among other rounded beach pebbles under a large flag at a depth of 4

feet (1.2m). The ‘cake-like concretions found in the shale of the district’, recorded by Westropp (1911) at penitential sites in the Burren, are also very similar in appearance to the High Island nodules (Pl. 6.2). There is no record of any of these being decorated and similar unmodified nodules have been found on at least two settlement sites in the same region, at Gragan West (Cotter 1988; 1990) and at Cahermacnaghten (Fitzpatrick 2007). 95E124:4:1 (Fig. 6.4)

Decorated nodule, complete. Disc-shaped ironstone nodule, smoothed all over. The faces have been slightly flattened and one is particularly marked by superficial scratches. A pair of closely spaced (c. 5mm), shallow, concentric circles has been incised at the edge of the flattened area on one face. A corresponding pair of circles on the opposite face is now either extremely worn or was never fully executed. A single incised circle runs near the outer edge of the stone on each face. The decoration varies from well executed to quite haphazard. The outermost circles are crookedFig. and6.4 uneven and surrounded by scratch marks; the lines defining the inner circles frequently split and merge. Divider/ compass marks are evident at the centre-point of both faces. Pinprick depressions on opposing sides of the stone may mark the points where the stone was secured while it was being worked. D: 81mm x 88mm; max. T: 48mm. Area 1, from deposit of soil and stone (F4) over paved surface (F29) at end of pseudo Grave 2. 95E124:4:4 (Fig. 7.5)

Decorated nodule, complete, made from ironstone. Smaller than 4:1 and with a more convex profile than either 4:1 or

the finds 151

4:5. The edges are quite worn and this, and some laminar flaking, has removed part of the decorated surface. There is a narrow polished ridge where the facets of the stone meet along the circumference of the stone; this is probably partly natural, but it may also have been enhanced by the stone being rotated against a hard surface (e.g. a bullaun stone). As in 4:1, compass/divider marks occur at the centre-point of each face. The decoration is as follows: two pairs of concentric incised circles occur on each face. One of the faces has an ‘extra’ circle running between the pairs and additional decoration in the form of a central, equal-armed (Maltese) cross filling the inner circle (D: 55mm). The cross is defined by inscribed nested triangles that fill the intervals between the arms. These have been done ‘freehand’, so that the number and height of the triangles varies. Overall, the incisions on this stone are far more regular and better executed than on 4:1, although, like the latter, the line segments double up or almost merge in places. D: 79mm x 82mm; max. T: 43mm. Area 1, from deposit of soil and stone (F4) over paved surface (F29) at end of Grave 3. 95E124:4:5 (Fig. 6.4)

Worked nodule, complete, made from ironstone. This is the thinnest of the three nodules and also has a pronounced flattened rim (W: 9mm) running around its entire circumference. The faces have been smoothed off and flattened out to a greater degree than on either of the other two nodules. The flattening is more pronounced on one face, where a poorly made central depression (D: 11mm; Depth: 2mm) is surrounded by a 20mm-wide band of rough pecking. The surfaces of the stone are criss-crossed by scratches and, in some cases, slightly deeper score marks. These appear to be incidental, however, and there is no evidence of any formal decoration. Pinprick depressions on opposing sides of the nodule probably mark the points where it was secured while being worked. D: 78mm x 80mm; max. T: 34mm. Area 1, from deposit of soil and stone (F4) over paved surface (F29) between Graves 3 and 4.

Tools Hones, manufactured type 95E124:208:1 (Fig. 6.5)

Hone, complete. Well-made on thin, tapering, rectilinear piece of ironstone of rectangular cross-section (this volume, p.159). There is an all-over polish and the angles at the broader end have been rounded off. Sharpening grooves visible on both faces are mainly superficial and most are little more than surface scratches. A set of short curved lines on one face may result from sharpening a metal tip as opposed to a blade. The object was found on

the right shoulder of a burial (F208) located inside the church. The perforated weight (208:2) was found on the left shoulder of the same individual; also associated was a bone pin (208:3) found over the right side of the pelvis. The burial was that of a male aged between twenty and 30 years (O’Donnabhain, this volume, p.232) and is dated to the mid-12th/early 13th century. L: 100mm; W: 300–360mm; T: 9mm. Area 2, from right shoulder of burial F208 (Grave 9). 95E124:430:2 (Fig. 6.5)

Hone, incomplete. Made from flat, tabular piece of finegrained sandstone. The piece tapers in width, but is broken off across the narrower end. Remains of the original cortex of the stone are preserved on one face. The long edges are worn from use and there is significant iron staining in one area. Rounding of the angles at the broader end of the implement may also result from wear, but it seems more likely that this was the ‘gripping’ end of the implement rather than the business end. L: 78mm; W: 23–30mm; T: 7mm. Area 4, from rubble. 95E124:464:1 (Fig. 6.5)

Hone, incomplete. Very well made from fine-grained sandstone. The piece tapers in length, is square in cross-section, but broken off near the narrower end. The working surfaces, including the butt end, are extremely flat and smooth. L: 49mm; W: 14–22mm; T: 13–22mm. Area 4, from fill layer within disused water-collection and drainage channel (F8025), east of Cell A. 95E124:4002:1 (Fig. 6.5)

Hone, incomplete. Very well made on tapering, rectilinear piece of fine-grained sandstone. The implement has flat, smooth faces, but is broken off across the narrower end. The faces and long edges are smoothened from use and there are some shallow sharpening grooves near the broken end. The piece was found in the vicinity of the passage leading to building F427. L: 103mm; W: 16–23mm; T: 14mm. Area 4, from low bank of soil and stone east side of passage (F4028) leading through monastic enclosure wall to sub-rectangular building (F427). 95E124:487:1 (Fig. 6.5)

Hone, incomplete. Square- or rectangular-sectioned hone made from psammite. The surviving piece consists of a triangular-shaped flake that has been broken off one side of a well-made implement. The straight butt end may be part of the original end of the implement, or the line of a previous break. The working surfaces are ground to a smooth finish and there is black staining visible on one side. Blackening of the broken away face is probably due to post-depositional factors.

152â&#x20AC;&#x192; high island: excavation of an early medieval monastery

95E124:464:1

95E124: 487:1

95E124:430:2

95E124:208:1 0

50mm

95E124:4002:1

Fig. 6.5â&#x20AC;&#x201A; Hones, manufactured type.

L: 50mm; W: 24mm; T: 20mm. Area 4, from the fill of a gully at the north end of the water collection-channel, north-east of Cell A. Hones made on cobbles/pebbles and tabular pieces of stone 95E124:478:3 (Fig. 6.6)

Hone/rubbing stone, incomplete. Made from a sandstone pebble, probably ovoid in shape originally. The piece is broken across its width, possibly as a result of exposure to heat. One facet of the stone is smoothened and blackened from use; a second facet also shows some wear polish. The original pitted surface of the pebble is preserved on the remaining faces. L: 40mm; W: 28mm; T: 39mm. Area 4, the piece was found in burnt residues outside Cell A. 95E124:4022:5 (Fig. 6.6)

Hone/rubbing stone, incomplete. Made from a flat, unmodified pebble of fine-grained red sandstone stained black in places. The piece is broken at one end; the opposite end is rounded and the pebble has a flattened oval cross-section. The surfaces have been worn to a very smooth finish and there are faint striations visible on one face. Slight pocking on the end of the stone suggests occasional use as a hammer/pounder. The stone may be fire-shattered. Max. L: 65mm; W: 62mm; T: 29mm. Area 4, found in the fill of the extended monastic enclosure wall (F4025). 95E124:4071:1 (Fig. 6.6)

Hone, complete. Irregularly shaped, abraded and weathered piece of fine-grained sandstone, rectangular in section. The shouldered sides taper so that one end is rounded and

the opposite end is reasonably straight. Smoothed and polished from use, mainly on the faces. Much of the surface of the implement is blackened and there are burnt residues adhering to the middle section on both faces. L: 142mm; W: 58mm; T: 23mm. Area 4, from deposit sealed by collapsed exterior face of extended monastic enclosure wall, west of Cell A. Cobble tools: rubbing stones, grinders, pounders, grinding slabs, burnishers 95E124:199:11 (Fig. 6.7)

Rubbing stone, complete. Made on a small natural quartzite pebble, roughly oval in shape and well-suited to being grasped between the thumb and forefinger. Two of the six facets are flattened and smoothened from rubbing; a third face shows slight evidence of use. D: 52mm x 37mm; T: 40mm. Area 1, stray find recovered from spoil removed from Area 1. 95E124:29:1 (Fig. 6.7)

Rubbing stone, complete. Made from flat, triangular fine quartzite pebble, rectangular in cross-section. Smoothed and polished from use on long sides and along contiguous portion of one face. Original surface of pebble preserved on opposing face. Flake scars on one of the long edges could result from secondary use of the implement as a hammer, but the fresh appearance of the scars suggests they are more likely to be of recent date. L: 82mm; W: 53mm; T: 20mm. Area 2, from paved surface around the church.

Fig. 6.5

the finds 153

95E124:478:3 95E124:4022:5

50mm

95E124:4071:1

Fig. 6.6 Hones, made on cobbles/pebbles and tabular pieces of stone.

95E124:334:1 (Fig. 6.7)

Rubbing stone, complete. Made from a small, circular quartzite pebble. Half of one surface and the edge of the stone contiguous to it are smoothened and worn from rubbing. The wear pattern suggests that the item being worked on was held at an angle. The rubbing stone was found on a surface contemporary with the smithing hearth (F352) beneath Cell B, but there is no trace of any iron or charcoal staining on any of its surfaces. D: 59mm; T: 41mm. Area 3, from deposit of gritty clay sub hearth-floor (F326) within Cell B, probably pre-dates cell. 95E124:265:1 (Fig. 6.7)

Grinder, complete. Roughly D-shaped flat cobble of quartz feldspar mica-schist. Found in two pieces, but subsequently spalled away at the working end. The stone is lightly pecked and faceted at the narrower (L: 30mm) end; the adjacent part of the surface of the cobble is stained black. The most likely use of the tool was for crushing in the manner of a pestle. As it does not seem to have been extensively used, the stone may have broken at an early stage. D: 105mm x 76mm; T: 32mm. Area 2, found in the backfill of the socket (F260) for the pre-church ‘slotted feature’. 95E124:256:1 (Fig. 6.7)

Pounder/grinder stone, incomplete. Made on a granite cobble that was probably circular in shape, and oval in cross-section originally. One end of the cobble and about half the attached face have broken off – the angle of the

break suggests that the missing end may have been used for heavy duty pounding or hammering. The opposite end has a flat, pecked facet. Both it and the adjacent surface of the cobble have been lightly smoothened and striated, probably as a result of a grinding action over a hard surface. Flaking away of part of the facet suggests this end of the stone was also used as a pounder. There are traces of black staining on the face of the stone. D: c. 132mm T: 81mm. Area 2, found within a deposit of stone and soil at the base of the east wall of the church. 95E124:4034:1 (Fig. 6.7)

Grinding slab, incomplete. Made on a slab of gabbro with naturally rounded ends. The stone is now broken so that its original length or width cannot be determined – an estimated width would be c. 160mm. The almost flat upper surface is worn smooth, but the ‘working area’ does not extend to the edges of the stone. The surface of the stone is stained black in places and that, and a number of shatter marks, suggest that the piece may have been burnt at some stage. Max. dimensions: L: 210mm; W: 118mm; T: 52mm. Area 4, from fill layer within the primary monastic enclosure wall. 95E124:450:1 (Fig. 6.7)

Grinding slab, incomplete. Made from an unmodified, oval cobble of coarse-grained granite. The stone is partly broken away at one end and spalled at the opposite end. A slightly dished natural hollow on one of the longer faces has been used as a grinding surface. Only one of the origiFig. 6.6 nal edges of the smoothed area survives, but the grinding

154â&#x20AC;&#x192; high island: excavation of an early medieval monastery

95E124:199:11

95E124:29:1

95E124:265:1

50mm

95E124:334:1

95E124:450:1 95E124:256:1

100mm

95E124:457:3 95E124:4034:1

Fig. 6.7â&#x20AC;&#x201A; Cobble tools: rubbing stones, grinders, pounders, grinding slabs and burnishers.

Fig. 6.7

the finds 155

surface cannot have been larger than 100mm x 80mm. The underside of the stone is smoothened, probably from friction during the grinding process. The damage to the stone is unlikely to be accidental and probably results from secondary use of the cobble as a hammer or pounder. L: 155mm; W: 115mm; T: 74mm. Area 4, from rubble. 95E124:457:3 (Fig. 6.7)

Grinding slab, complete. Made from a flat, tabular piece of micro-granite, roughly rectilinear in shape, but with one irregular edge. The stone has a naturally smooth, almost glassy, veneer. This has been ground away on one of the larger faces, exposing a more pitted surface, which is charcoal stained in places. The working area is confined to the face of the stone and does not extend to the sides. The piece is very similar to some of the grinding slabs recovered at the stone forts of Dún Aonghasa and Dún Eoghanachta on the Aran Islands (Clarke 2012a, b). L: 200mm; W: 145mm; T: 50mm. Area 4, found in a possible hearth deposit within the fill of the extended monastic enclosure wall. Rotary quern 95E124:10:1 (Fig. 6.8)

Rotary quern, incomplete. Slightly less than one half of the upper stone of a rotary quern found in the masonry of the east wall of the church. Probably c. 280mm in diameter originally. The quern is the standard disc-shaped type and is carved from quartzite. The remains of punch dressing are visible on the sides. The grinding surface is slightly uneven and a large chip is missing from one edge. Part of the central funnel-shaped perforation survives. On the underside of the stone, a flattened angle on the broken edge of the perforation is reminiscent of a rind socket, but there is no corresponding feature on the opposite side of the hole. The carving of the quern must have taken a great deal of effort and it is possible that the stone was reused after the object had broken. D: c. 280mm; T: 53mm; Perforation, D: c. 43mm. Area 1, within east wall of church, above headstone (Cross-slab 8, Section 7) for Grave 4. Perforated and notched stones 95E124:208:2 (Fig. 6.9)

Complete. Probably a line sinker, made from a rough, sub-circular amphibolite pebble of irregular cross-section. One face is reasonably flat and striations may result from the grinding-down process. In contrast to the weight itself, the central hourglass perforation is very skilfully made; the regularity of the reaming marks suggests a lathe might have been used. An elliptical sinking beside the hole is probably a ‘first’ attempt at perforation. There is an irregularly shaped notch (3mm deep) on one edge of the stone

95E124:10:1

100mm

Fig. 6.8 Rotary quern.

and a very shallow reamed groove diametrically opposite. These would not be very effective in preventing a line from slipping; they may mark the points where the stone was clamped while it was being perforated. The weight would be suitable for hand-line fishing, but is probably too light for a net sinker. Another suggestion (C. Maddern, pers. comm.) is that it may have been used as a plumb-line. Weight: 124g. D: 71mm x 67mm; T: 10–20mm. Perforation, D: 11mm. Area 2, found on the left shoulder of burial F208 in Grave 9 (see hone stone 208:1). 95E124:438:4 (Fig. 6.9)

Complete. Made from a tabular piece of quartz feldspar mica-schist with one rounded and one shorter butt end. The faces, one of the long edges, and the ‘butt’ end have all been smoothened. There is a central hourglass perforation: two worn shallow grooves on one of the long sides may be line marks. The stone may have been used occasionally for sharpening; as well as score marks there are two small circular depressions (D: 2mm) that seem to have been made with a metal tip. The stone is not necessarily too heavy for a line sinker – on the Aran Islands similar heavy stones are used for hand-line fishing from the cliffs. It seems too well made for that purpose, however, and it may have been designed for more conspicuous use, for example in a building. Weight: 435g. L: 125mm; W: 85mm; T: 15–20mm; Perforation, D: 15mm. Area 4, found in the fill of the extended monastic enclosure wall. 95E124:4022:3 (Fig. 6.9)

Perforated disc. There are no fresh breaks on the piece, but the fact that one edge is flat and smoothed off suggests that either part of the circumference has spalled off or the disc has been cut from a larger object. It is made on a thin piece of mica-schist; both faces are flat and relatively even.

156â&#x20AC;&#x192; high island: excavation of an early medieval monastery

95E124:438:4

95E124:208:2

50mm

95E124:478:2 95E124:4022:3

95E124: 4022:4

95E124: 623:2

Fig. 6.9â&#x20AC;&#x201A; Perforated and notched stones.

Fig. 6.9

the finds 157

Perforated disc, extremely worn with surface spalling, possibly heat-shattered. The flat, circular disc is made from slaty amphibolite. Only a very small area of the original surface and roughly half of the central cylindrical perforation are preserved. The surviving portion of the original edge is smooth and polished, but not worn flat as one might expect if the piece was a sharpening wheel. As a significant part of the implement has spalled away, however, this remains the most likely interpretation of its function. D: 76mm; max. T: 6mm. Perforation, D: 15mm. Area 4, found in burnt residues outside Cell A.

original surfaces are smooth; the flatter face appears to be the more polished and its outer edge is also worn. The opposite face is too abraded to identify any corresponding wear pattern. There is a well-made (probably central) hourglass perforation. On the convex face, two fingershaped grooves (11mm wide x 2mm deep) extend from the perforation to the edges of the stone. Two, or possibly three rough depressions (D: 3mm; depth 1mm) on the side of the stone are also artificially made. The object is of unknown function; the wear pattern suggests it may have been inserted into a wooden housing, like a cog or a coupling. The hourglass perforation (narrowing to only 6mm at the centre) probably rules out the possibility that the piece was mounted on an axle or spoke, or was used as a block or pivot. The best that can be said is that the piece may have had some industrial use. Surviving L: 61mm; W: 55mm; Perforation, D: 16mm. Area 4, found in the fill of the extended monastic enclosure wall.

95E124:4022:4 (Fig. 6.9)

95E124:623:2 (Fig. 6.9)

The central hourglass perforation is slightly worn along one edge but even so, its sharp edges would seem to make the disc unsuitable for use as a line sinker. Weight: 101g. D: 86mm x 68mm; T: 5–8mm. Perforation, 6mm x 8mm. Area 4, found in the fill of the extended monastic enclosure wall. 95E124:478:2 (Fig. 6.9)

Perforated stone, incomplete. Well-made object of quartz feldspar mica-schist, broken away now at one end. Blackened and partly shattered from exposure to heat. The surviving portion is heel-shaped, with one flat and one slightly convex face – when complete, the object may have been elliptical in shape and c. 86mm in length. The

Cross ? Small, rectilinear piece of mica-schist with semicircular notch (D: 6mm) at one angle. There are no fresh breaks, but presumably the object was originally bigger. It may be a broken-off piece from a rudimentary ‘crossshaped’ grave-marker of a type that is common in many west of Ireland burial grounds. Alternatively, it may be

Table 6.2 Inventory of flaked lithics.

Flint

Description

Dimensions

Context

95E124:211:1

Cushion-shaped gun flint made from dark-grey flint. Worn from use along one edge (identification by Edward Bourke).

23mm x 20mm, T: 8mm

Area 2

95E124:830:1

Chip of light-grey-coloured flint; roughly triangular in section. Small projecting point (L: 4mm; T: 3mm) at one end. No definite signs of wear, but could have been used as a reamer.

25mm x 10mm, T: 8mm

Area 8

95E124:351:1

Tiny triangular-shaped chip. No signs of working (not illustrated).

10mm x 7mm, T: 4mm

Area 3

95E124:214a:2

Flat rhomboid chip, one face preserves the outer side of the original stone. No signs of working.

22mm x 12mm, T: 16mm

Area 2

95E124:215:2

Narrow rectilinear blade-like piece, triangular in section. No signs of use, but business end may be missing.

13mm x 10mm, T: 5mm.

Area 2

95E124:425:3

Small flake fragment, triangular in section. No signs of use. May be part of same piece as 448:3.

15mm x 17mm, T: 5mm.

Area 4

95E124:448:3

Roughly square-shaped flake fragment. No signs of use. Possibly part of the same piece as 425:3.

19mm x 17mm, T: 5mm

Area 4

95E124:889:2

Small thumbnail-shaped piece, broken away along one side. Possible retouch along the adjoining concave edge (L: 10mm). Could have been used as a scraper.

15mm x 12mm, T: 5mm.

Area 8

158â&#x20AC;&#x192; high island: excavation of an early medieval monastery

95E124:211:1 95E124:214a:2

95E124:830:1

95E124:889:2

95E124:448:3

95E124:425:3

95E124:215:2

25mm

Fig. 6.10 Flaked lithics and chert/flint pieces.

95E124:801:1

95E124:66:1

95E124:199:14

95E124:66:1 0

250mm

50mm

Fig. 6.11â&#x20AC;&#x201A; Possible stone artefacts and miscellaneous.

Fig. 6.10

the finds 159

part of a small portable cross similar to examples found at Illaunloughan Island (White Marshall and Walsh 2005, fig.5) and Skellig Michael (Horn et al. 1990). L: 48mm; W: 36mm T: 1–7mm. Area 6 from postmonastic deposit of fragmented schist. Flaked lithics See Table 6.2 and Fig. 6.10. Possible artefacts 95E124:201:18

(Not illustrated) Possible burnishing stone, incomplete. Oval-shaped white quartzite pebble broken at an angle. A very small area of shine at the edge of the break scar may indicate that the stone was used for burnishing, but this is very tentative. Quartz pebbles are difficult to break and it is hard to imagine a scenario involving the deliberate breaking of this stone – its small size would rule out use as a hammer. D: 83mm x 49mm; T: 43mm. Area 2, found in rubble. 95E124:322:1

(Not illustrated) Possible rubbing stone, incomplete. Roughly half of an oval-shaped psammite cobble that has probably been fireshattered. The surfaces are smooth and partly burnt. The only possible evidence of use is an area of scarring and shine on the flatter ‘under’ surface of the stone. This may indicate use as a rubbing stone, but the identification is very tentative. D: 90mm x 67mm; L: 93mm. Area 3, from a (subsidiary) hearth inside Cell B. 95E124:66:1 (Fig. 6.11)

Cobble fragment. Middle section of sandstone cobble of ovoid cross-section; both ends broken away. The stone is in poor condition, either as a result of weathering or exposure to fire. The exterior surface is smooth, but there is too little of the object surviving to say if this is entirely natural or results from usage as a tool (e.g. as a rubbing stone). D: 67mm x 52mm; L: 40mm. Area 1, found in the fill of Grave 3. 95E124:801:1 (Fig. 6.11)

Cobble fragment. A spall from a burnt cobble incorporating a facet with original cortex. No definite trace of modification or use. L: 85mm; W: 66mm; T: 25mm. Area 8, from rubble. Grave furniture

95E124:207:1 (Not illustrated) White quartzite pebble. Roughly circular and of flattened oval cross-section; no trace of modification or wear evi-

dent. An example of one of the many quartzite pebbles recovered at the site, particularly with the graves. D: 67mm x 74mm. Area 2, soil layer below burnt deposit (F221) within the church. Miscellaneous 95E124:199:14 (Fig. 6.11)

Granite globe. Roughly round in shape, although not a perfect sphere. Likely to have been modified to create globe-like appearance. D: 0.4m x 0.35m, found outside excavation area south of the church, between the inner church enclosure wall and the outer monastic enclosure wall (Fig. 5.1).

6.3 Stone identification and X-ray analysis of selected small finds Forty-nine of the smaller lithic finds from the excavations on High Island were submitted to Brian McConnell of the Geological Survey of Ireland for geological identification (excluding the larger stone crosses). With the exception of four items (see below), all items were identified by nonintrusive visual examination and, where required, use of microscopic magnification x50. The items examined were a combination of stone tools, crosses and structural elements from monastic buildings. The bedrock of the island is composed of metamorphic rocks that are part of the Dalradian sequence of Connemara. The predominant rock type is mica-schist, with interbedded quartzite and a smaller outcrop of marble. Small ‘feldstone’ dykes were also noted on a 19th-century map of the island (Fig. 1.4). Granite occurs on the island as substantial erratic boulders and as smaller fragments in the glacial till, but does not form part of the bedrock. In terms of their respective provenances, the lithic items can be divided into three categories: 1. Probable bedrock source, though possibly reworked as a beach pebble first. 2. Non-bedrock, but possibly from the island as beach or glacial pebble, or from a nearly island or mainland source. 3. Exotic or non-local. Category 1. refers to items composed of a rock type that forms part of the island bedrock. The items may have been sourced from the bedrock, or from a beach or glacial till pebble of the same type of rock. (The following items belong to this category: 2:3, 10:1, 29:1, 199:1, 199:11, 201:18, 209:1, 209:2, 265:1, 322:1, 334:1, 402:1, 438:4, 487:1, 4022:3, 4022:4, 623:2, 833a:1.) Category 2. includes those items composed of nonisland bedrock type, but of a rock type known in the

160 high island: excavation of an early medieval monastery

vicinity, on neighbouring islands or on the nearby mainland. These items were most probably picked up by the monks as beach pebbles or found in the glacial till on the island or nearby. Despite the fact that these rock types are not found in the island bedrock, they could have reached the island during glacial movements. Glacial striae on the island indicate ice movement towards the north-west, so metamorphic and granite rock debris could have been carried over from Connemara. (The following items belong to this category: 66:1, 208:2, 256:1, 430:2, 450:1, 457:3, 464:1, 478:2, 478:3, 499:2, 4002:1, 4022:5, 4034:1, 4071:1.) Category 3. refers to items of a rock type not found in the local area geology and unlikely to have been introduced in the glacial till. (The following items belong to this category: 4:1, 4.4, 4:5, 38:1, 208:1, 211:1, 214a:2, 215:2, 351:1, 425:3, 448:3, 4030:1, 4030:3, 830:1, 889:2.) Of the 49 items examined, it is suggested that just over two-thirds could have been sourced by the inhabitants of the monastery (nineteen of possible island origin (Category 1.)) and fifteen from a local area origin (Category 2.); onethird (fifteen items) were probably imported. Identification and description of all items can be found in Section 6. Identification of the three altar stones (4:1, 4:4, 4:5) and the hone (208:2) was not possible by visual analysis alone. The items were subjected to non-destructive chemical analysis by X-ray fluorescence, and the results are outlined below.

X-ray analysis of selected small finds The spheroidal nodules (4:1, 4:4, 4:5) and hone (208:1) The dark, fine-grained artefacts (4:1, 4:4, 4:5) and 208:1 were difficult to identify by visual means alone, so nondestructive chemical analysis by X-ray fluorescence was undertaken. Samples were irradiated for two minutes with a hand-held energy-dispersive X-ray spectrometer, with automatic semi-quantitative calculation of element concentrations (Table 6.3). The machine was calibrated for heavy metals, so light elements could not be determined.

All samples have high iron content, 25–38% Fe, and are also enriched in Mn and Ba, but low in Pb, Cu, Zn and Ni. These analyses indicated that the three elipsoidal stones are iron nodules. These form by precipitation of iron carbonate or oxide in pore spaces during diagenetic alteration of organic-rich sedimentary rocks. They are a common feature of “Coal Measure” sedimentary successions (in Leitrim and Carlow, for example), but also occur in similar, non-coal-bearing rocks, for example in South Clare. A nodule from coal measures from the Geological Survey of Ireland (GSI) collection is presented in Table 6.3 for comparison. The item identified as a hone (208:1) gave a very similar spectrum to the nodules, although its flat, rectangular shape is different. It is likely that this is also from an iron nodule.

6.4 Metal The coastal location, coupled with unsuitable soil conditions, resulted in a highly corroded and poorly preserved assemblage of metal artefacts. The assemblage is described under each metal type in the order of: copper alloy, iron, and miscellaneous modern finds of copper/brass. The iron nails are described and catalogued at the end of this section. Copper alloy The only copper-alloy item that could be identified to object type was a small, decorated strap (95E124:201:8/ 95E124:202:3) that may have been part of a book clasp. The object, which was broken into two pieces, came from post-monastic period deposits datable to after the late 12th or early 13th century and which had been subject to disturbance from burrowing rabbits inside the church. The remaining items are small, undiagnostic scraps and crumbs of copper/copper alloy, which lack any original edges or features; the largest (95E124:8:3) measures 10mm x 5mm. A number come from early contexts. While the scraps are unidentifiable, there is no evidence to indicate that any substantial objects might have decayed.

Table 6.3 Results of X-ray fluorescence (in ppm) on the spheroidal nodules (4:1, 4:4, 4:5) and the hone (208:1). <LOD = below detection limit; - = not analysed.

Sample 4:1 4:4

Pb 29.14 28.94

As 20.58 <LOD

Zn Cu Ni Fe Mn Ba Ca <LOD 49.92 <LOD 369674.13 1476.57 1606.58 <LOD <LOD 192.41 384798.75 1528.62 2050.88 <LOD

S <LOD

P <LOD

4:5 208:1

32.4 19.57

<LOD 14.44

92.18 <LOD 194.44 259173.7 <LOD <LOD <LOD 319185.72

<LOD <LOD

GSI 39.66 <LOD NODULE LCF

46.16

49.24

1028.71 2555.82 <LOD 1042.93 1962.79 <LOD

206.44 271222.16 8760.09 1078.52

the finds 161

Fig. 6.12 Copper-alloy and iron artefacts and miscellaneous modern find.

95E124:600:1

95E124: 201:8/202:3

50mm

95E124:201:2

95E124:201:8/95E124:202:3 (Fig. 6.12)

Fitting or clasp, found in two parts. L-shaped strip with loop terminal at one end and spatula-shaped expansion (W: 8mm) at the opposite end, which is probably broken. The object is cut from a thin strip of rectangular-sectioned metal; there are short projections at the neck immediately below the perforated terminal. It is not clear if the piece was L-shaped originally, or if it has been bent subsequently, but the former seems more likely. Paired incised lines define the long edges of one face; the surface of the expanded end has been eroded by bronze disease and it is not clear how, or where the decorative panel terminated. The most likely interpretation is a book clasp. It is unlikely to have been a looped shank (e.g. for a key) as only one side is decorated. Stem L: 70mm; W: 5mm; T: 1.5mm; Loop D: 11mm; Perforation D: 5mm. Area 2, the two fragments were found in a rubble deposit (F201) and in the mortar floor (F202) inside the church. Scraps with no diagnostic features (Not illustrated)

Find no. 95E124:224:1

Area Context Area 2 Silver/grey schist clay over burnt debris (F221) 95E124:232:1 Area 2 Charred layer possible ‘old ground surface’. 95E124:55:1 Area 1 Brown soil within deposit sub-church 95E124:8064:1 Area 8 Fill of wall F813, south of Cell B 95E124:8:3 Area 1 Cross-slab base (cat. no. 51) found in east wall of church enclosure

Iron artefacts The iron artefacts are extremely corroded, the soil conditions not being conducive to the preservation of iron. Their fragmentary condition may also reflect the degree to which the objects were worn before being discarded. Even after X-ray analysis, most were not identifiable to object type. The majority of the pieces were found in rubble layers. A few possible knife blades and a collection of carpentry nails (see catalogue at end of this section) were the only recognisable iron artefacts. Some of these may have been made in the smithing hearth (F352) found beneath the floor of Cell B. A small clay crucible (347:1) found in the fill of the same hearth suggests that non-ferrous material was also being worked; analysis of the crucible (Young, this volume, p.169) did not shed any further light on the nature of that activity. 95E124:201:2 (Fig. 6.12)

Blade, incomplete. Fragment of a rectilinear blade now in two pieces. Neither the haft end or the tip is preserved. The blade is triangular in section; the shouldered edge appears more worn than the cutting edge. Possibly a ‘strap’ razor of relatively recent date. L: 106mm; W: 26mm; T: 2mm. Area 2, from rubble – the same context produced a clay pipe stem. Artefacts (Not illustrated) 95E124:243:1

Blade, incomplete. Two pieces, the larger fragment is from the tip of a narrow blade, worn away at an angle on the cutting edge. The smaller (18mm) fragment is probably from the same implement. L: 33mm; W: 9mm; T: 1.2mm. Area 2, from rabbit disturbance inside the church.

Fig. 6.1

162 high island: excavation of an early medieval monastery

95E124:445:1a/b

Possible blade fragments. Two flat, blade-like pieces, crosssection not visible. The largest piece (L: 40mm, H: 12mm) is rectilinear in shape for most of its length. Area 4, from the fill of the water-collection and drainage channel (F8025), outside Cell A. 95E124:436:3

Possible blade fragments. Flat, rectilinear strip and small chip, possibly a wrap-over blade that has split. L: 36mm; W: 22mm; T: 1.5mm. Area 4, from rubble. 95E124:805:3

Possible blade fragments. Flat, rectilinear strip (27mm x 21mm), cross-section not visible. Two other smaller pieces may be part of the same object, but no diagnostic features are evident. Area 8, from rubble. 95E124:401:1

Decorative fitting, incomplete. Narrow, flat, curved strip with countersunk panel (W: 3mm x depth: 2mm) on upper face. The remains of at least two staple holes (3mm x 1mm) are visible on the panel. Possibly part of a handle, or a furniture fitting. L: 75mm; W: 8mm. Area 4, from rubble. 95E124:425:1

Animal shoe, incomplete. Part of the front/side section of a small, corroded iron shoe, flat on both faces. Two nail holes revealed by X-ray. Possibly a donkey shoe. L: 90mm; W: 30mm; T: 7mm. Area 4, from rubble. Unidentified objects (Not illustrated) 95E124: 201:4

Four fragments: the largest piece is rectangular in shape and curved in profile. None of the edges seems original; there are traces of bronze adhering to one of the longer edges. The piece looks like it has sheared off a round-sectioned rod. The remaining fragments are small (10mm) crumbs of corroded iron. Area 2, from rubble. 95E124:221:1

Rectilinear fragment, broken at both ends and of irregular quadrilateral cross-section. One end tapers to a bent point, the opposite end is angled. The piece is too thick to have been a blade. If it is a nail shank, it is grossly misshapen. L: 40mm; W: 12mm; T: 9mm. Area 2, from the burnt debris beneath the church floor. 95E124:313:1 and 314:1

Thin, flat, rectilinear strap with cutaway shoulder at one angle and a projecting spur (12mm x 7mm) at the diametrically opposite angle. The piece does not have the tapering profile of blade 201:2, but is of broadly similar dimensions

to it. It may be an unfinished blade or an offcut. L: 48mm; W: 26mm; T: 1mm. Area 3, from rubble layer in Cell B. 95E124:457:2

Four heavily corroded fragments (largest fragment L: 40mm, W: 26mm). It is unclear if the object was originally roundended and hollow, if it has split post-deposition, or if is the remains of a shank (D: 6mm) from which the iron has bolted. Area 4, from burnt debris within fill of extended monastic enclosure wall. 95E124:199:12/13

Square-sectioned rod (L: 25mm, T: 6mm) with a small projection (3mm x 2mm) at one end. Likely to be part of a nail shaft, but the projecting tip could also have been used for some other purpose. 199:13 is a small (L: 15mm), tapering fragment, possibly part of the same object. Stray finds from rubble removed from general vicinity of church enclosure wall. 95E124:478:5

Corroded piece of iron, shape suggests it may be a bent tacklike nail, with head diameter of c. 20mm. Area 4, from burnt residues outside Cell A. The following consist of small scraps/lumps of iron without any distinguishing features. Most are less than 10mm in size and some (e.g. 4020:2, which consists mainly of iron corroded onto burnt stone) may be associated with smithing: 95E124:4:8, 29:2, 93c:1, 232:1, 430:5, 435:1, 436:1, 436:2, 438:3, 448:1, 448:2, 457:5, 457:6, 461:1, 4020:2. Miscellaneous modern finds

Copper/brass 95E124:402:2

(Not illustrated) Penny, copper, English of George V, dated 1922. Area 4, from rubble. 95E124:600:1 (Fig. 6.12)

Sheet copper/copper alloy, rolled piece. Found in sod layer. L: 60mm; D: 7mm. Area 6, from rubble. Buttons (Not illustrated) Five brass buttons were found in late contexts. Four have backmarks, and the fifth has the number ‘64’ on its upper face. 802:7 may be a uniform or club button – the back has a vine pattern with shamrocks and fleur-de-lis. The practise of using die-stamped backmarks began in the second half of the 18th century and continues to the present day. 95E124:600:3

Brass button with backmark ‘LLOYD STANDARD COL-

the finds 163

OUR’, attachment loop broken off. The button appears to be of relatively recent date. D: 20mm. Area 6, from uppermost level of sod. 95E124:801:5

Brass button, identical to 600:3, but with attachment loop preserved. D: 20mm. Area 8, from rubble. 95E124:201:7

Brass button/stud with ‘64’ on front face; no backmark. D: 17mm. Area 2, from rubble. 95E124:801:2

Brass button/stud. Backmark includes the word ‘QUALITY’, but the remainder is illegible. From rubble. D: 14mm. Area 8, from rubble. 95E124:802:7

Brass button, attachment loop broken off. Back is decorated with fleur-de-lis and shamrocks and is branded ‘STAND COLOUR’. Possibly a club or uniform button. D: 20mm. Area 8, from rubble.

Classification of iron nails Thirty-seven recorded finds include one or more nails or parts of nails; a few other highly corroded scraps may also be nail fragments. Apart from a modern steel nail, only two nails (703:1 and 873:1) are possibly complete and neither comes from a definite monastic context. The nails have been divided into three types, all commonly found on early medieval sites, although none is chronologically significant. Apart from a few distinctive forms, handmade nails remained largely unchanged until their replacement by machinemade nails in the 19th century. The vast majority of the High Island examples belong to Type 1 (heavy carpentry nails), and almost all of these seem to have had heavy circular or oval heads (Type 1a). Lighter carpentry nails (Type 2) are poorly represented, with only three possible examples recorded. Only one likely example of Type 3 (a short tack) has been identified, although some of the nails lacking shanks could possibly have been of this form. The remainder of the assemblage consists of broken shank pieces, but the poor condition of the pieces precludes any further analysis.

Table 6.4 Type 1 nails (not illustrated).

Reg. no. 95E124:201:20 95E124:202: 2

Type 1a 1

95E124:430:4

95E124:436:5

95E124:457:1

95E124:703:1

95E124:873:1

95E124:4057:1

1a?

Description Spalled off top of round-headed nail. Two small pieces, X-ray suggests the larger may be part of head/shank junction of a heavy nail and the smaller is a shank piece. Corroded shank and broken-off head of heavy nail. Head sub-circular, square-sectioned shank. Heavy square-headed nail with attached portion of very corroded shank. Very corroded piece, but may be head/attached portion of shank of heavy nail. Rectangular-headed heavy nail, bent at tip, but possibly complete. Tapering shank has rectangular section. Large, ovoid head of heavy nail; very corroded shank, bent at tip, but possibly complete. Corroded and bent nail, resembling fiddle-key type, but too robust for this form; the head was probably flattened when the nail was being removed.

Head D: 18mm Estimated D: >17mm

Shank L: 13mm, T: 4mm

Context Area 2, from rubble. Area 2, from mortar/ clay floor of church.

D: c. 28mm

L: 33mm T: 8mm

Area 4, from rubble.

W: 20mm, T: c. 2mm

L: 30mm Area 4, from rubble.

c. 22mm x 17mm

L: 37mm

25mm x c. 18mm, T: 5mm 28mm x 17mm c. 26mm x 17mm

Area 4, from burnt deposit within fill of extended monastic enclosure wall. L: 40mm Area 7, from silty clay on T: >7mm floor of wall chamber. x 6mm L: 42mm Area 8, from clays overlying path around Cell B. L: Area 4, from deposit 47mm, south of and abutting D: 6mm primary monastic enclosure wall (F4040).

164 high island: excavation of an early medieval monastery

6.5 Archaeometallurgical residues Tim Young

Introduction Material of metallurgical origin indicates that a pit below Cell B (Fig. 5.36, no. 1, F352) was associated with the smithing of iron and probably also for the working of non-ferrous metals. The pit itself yielded an assemblage of fines, indicative of accumulation within a smithing hearth. The morphology of the feature is compatible with it being a hearth. The adjoining excavation, Area 8, yielded a small plano-convex smithing hearth cake, compatible with the sort of macroscopic slag that might be expected to have been generated alongside the within-hearth fines. The same area also yielded a fragment of the tip of a tuyère, also of a type compatible with a smithing hearth. The hearth pit produced remains of an incomplete thin-walled crucible, apparently bell-shaped, but possibly part of lidded form, in a soft white (pipe clay) fabric with an organic (hair?) temper and sporadic large grains of mica-schist. The crucible may have failed before use. The submitted material was, however, dominated by corroded iron fragments (mainly nails) and examples of natural iron mottling and panning.

Methods All material was examined under a low-magnification stereo microscope. Material was selected for further analysis as appropriate. Crucible material was examined to determine the chemical composition of surficial slag phases using the Energy Dispersive X-Ray Fluorescence (ED-XRF) Eagle II instrument at English Heritage, Fort Cumberland, Portsmouth, with the kind assistance of Dr Justine Bayley. The samples were analysed in a normal atmosphere to produce only qualitative indications of metal presence.

Whole-specimen chemical analysis for major elements was undertaken using fused beads on the Open University Earth Science Department’s Wavelength-Dispersive X-Ray Fluorescence (WD-XRF) system. Whole-specimen chemical analysis for minor and trace elements was undertaken on the Inductively-Coupled Plasma Mass Spectrometer (ICP-MS) in the School of Earth, Ocean and Planetary Sciences at Cardiff University. The chemical analysis of the material mentioned in this report is attached at the end of this section (Table 6.9). The chemical analyses and catalogue of all submitted fragments has not been included in this report; for anyone wishing to view this material, please consult the Archive of the National Monuments Services, Department of Arts, Heritage and the Gaeltacht in Dublin. A catalogue of the iron nails is included in Section 6.4.

Results Natural Materials The collection contained a large number of specimens of iron oxide concretionary materials, ranging from diffuse mottles through to substantial iron-rich crusts (in some cases well-enough formed to be considered as bog iron ore), as well as some magnetic iron oxide pieces, probably derived from weathered, copper-bearing mineral veins. Analysis was undertaken of a specimen of a welldeveloped iron pan crust (F4039:1) and of a piece of vein mineral (F37) to determine their composition, in order to aid the investigation into whether the iron slags recovered from the site might have been from smelting. Iron pan The major element analysis of the material shows a moderate degree of iron enrichment (36.3% of iron quoted as Fe2O3), with a low manganese content (0.24% MnO; bog

Table 6.5 Type 2 nails (not illustrated).

Reg. no. 95E124:299:1

95E124:452:1

95E124:811:1

95E124:825:1

Type Description 2 Two pieces, X-ray shows the largest is square-sectioned shank with attached part of small, flat head. Second piece is a stone with iron adhesions. 2a Round/oval-sectioned shank, flattened expansion at one end may be the remains of the head. 2a? Four pieces, three of which are shank fragments (max. L: 24mm). The fourth piece may be part of a circular head with attached bent shank. 2a Round/oval-headed light form of nail embedded in iron corrosion.

Dimensions c. 10mm x 7mm L: 20mm, T: 5mm

Context Area 2, unstratified.

D: c. 10mm

L: 25mm D: 6mm

D: 22mm

Area 4, from post-abandonment rubble. Area 8, from rubble.

D: c. 17mm

L: 24mm T: 4mm

Area 8, from rubble.

the finds 165

Table 6.6 Unclassified and modern nails (not illustrated).

Reg. no. 95E124:202:1

Type Description Possibly 1a or 3 (a) Head and (b) portion of rectangular-sectioned shank. Head may have been circular.

95E124:436:4

Possibly 1a or 3 Circular or oval-shaped head and Head, 30mm x 25mm small attached portion of shank Possibly 1a or 3 Oval head and short attached Head, 31mm x 25mm portion of shank. Shank: L: 8mm, D: 5mm

95E124:803:1

95E124:825:2

95E124:4014:1

Possibly 3

Two pieces, the smallest (a) seems to a tapering shank, (b) is very corroded, but may be the relatively large head of the same nail. Possibly 1a or 3 Large oval head and short (<10mm) section of shank.

Dimensions (a) Head, D: 15mm (b) Shank, L: 26mm, T: 4mm x 3mm

Context Area 2, from mortar/clay floor of church.

Area 4, from rubble. Area 8, from clay packing around Cell B. Head, estimated D: 20mm. Area 8, from Shank, L: 27mm, T: 7mm rubble.

Head, 27mm x 21mm

95E124:4022:1

Unclassified

Very corroded fragment broken in two pieces, may be head/ shank fragment.

Overall L: 35mm

95E124:35:2

Modern steel nail

Six-sided head, thin tapering shank.

Head, W: 23mm Shank: L: 35mm; T: 6mm

Area 4, from layer north-west of extended monastic enclosure wall (F4025). Area 4, from fill of extended monastic enclosure wall (F4025). Area 1, from north wall of church enclosure.

Table 6.7 Shank pieces (not illustrated).

Reg. no. 95E124:1:3 95E124:1:4 95E124:1:5 95E124:201:17 95E124:214a:1

Description One piece, square-sectioned shank.

Dimensions L: 35mm T: 5mm One piece hollow round-sectioned shank, L: 28mm most of metal has bolted. D: 6mm One piece of square-sectioned shank/shank L: 30mm head junction. T: 5mm One piece. Bent, tapering nail shank, L: 46mm square section. T: 7mm One corroded lump; X-ray shows shank Shank, with slight expansion, possibly part of nail L: 29mm head, at one end. T: 5mm

Context Area 1, from rubble. Area 1, from rubble. Area 1, from rubble. Area 2, from rubble. Area 2, found in pseudo Grave 10 (F214) inside church.

166 high island: excavation of an early medieval monastery Continued Table 6.7 Shank pieces (not illustrated).

Reg. no. 95E124:215:1

95E124: 215:5

Description Four pieces, three of which appear to be shank fragments; X-ray shows the largest (L: 33mm) to have protrusion, possibly part of head, at one end. The fourth piece is a mortar fragment with iron staining. Fragment of rectangular-sectioned shank now broken into two pieces.

95E124:221:2

One piece, X-ray shows wide (9mm) shank, possibly rectangular-sectioned and blunted at one end. 95E124:445:1 c/d/e Three small pieces, probably corroded shank fragments. 95E124:457:1 One broken piece.

Dimensions Shank, T: c. 6mm

Context Area 2, from soil deposit under mortar/clay surface (F202) inside the church.

L: 20mm T: 5mm x 3mm L: 50mm

Area 2, from soil deposit under mortar/clay surface (F202) inside the church. Area 2, from burnt debris over prechurch ‘paved area’ (F245).

L: 25m T: 6mm L: 47mm

95E124:457:4

One shank piece and three associated scraps.

L: 30mm D: 10mm

95E124:478:1

L: 40mm D: 6mm D: 7mm

Area 6, from rubble.

95E124:803:2

Two pieces, largest is part of roundsectioned shank; the second is a small (20mm) corroded lump. Three small corroded lumps of iron; roundsectioned shank visible on one piece. Broken piece of tapering shaft.

Area 4, from fill of water-collection and drainage channel (F8025). Area 4, from burnt debris, fill of extended monastic enclosure wall (F4025). Area 4, from burnt debris, fill of extended monastic enclosure wall (F4025). Area 4, from burnt residues outside Cell A.

95E124:805:2

Bent shank now broken into two pieces.

L: 28mm D: 9mm L: 62mm D: 6mm

Area 8, from clay packing around Cell B. Area 8, from peat within rubble.

95E124:806:1 95E124:811:3

One piece, broken at both ends. Two pieces of broken square-sectioned shank.

95E124:618:1

95E124:822:1

95E124:825:3 95E124:829:1 95E124:4020:1

(a) L: 37mm Area 8, from rubble. T: 5mm; (b) L: 19 mm T: 6mm Area 8, from peat/rubble. Square-sectioned shank piece with attached Shank: L: 50mm corroded lump of iron (33mm). One end of shaft is expanded and flattened. Possibly T: 5mm a nail. One piece, broken across one edge. L: 29mm Area 8, from peat/rubble. D: 5mm Tapering shank in two pieces. L: 38mm Area 8, from peat. T: 9mm One broken piece. L: 31mm Area 4, from fill of extended D: 7mm monastic enclosure wall (F4025).

the finds 167

Possible nails

Reg. no. 95E124:438:2

Description One small rod-like piece.

95E124:436:6

One piece of circular or Possibly part of a ovoid section. heavy nail. Four small pieces. Possibly nail shank(s). Four small pieces; Probably part of a the largest is slightly nail. domed. Four pieces, one of Possibly part of a nail. which appears to be part of a rounded shank.

95E124:888:2 95E124:4030:2

95E124:4:7

Classification Possible nail shank.

iron ores typically have a fairly high manganese content). The loss on ignition was very high (26.3%, if all iron was present as FeIII), suggesting that the iron was present in strongly hydrated minerals. The silica:alumina ratio is much lower (1:4) than for the slag material described below, suggesting a clay-rich mineralogy for the host sediment. The material shows most elements broadly in line with their contents in the slag (although diluted by additional iron in the slags), indicating that despite some mineralogical differences, the host of the iron crust and the material contributing silicate material to the slags were broadly similar. Phosphorus is high in the iron crust (5.04% quoted as P2O5). Iron oxides (like goethite) are known for their ability to scavenge phosphorus under certain circumstances. The rare earth elements (REE) show an upper crust normalised profile (Fig. 6.13), with a humped profile centred on europium. The light REE show a slight relative depletion, but with a more marked relative depletion for the heavy REE. Table 6.8 Summary of slag fines recorded from the smithing pit F347/F351.

No.

Wt.

Sintery material Prills

136 21

97.95g 34.70g

Blebby slags Slag spheroids Slag films Flake hammerscale Other fines Corrosion around iron

124 33 7 9

114.85g 4.22g 1.31g 0.6g 6.36g 1.47g

Dimensions L: 35mm, W: 7mm

L: 52mm, D: 25mm x 15mm

Context Area 4, from fill of extended monastic enclosure wall (F4025). Area 4, from rubble.

Area 8, from fill of leacht (F887) outside Cell B. L: 18mm, W: 14mm Area 4, from fill of primary monastic enclosure wall (F4040). D: 7mm x 5mm Area 4, from soil deposit north side of church, over paved surface (F29).

Magnetic vein ore The major element analysis shows that iron is the dominant element (89.1% calculated as Fe2O3). The loss on ignition is 8.7% (when iron is calculated as FeIII). This suggests that the iron mineral is mainly hydrated (the LOI for pure goethite or lepidocrocite, the common iron oxyhydroxides, FeO.OH, is 10.1%), and the fairly strong magnetic nature of the specimen suggests either the presence of a small proportion of magnetite (Fe3O4) or of maghemite (γ-Fe2O3). The upper crust normalised REE profile (Fig. 6.13) shows very low levels of REE (∑ REE is only 6.92ppm), with a fairly constant increase in depletion from the light REE to the heavy REE.

Archaeometallurgical materials The ‘Hearth Pit’: slags Dull, blebby slags, many of which appear to be dominated by lining material, dominate the assemblage of residues from the pit. Some of the blebby material is fused and mingled with the porous, sintery-appearing type of residue, which may form on the base of both smithing hearths and smelting furnaces. Alongside these mainly low-density slags there are prills and spheroids of denser slag, together with some shiny dense slag films. Residues resembling these may be found in both smithing hearths and smelting furnaces. Although the highly magnetic nature of this material and the presence of a small quantity of flake hammerscale favoured an origin in a smithing hearth, some analytical work was undertaken to allow a comparison of the composition of the potential iron ore resources of the island with the slag composition, in order to determine whether an alternative interpretation of the slags as smelting residues would also be possible. The major element analysis of the slags shows that

168 high island: excavation of an early medieval monastery Fig. 6.13 Upper crust normalised rare element profiles for samples from the metal-working pit and two potential ore types. (Normalisation factors after Taylor and MacLennan 1981.)

the sintery material, the prills and the small porous slag piece (which had tentatively been identified as possible ore), are all slags and have a closely similar composition. These three have an extremely high iron content (77–81% calculated as FeO) and a fairly high phosphorus content (0.9–1.3% P2O5). Silicon and aluminium are extremely low (SiO2 5.4–9.8%; Al2O3 1.7–3.1%). The silica:alumina ratio is remarkably constant (2.6:3.1). The lining slag, in contrast, has a low iron content (32.7% expressed as FeO) and much higher silica and alumina contents (38.9% and 11.7%, respectively). The silica:alumina ratio is close to those of the other slags (3:31). The other elements expected to be derived from the lining (magnesium, calcium, sodium, potassium, titanium) are all elevated in the lining slag. These major element data suggest that the slags are effectively mixtures of a slightly variable, but broadly com-

95E124:852:1

95E124:347/351:1

Fig. 6.14 Crucible and tuyère tip.

parable silicate component, with iron oxide. The extremely iron-rich nature of the slags indicates a low degree of fluxing by silicate material, a feature common when a tuyère is used in a smithing hearth, but which is not encountered in bloomery iron smelting. The minor and trace element content of the slags is largely derived from the lining of the hearth (or melting of the tuyère tip). The majority of the elements appear to maintain a fairly constant ratio with aluminium. The upper-crust normalised REE profiles (Fig. 6.13; normalisation factors after Taylor and McLennan 1981) show a fairly horizontal profile for the lining slag, compatible with its origin from melted soil or sub-soil. The profiles for the slag prill, the sinter and the porous slag (which had been considered as a possible piece of ore) are very similar (i.e. horizontal), suggesting their silicate components may also be derived from sediment. In contrast,

25mm

the finds 169

both the potential iron ore materials show a slight relative heavy REE depletion. The iron pan material has an overall REE content close to upper crust average, but slightly depleted in the heavy REE elements. The magnetic iron ore sample shows a very low REE content, varying from 0.07 of upper crust levels for La to 0.03 for Lu. These properties indicate the slags are unlikely to have been derived from smelting either of the ores. The ‘Hearth Pit’: crucible 95E124:347/351:1 (Fig. 6.14)

The hearth pit contained four sherds, three of them conjoining, of a small crucible: 25mm high in a very finegrained pale fabric (pipe clay), very pale buff near the surface, but pale grey in the core (F347/F351:1) (Fig. 6.14). The fabric bears sporadic large particles (up to 3mm) of micaceous schist (compatible with a local origin) and a widespread organic temper, apparently of hair. The body of the crucible shows two layers, with a tendency to delaminate. This appears to be an incoherency in the clay during fabrication rather than a deliberate feature. A soft pale fabric is not common in Irish crucibles, but has been recorded in early medieval material from Dunmisk, Co. Tyrone (Ivens 1989). The main sherd includes the base of the crucible, which has a maximum thickness of about 6.5mm. The wall of the crucible thins rapidly upwards, being less than 3mm thick at 18mm above the base and 2mm thick immediately below the rim. The rim is rather widely splayed, with an apparent diameter of about 32mm. Only a short length of rim is preserved, however, so it is possible that this flared profile is a result of the preserved section passing up through a pouring lip. The crucible is not complete enough to allow firm identification of its morphology. It is possible that the flared profile represents the pouring lip of a lidded crucible, rather than an open form. In particular, the profile of the present material bears close comparison with that of the pinched crucibles illustrated from Dunadd (Lane and Campbell 2000; Type D, illustration 4.48, no. 187). A sherd from close to the rim was examined for any metal traces using the ED-XRF, but no metal contamination was found either on the inner or outer surface. Both inner and outer surfaces of all the sherds were visually inspected for indications of slag or vitrified deposits, but none was found. It seems likely that the flaw in the centre of the fabric may have caused the crucible to fail before use at high temperature. Other materials Only a very limited number of archaeometallurgical materials were recorded from other contexts:

95E124:618:2

(Not illustrated) 4.9g of moderately dark, vesicular, dense slag, attached to a small area of oxidised lining. (Area 6, from rubble) 95E124:8099:1

(Not illustrated) Probably the majority (c. 80%) of a small, dense smithing hearth cake (SHC), top smooth, base rough, with slight dimpling onto fuel bed at margins. 330g. (Area 8, found in spoil) 95E124:852:1 (Fig. 6.14)

34g fragment from the lower side of the tip of tuyère. The curvature is very low, suggesting that this is the flattened base of a tuyère with a sub-circular cross-section. (Area 8, from silt-fill of rock-cut trough, F845) 95E124:888:2

(Not illustrated) Four tiny pieces (total 1.58g) of vesicular lining slag bearing sand grains. (Area 8, from a deposit of dry, fragmented schist within leacht, F887) Interpretation The chemical studies clearly indicate that the slag fines are from iron-working, not iron smelting. The ‘hearth pit’ and its residues can be compared with a similar Iron Age assemblage from feature 1056 at Ballydavis 2 Site A, Co. Laois (unpublished report, Young 2005). The fines assemblage is also close to those recovered from hearths and dumps at Coolamurry Site 7, Co. Wexford (unpublished report, Young 2006), which are of uncertain age. Assemblages with prills and spheroids, representing slag dripping through the fuel in the hearth, indicate a fines assemblage generated within the smithing hearth, rather than the common style of smithing fines assemblage, dominated by hammerscale, which is generated on the floor of the workshop. That does not necessarily indicate that the deposit is actually a smithing hearth; it may simply be a deposit formed from cleaning-out the smithing hearth. Consideration of the form of the feature, with a somewhat elongated shape, a shallow depth and a well-marked edge (at least on some sides) does suggest, however, that it may indeed have been a smithing hearth. There is insufficient certainty in the plan of the feature to be clear of the orientation of the hearth, but it is possible that it was blown from the north, with the area south of the possible stone partition being the base of the hearth proper, with the small depression being associated with preparation of a hollow for slag accumulation immediately in front of the tuyère. In this scenario, the stakehole north of the partition might be associated with mounting the bellows. The SHC is rather small (with an estimated maximum original weight of about 410g), suggesting it was produced

170â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Table 6.9 Chemical analyses of samples from the metal-working pit and two potential ore types. ((a) major elements, expressed as weight% oxide. Iron is expressed as the FeII oxide. LOI = loss on ignition. For the iron pan and magnetic ore specimens, iron is also quoted as the FeIII oxide, with corrected LOI. (b)â&#x20AC;&#x201C;(d) trace elements, expressed in parts per million (ppm). nd = not determined, < = below detection.) (a) Sample

SiO2

Al2O3

FeO

MnO

MgO

CaO

Na2O K2O

TiO2

P2O5

LOI

Total

Fe2O3

LOI

4039:1 Iron pan

15.27

10.90

32.54

0.24

1.38

1.42

1.11

0.96

0.34

5.04

29.94

99.13

36.16

26.32

F37 Magnetic ore 0.49

0.13

80.15

0.02

0.20

0.16

0.04

0.05

0.02

0.14

17.59

99.00

89.08

8.67

F351 Prilly slag

7.52

2.93

81.37

0.11

0.59

1.38

0.20

0.16

0.19

1.15

4.36

99.95

F351 Lining slag

38.86

11.74

32.66

0.14

1.47

3.46

2.52

3.19

0.73

1.30

3.39

99.47

F351 Sintery slag

9.81

3.05

76.88

0.12

0.69

1.53

0.25

0.30

0.21

1.30

6.15

100.28

F351 Porous slag

5.43

1.65

80.85

0.10

0.64

0.82

0.19

0.20

0.11

0.86

7.47

98.34

(b) Sample

4039:1 Iron pan

2.18

70.61

34.5

40.71 52.9

72.8

8.00

12.55 367.8

13.4

349.5

6.52

1.54

1.42

0.40

311.5

F37 Magnetic ore 0.32

0.77

34.3

22.05 131.6 52.3

1.82

0.95

55.9

0.9

323.8

1.42

1.17

16.44

0.28

103.1

F351 Prilly slag

0.52

21.01

8.1

0.96

6.18

2.04

95.1

5.9

288.4

2.79

0.47

0.18

0.55

150.9

F351 Lining slag

2.65

100.09

58.1

31.27 73.8

103.2 18.09

82.59 269.6

25.9

591.4

22.98

2.61

3.49

1.48

822.4

F351 Sintery slag

0.50

26.65

14.9

2.22

34.7

18.5

5.11

3.55

84.7

4.7

372.0

4.47

0.96

0.80

0.38

155.6

F351 Porous slag

0.12

7.40

8.8

1.06

58.9

30.9

2.59

1.57

41.2

2.6

352.6

2.56

0.87

2.96

0.31

148.2

4039:1 Iron pan

26.02

48.9

25.9

67.23

7.25

27.34 5.47

1.19

4.71

0.65

3.40

0.51

1.25

0.19

1.11

0.12

F37 Magnetic ore 2.05

2.77

0.32

1.08

0.19

0.04

0.16

0.02

0.11

0.03

0.06

0.01

0.06

0.01

F351 Prilly slag

6.86

14.25

1.62

5.82

1.10

0.23

1.00

0.15

0.90

0.17

0.52

0.08

0.55

0.06

F351 Lining slag

45.45

93.25

11.26

40.88 7.00

1.49

5.96

0.83

4.59

0.82

2.34

0.38

2.52

0.36

F351 Sintery slag

5.57

11.25

1.28

4.85

0.86

0.20

0.82

0.12

0.77

0.15

0.41

0.07

0.46

0.09

F351 Porous slag

3.34

6.07

0.71

2.67

0.48

0.10

0.46

0.06

0.39

0.08

0.23

0.03

0.22

0.04

(d) Sample

4039:1 Iron pan

9.26

0.49

7.36

9.19

9.26

F37 Magnetic ore 8.17

0.15

17.69

0.52

0.15

F351 Prilly slag

7.95

0.36

5.52

3.25

1.34

F351 Lining slag

16.08

1.77

5.21

11.20 6.09

F351 Sintery slag

9.86

0.39

6.34

2.54

1.24

F351 Porous slag

9.03

0.22

8.76

1.17

0.44

the finds 171

during blacksmithing, rather than during the refinement of raw iron. The assemblage contains little capable of providing a date for the assemblage. The crucible is not sufficiently complete to be identifiable, but some features suggest that it may be a form that would have been lidded; it is not clear whether, if lidded, it would have had a separate lid luted on (commonest in 6th to 8th centuries), or have been a pinched form with integral lid (which appear in the 7th century and possibly survive as late as the 10th). The presence of a tuyère suggests an early medieval date, for they are not well documented in the Iron Age. The crucible suggests a similarity with Dunadd Type D, although this is far from certain. At the Scottish site of Dunadd, the Type D vessels appear to have been used for silver and gold. Irish finds of pinched crucibles of this type include material from Ballinderry II crannóg (Hencken 1942), Carraig Aille II (Ó Ríordáin 1949) and Correneary (Davies 1942). The type is also known from the Brough of Birsay, Orkney (Curle 1982), and from various sites in Scandinavia. The significant quantity of iron panning and mottling is likely to be due to oxidation of iron-rich water within the sediment. Weathering of iron-rich rocks has probably generated most of the iron (although degradation of iron-working residues and iron artefacts may also have made a small contribution). The iron can be carried in solution when reduced (as FeII), but on oxidation precipitation of FeIII oxides will occur. Such oxidation may occur where reduced soil waters flow into a more oxygenated

95E124:482:1

25mm

environment, and in particular may occur at the limits of oxygenation during summer drying. Summary Although the fines that accumulate amongst the ash and fuel in the base of a smithing hearth are very similar to those in the base of a smelting furnace, the detailed analysis undertaken on the material from F347/351 provides strong evidence that these were formed in a smithing hearth. Circumstantial evidence for identification of the area as a smithy comes from other finds from the adjacent Excavation Area 8, which yielded several fragments of macroscopic smithing slag and part of the tip of a tuyère. The nature of the pit itself is unclear; it may be part of a smithing hearth, or it may be a pit used for disposal of the hearth ash. The presence of fragments of a crucible in the pit suggests that the smithy may have also undertaken some work with non-ferrous metals. The crucible-type is uncertain, but an extremely tentative reconstruction of its form may suggest it was a lidded type, perhaps a pinched, integral-lidded form, which would have a 7th- to 10th-century date range.

6.6 Coins The two silver coins have been identified by Michael Kenny of the National Museum of Ireland as a penny of possible Scandinavian origin, probably minted in the period between the reign of Edward the Confessor (1042– 1066) and the advent of the bracteates coinages of the early 12th century (95E124:320:1) and a later halfpenny of King John (1204–1211) minted in Dublin (95E124:482:1). PreNorman coinage is a relatively rare find on rural sites in the west of Ireland, but a Hiberno-Norse silver penny of the period 1020–1035 found at the small island monastery at Illaunloughan, Co. Kerry (White Marshall and Walsh 2005, 176), is an obvious parallel. A slightly later version of the same coinage was also recovered at the much larger monastery at Whithorn, in southwest Scotland (Hill 1997, 345). The origin of such coins may have been as gifts or tribute, indirect exchange or payment. As a store of value, coins would also have been carefully hoarded and this may be reflected in the (possible) find-spot of the High Island silver penny (320:1). Coins were in much wider circulation in the Anglo-Norman period and the silver halfpenny may have been a casual loss. 95E124:320:1 (Fig. 6.15)

95E124:320:1

Fig. 6.15 A silver halfpenny (top) and silver penny (bottom).

Base silver penny, 11th or early 12th century, in reasonable condition. The obverse has pseudo-lettering and a poorly executed human face with traces of a head, nose and eyes. The reverse has a voided cross within a circle. The origins of the coin are discussed below (Section 6.7)

172 high island: excavation of an early medieval monastery

in Limerick and two in Waterford. The King John Irish coinage of silver pennies was intended mainly for overseas use, but a number of halfpennies and farthings were also struck for the purposes of normal trading in Ireland (Seaby 1970, 22). D: 13mm. Area 4, found in the fill of the water-collection and drainage channel (F8025), east of Cell A.

6.7 A rare Viking-age penny from High Island Michael Kenny

Pl. 6.3 (a) Viking age penny (320:1): (a) obverse; (Photographic Unit, NMS).

Pl. 6.3 (b) Viking age penny (320:1): (b) reverse (Photographic Unit, NMS).

by Michael Kenny. D: 17.5mm. Area 3, found amongst collapse at the back of Cell B, in the location of a recess or aumbry in the west wall of the cell. While it is possible that the coin may have been secreted within the recess, this cannot be confirmed. 95E124:482:1. (Fig. 6.15; Pl. 6.3)

Silver halfpenny, the lower half of each face is worn. King John (1204–1211), minted in Dublin by Roberd Willem. The obverse has a crowned bust in triangle, no sceptre, other details unclear. Legend: Johannes Rex. Reverse has cross patee over a crescent within a triangle, with a star in each angle. O’Sullivan (1961, 16) noted that the National Museum Collection included 40 specimens of this halfpenny issue. Of these, 30 were minted in Dublin, five

The discovery of an unusual Viking-age coin in collapse at the back of Cell B has raised a number of interesting issues. The first relates to the status and function of the settlement during the Viking period and must be examined in the context of the overall archaeological evidence. The second is the question of the coin’s numismatic provenance: where, when and by whom was it issued and how did it find its way to such a remote location? Stylistically, it is quite unlike anything in the main Hiberno-Norse issues and indeed there is only one other similar coin known, a rather mysterious penny in the collection of the National Museum of Ireland. This latter coin, a silver penny from the Parsons Collection, came to light around 1920 and was acquired by the National Museum at a later date. This is the so-called ‘O Neill penny’, hitherto thought to be unique. It was named so by its then owner, Alexander Parsons, who read the obverse legend as NOIL REX M, suggesting that it was struck under the authority of a southern Uí Neill king – he read the REX M as the abbreviated Latin for King of Munster. From the retrograde reverse legend, which reads anti-clockwise, BLIINPISE ON LI, he suggested that the coin was struck at Limerick. Parsons published an article on the coin in the British Numismatic Journal at the time (Parsons 1921/2, 59–71), but interestingly made no reference to it in his later work on the chronology of what was then known as the Hiberno-Danish coinage. Later numismatists tended to disagree with Parsons. William O’Sullivan, in his work on Hiberno-Norse coinage, referred to the penny as being ‘to a certain extent a mystery coin’, copied from a coin of Edward the Confessor (O’Sullivan 1949, 33–4). Michael Dolley was blunter, stating that: ‘the coin is just not Irish. The prototype is an English penny of Edward the Confessor, struck a few years before the middle of the 11th century. The attribution to Limerick and to an O’Neill is one that flies in the face of history and of philology and one has only to place the coin in question

the finds 173

beside Hiberno-Norse which belong indubitably to the period 1065-1095 to realise that the Parsons coin has no place in the Irish series. That the piece and provenance are non-Irish now seems generally agreed and there can be little doubt that the coin belongs to the period of civil war when Svend Estridsen and Harald Hardrada were hammering out a modus vivendi between Dane and Norwegian’ (Dolley 1966). Dolley was supported by W.A. Seaby, who stated that the piece was not Irish but an apparently Norwegian coin from the period c. 1045 (Seaby 1972, 271). The fact that it did not have a provenance made further research difficult, as did the fact that it was struck with inverted and retrograde lettering. As William O’Sullivan had so carefully put it: ‘a representative number of specimens would give different spellings of moneyers’ and issuers’ names and the place of minting, from which it would be possible to give a fairly certain identification as to issuer, moneyer and place of issue’ (O’Sullivan 1949, 34). So the situation remained until 1999, when the High Island coin came to light. The coin that was discovered is virtually identical to the Parsons coin. The obverse and reverse legends are similar, as is the overall design (Fig. 6.15; Pl. 6.3 (a), (b). At first glance the coins would appear to have been struck from the same die. There are some traces on the High Island coin, however, which suggest that it could be a crude reworking of the Parsons die. Analysis shows it to actually have a higher silver content, 76.9% as against 65% for the Parsons coin. It has a correspondingly low copper content, 16% against 23.2% for Parsons. Since the island is rather inaccessible, it is highly unlikely that the coin is anything other than a genuine find from the period in question. The new find was examined shortly after it was found by the late Mark Blackburn of the Fitzwilliam Museum, an authority on Anglo-Saxon and Hiberno-Norse coinage. While his instinct was that it was probably Scandinavian, he did not express a strong opinion either way. More recently the coin has been examined by Andrew Woods from the same museum, who has been engaged in a study of the later Hiberno-Norse issues, especially the bracteates coins of the 12th century. He in turn brought the coin to the attention of numismatists in Scandinavia. While nothing quite like it has appeared in the Scandinavian hoards, one of those experts who had access to the illustrations, Kenneth Jonsson, has expressed the opinion that it might be Norwegian (K. Jonsson to A. Woods, pers. comm., 2011). In the absence of a larger body of provenanced finds, the question of whether the two pennies are Irish or Scandinavian must, for now, remain unanswered. On

the one hand, the fact that both have surfaced in Ireland – one from a precise archaeological site – suggests that they might be Irish. On the other hand, there is nothing else like them in the Hiberno-Norse series. In either case, the prototype is probably an Anglo-Saxon penny of Edward the Confessor. That in itself is of little assistance since many varieties of the Anglo-Saxon penny were copied in both Ireland and Scandinavia (see, for example, Ahlstrom et al. 1976, 12–16; Dolley 1966, 92–145). The date range is probably between the reign of Edward the Confessor (1042–1066) and the advent of the bracteates coinages of the early 12th century. Addendum: the above note was written on the understanding that the provenance and authenticity of the coin were beyond question. This has been thrown into doubt, however, by a recent discovery. A virtually identical coin turned up at the Fitzwilliam Museum in Cambridge in 2012. According to the owner, it came directly from an individual who was known to strike high-quality reproductions of early and rare coins. If the Cambridge coin is a modern forgery/reproduction, it also puts a major questionmark over the High Island coin and how it got to its final resting place. Recent and more detailed examination of the metallic content of the High Island coin has raised further questions. At this point, as investigations continue, it is difficult to say what credence should be given to the coin or its use as archaeological evidence.

6.8 Bone 95E124:208:3 (Fig. 6.16)

Bone pin, complete. Found with the mid-12th to early 13th-century burial (F208) and therefore associated with the hone (208:1) and weight (208:2). The pin has a tapering, curved shank; the tip and rudimentary crutch-shaped head are both worn. The back of the head splintered in antiquity. The piece is made from the shaft of a bovine long bone (identification by Margaret McCarthy) and the longitudinal cut marks are still visible. Three worn grooves on the right-hand edge of the shaft are probably a result of friction from the thread used to secure the pin while it was being worn. The wear pattern indicates that the pin was habitually worn on the right-hand side, which accords well with its find-place – over the right side of the pelvis. Further evidence of use is indicated by the fact that the pin continued to be worn even after the back of the head had splintered in antiquity. L: 80mm; D: 5mm; D of Head: 6mm x 4mm, T: 1.5mm. Area 2, found with burial F208 (Grave 9) inside the church.

174 high island: excavation of an early medieval monastery Fig. 6.16 Bone pin. 95E124:208:3

25mm

Fig. 6.17 Glass beads.

95E124:811:2

95E124:33:1

95E124:829:2

25mm

6.9 Glass beads

6.10 Miscellaneous modern

Glass beads are a relatively common find on medieval sites and the cobalt blue colour of 33:1 is typical of the late Iron Age and medieval periods. Beads of red and blue glass and of bone were found with some of the early phase child burials on Omey Island (O’Keefe 1993, no. 11). There was no definite evidence to link any of the High Island beads with burials, but all three may have come from secondary contexts; 33:1 was found on top of the enclosure wall; the two other coloured beads (829:2 and 811:2) came from the upper levels of the post-monastic period deposits suggested to be early 15th–late 20th century.

Glass (Not illustrated) Four small (25mm) scraps of green bottle glass were found, two of which came from the interior of Cell B, Area 3. Rabbit burrowing probably accounts for the intrusive context of 319:1. The glass (and the pottery from the same location) is probably associated with mining activity on the island during the 19th century.

95E124:33:1 (Fig. 6.17)

95E124:311:1

Glass bead, complete. Small, disc-shaped glass bead of cobalt blue colour. Central cylindrical perforation. D: 6mm; L: 3.5mm; Perforation D: 1.5mm. Area 1, found on top of the west wall of the church enclosure.

Area 3, found in rubble. 95E124:814:1

95E124:811:2 (Fig. 6.17)

Area 3, found in hearth residues above paved floor (F326) (intrusive).

Glass bead, complete. Small, barrel-shaped glass bead of milky blue colour. Central cylindrical perforation. D: 4.5mm; L: 4mm; Perforation D: 1mm. Area 8, from rubble/peat layer.

95E124:201:19

Area 2, found in rubble.

Area 8, found in hearth residues within peat/rubble. 95E124:319:1

95E124:829:2 (Fig. 6.17)

Glass bead, complete. Barrel-shaped bead of amber-coloured glass; central cylindrical perforation. D: 9mm; L: 10mm; Perforation D: 3mm. Area 8, from post-monastic peat/rubble.

Fig. 6

the finds 175

6.11 Pottery and clay pipes Clare McCutcheon

Introduction A total of twelve sherds of pottery were presented for study. Following identification and reassembly within and between contexts, this was reduced to eight sherds, representing three vessels, dating from the later 18th to the early 20th century. Methodology The material was identified visually and the information is presented in Table 6.10. This lists the quantity of sherds in each fabric type. In addition, the minimum number of vessels (MNV) is based on the rim/handle sherds, and the minimum vessels represented (MVR) is based on the more subjective visual differences in sherds present. The forms of the vessels are also listed, with the date range of the distribution of the fabric type. The pottery types (Not illustrated) Creamware, also known as Queen’s ware, was developed by Josiah Wedgwood c. 1760 to compete with the very large quantities of Chinese porcelain arriving into the European market, but it actually ended up putting tin glazed earthenware out of business. Creamware is distinguished by the slightly green tinge in the glaze collected at bases, etc. Table 6.10 Pottery identification,

Fabric type

Sherds MNV MVR Form

Creamware

Banded ware

Stoneware Total

4 8

1 3

Date

Footed 18th bowl Footed 19th mug Jar 19th–20th

Wedgwood further developed pearlware as an improvement on creamware in 1779, using cobalt blue in the glaze to whiten the ware. Pearlware is also the basis for a number of decorative motifs, such as transfer printing, mochaware and banded ware (Savage and Newman 1985). Banded ware is a decorative variation on a pearlware vessel, where a series of parallel horizontal lines in various colours are added around the vessel. Stoneware, originating in the 15th and 16th centuries in Germany, was later produced in large quantities in London. The value of this ware is its strength and impermeability, replacing glass bottles for many of the emerging polishes, inks and medicines of the

later 18th and 19th centuries. The surface is ‘glazed’ with the addition of salt during the firing process, while the possibility of firing the clay to a higher temperature than earthenware made it impermeable in and of itself. Clay pipes (Not illustrated) Two bowls and two stems were recovered from the site. One bowl (801:4) is a small, unmarked, Bristol-style bowl, very worn and well-smoked, as evidenced by the blackened interior. It is not possible to say whether the bowl was spurred or flat based, and the top of the rim is also broken, but might have been rouletted. This bowl probably dates to the later 17th/18th century. The second bowl is a portion of a 19th-century bowl with the mark of E. Hynes, Maker, Galway. The name of Hynes is well associated with Galway and pipes are known to have been made by Mary J. Hynes from 1881 and her son, Michael Hynes, ‘who appears to have been the last pipe-manufacturer in Galway’ (Norton 2004, 428). In addition, the rim of this large bowl was rouletted, as was the illustrated pipe of M. Hynes found at Eyre Square (ibid., fig.6.5:4f).

176

7 the cross-sl abs

Christine Maddern

Previous commentators

The first 19th-century commentator on the cross-carved and cruciform slabs found on High Island (Ardoileán) appears to have considered them of minor importance compared to the enclosure and the buildings. George Petrie, writing in 1820, gives a detailed description of the enclosure buildings, but passes rapidly over the cross-carved slab used as a lintel to the church doorway and the stones at the ends of the ‘ancient stone sculpture, like a pagan kistvaen, composed of large mica slates, with a cover of limestone’, merely noting that these stones were ‘rudely sculptured with ornamental crosses and a human figure’. Likewise, the ‘several rude crosses’ in the enclosure and the ‘several rude altars, or penitential stations, on which are small stone crosses’ receive only passing mention, possibly because Petrie’s single visit was necessarily fleeting (Petrie 1845, 424–7). Although lack of time may account for the short shrift given to the slabs, his comments appear to reflect the aesthetic taste of the time, which tended to compare early Irish carvings – and indeed, figural painting in manuscripts such as the Book of Kells (Dublin, Trinity College MS A.I.6 (58)) of around AD 800 – with classical exemplars, and therefore to find them crude and unrealistic. William Wakeman, who had visited the Island in 1839 with John Donovan as part of their work on the Ordnance Survey, drew ten of the cross-slabs, resulting in the largest graphic collection of these slabs for over 100 years. Twentyfour years later, Wakeman published his first account of the antiquities on the island, including a note of the burial enclosures at the east end of the church, the western end stone of which included a cross within a circle (Wakeman 1863, 220). His second account described the cross-carved slabs as being ‘very interesting, as, although extremely rude, they are all more or less decorated in the earliest style of Irish Christian art’ (Wakeman 1867, 368). Shortly after Wakeman’s second account, George Kinahan reflected the growing interest in and appreciation of the cross-carved slabs in commenting that the landing-place cross (Crossslab 22) ‘is very perfect’, and that the cross by Brian Boru’s well (Cross-slab 21) was ‘handsome’ (Kinahan 1869, 554). Twenty-seven years later the continuing interest in High Island cross-carved slabs was further reflected by Robert Macalister, whose observations included the fact

that the cross-carved lintel of the church doorway (Crossslab 16) was ‘unusually thin for such a position…it is…not the original lintel’ and that ‘three out of the five examined by us were cross-slabs rather than crosses’ (Macalister 1896, 198–9, 205). Macalister went on to describe (and in some cases sketch): the cruciform slab previously referred to that was reused as a lintel in the church doorway (Crossslab 16); the slab at the landing place (Cross-slab 22); the slab at Brian Boru’s well (Cross-slab 21); the slab sited at a point some 6.5m south of the church (Cross-slab 20), whence, Macalister notes, it had been removed by Kinahan; and a cruciform slab (Cross-slab 19) found in rubble and set upright by Kinahan, which excited Macalister’s interest because of the four holes that delineate the cross, only two, placed diagonally, had been completely drilled through the slab (Macalister 1896, 205–7). For archaeologists, one of the most important tasks in dealing with a site is to date artefacts and place them in chronological sequence. None of the antiquaries cited above ventured to date the slabs (perhaps wisely, in light of the limited knowledge available at the time), although Macalister would later date some of the comparable crosscarved slabs at Clonmacnoise (Macalister 1909, 96–110). His findings are considered later in this section. In the case of cross-carved slabs, dating has proved to be a stumblingblock to modern archaeologists and iconographers alike. In 1990 Michael Herity summed up the fruits of observations made during twenty visits to the island, spread over seven years (Herity 1990a, 71). These included detailed descriptions of the crosses and cross-slabs, which Herity considered to date generically to the end of the 7th century (Herity 1990a, 85–93, 97–9); the ‘ornamented tomb shrine’ at the east end of the church he dated specifically to AD 625–725 (Herity 1990a, 79). These very early dates do not, however, agree with evidence from the most recent excavations, illustrating the difficulties faced by archaeologists working with incomplete data. Jim Higgins was more circumspect in describing and dating cross types in his compendium of cross-slabs in Co. Galway, including some of the High Island slabs. For example, he separately dated each main face of the slab removed to Brian Boru’s well by Kinahan in the 1860s (Cross-slab 21). The first, simpler linear cross (Cross-slab

the cross-slabs 177

21B) he assigned to his Type I or III; the other face (Crossslab 21A), also a linear cross but of a developed form resembling expansional crosses found elsewhere on the Island, he assigned to his Type III (Higgins 1987, Pt. ii, 294). He had previously set out a scheme for dating types. Of these, he dated Type I from the 6th to the 7th century, with the caveat that ‘they seem to have been in use in association with other cross forms down to the 9th century’. He added that, in his opinion, ‘there is no evidence to show that the Co. Galway stones were in use between, say, the 10th and 12th centuries’ (Higgins 1987, Pt. i, 178). Type III crosses he considered to have ‘by far the widest date range of the Co. Galway monuments’, citing the 8th and 9th centuries as the main period, continuing into the 9th and 10th centuries; one could be dated to the 12th century (Higgins 1987, Pt. i, 178). Free-standing cruciform-shaped and decorated stones (Higgins’ Type IV) ‘because of the relatively close parallels which [they have] can be seen as part of a larger western coastal group… [and] are strongly suggestive of a date between the second half of the 7th century and 700 or even 750’. Faced with cruciform crosses with no decoration (such as the slab designated Cross-slab 19 in this section), Higgins commented that the dating ‘seems, on the basis of the few datable parallels available, to be very wide indeed. Some may be as early as the 8th century, though crosses of this type continued…into the Early Historic period…their plainness makes it virtually impossible to distinguish between early and late examples…plain crosses are known from all over Ireland…but none have been found in a closely datable context’ (Higgins 1987, Pt. i, 178–80). The most recent commentators, Jenny White Marshall and Grellan Rourke, concurred with Higgins’ comments about the possibly wide date range of some categories of cross. For example, short-armed cruciform crosses that may have been carved as early as the 8th century have modern parallels dating to the 19th and 20th centuries (White Marshall and Rourke 2000, 162; Higgins 1987, Pt. i, 71–3). This principle they accepted as applying to many categories of cross on High Island, though some they considered could be more closely dated. For example, following Lionard (Lionard 1961, 117), they suggested that outline crosses with hollows at the junctions of the limbs (referred to as armpitted crosses in the present work) were likely to have originated in Northumbria in the late 7th century and to have been imitated in Ireland by the 8th century (White Marshall and Rourke 2000, 163). In this respect Lionard followed the views of Françoise Henry (Henry 1933, 101–13, 166–77) and Robert Stevenson (Stevenson 1956, 85–6). The exchange of ideas between Ireland and Northumbria is not just a probability but a certainty, given the documented instances of Irish travellers and settlers in

Northumbria and vice versa, such as the numerous mentions in Bede’s Historia Ecclesiastica Gentis Anglorum (for examples of migration in both directions, see Colgrave and Mynors 1969, 221, 347). The influence of Irish sculptors or patrons is also evident in the designs of cross-carved slabs in Wales. Redknap and Lewis cite the Vita Cadoci (Life of Cadog), which speaks of a skilful Irish master-builder working in Glamorgan (Redknap and Lewis 2007, 121), and several High Island slabs are carved with designs comparable to those found in areas of Wales settled by Irish emigrants (see ‘Typology’, below). However, there is no direct information about how or when specific motifs were transferred from place to place. Some sculptural motifs used at the early Northumbrian monastery on Lindisfarne may, in fact, have been based on models carved on Iona, her mother-house, and be of early Irish origin. (It should be noted that few slabs from Iona have yet been definitively dated.) As Peter Harbison notes in the case of high crosses, a comparison between English and Irish crosses does not suggest that carvers, models or prototypes were of English origin (Harbison 1992, vol. 1, 345). To return to the example of the armpitted cross, wherever and whenever it originated, the motif continued to be used in the Insular world for some 400 years – and possibly longer – after the most cited and only securely dated early example: the outline cross that was carved on St Cuthbert’s wooden coffin in AD 698, an example that Richard Bailey considers to imply that free-standing crosses of this shape ‘were familiar objects in late 7th-century Northumbria’ (Bailey 1989, 238–43). Thus, it can be seen that both identifying the sources of cross types and dating them is an inexact art; to quote Kathleen Hughes and Ann Hamlin, ‘the dating of crosses, as of churches, is bedevilled by the lack of chronological fixed points’ (Hughes and Hamlin 1977, 91). The tools available to the epigraphist in cases where slabs are incised with text – i.e. dating by letter-form and the analysis of any grammatical conventions, spelling variants and scribal contractions used – are not available in the case of High Island. The same is true of the materials available to the historian, identification of distinctive names from manuscript sources, such as the Annals, being denied since none of the surviving slabs is inscribed with dedications. The remaining tool, typological analysis, has therefore been used to attempt to place the cross-carved slabs in roughly chronological order. This requires the comparison of each individual cross-form with comparanda not only at other Irish monasteries but at ecclesiastical sites in other Insular areas – the historic area of the kingdom of Northumbria, the Isle of Man and areas of present-day Scotland, Wales and south-west England settled by Irish emigrants from the 6th century onwards (see, for example, Edwards 2007, 5). Some of these comparanda have been dated by inscrip-

178 high island: excavation of an early medieval monastery

tion, but it must be stressed that this method cannot give exact dates, although it can suggest date spans during which particular designs were most likely to be used. It also gives some hint of the large number of contacts that must have taken place between ecclesiastic communities during the early medieval period, and the cosmopolitan nature of these contacts on High Island. Finally, iconography is used to supplement the typological analysis. The symbolic and theological meanings of the stones are explored according to the shapes of the crosses. Using this tool it is possible to relate some of the crosses to theological and liturgical ideas in early medieval Ireland that can also be roughly dated.

Typology

In this section the carved crosses have been arranged in three broad groups: linear, outline and expansional. Many of these have been incised on cruciform-shaped slabs, emphasising their theological content. These groups, together with their approximate dates, are set out in Table 7.1. This table also lists, for completeness, miscellaneous stone fragments that have been carved or incised, but not with a recognisable cross form. These are undatable except within the general time-span of the early medieval occupation of the island, that is, from the 7th century at the earliest to the 12th century at the latest.

Group 1: Linear crosses Linear crosses are those formed by single incised lines. At High Island they fall into three subgroups: linear crosses with various kinds of terminal (or none) but no circular details (Subgroup 1(a)); linear crosses whose central crossing has been encircled with a ring that does not encompass the limbs of the cross (Subgroup 1(b)); and linear crosses that are completely encircled (Subgroup 1(c)). Subgroup 1(a) consists of sixteen examples incised on fifteen separate slabs; 1(b) of five examples; and 1(c) of eleven or possibly twelve examples, carved on nine separate slabs. An exception, Cross-slab 5A, is incised with three rings at the crossing of a cruciform slab and seems likely to have borne a linear Greek cross inside the innermost ring, similar to that incised on the other broad face. Subgroup 1(a) All the crosses in this group are Greek (equal-armed) except Cross-slab 18, a pillar stone that was reused as a lintel in antiquity, which when viewed in its original upright mode has upper and lower limbs that are longer than the lateral ones. Despite their apparent simplicity, there is some variety in the details of these linear crosses. Four Subgroup 1(a) crosses (Cross-slabs 30, 33, 36 and 40) have no terminals, while two (Cross-slabs 18 and 22C) have crosslets on each limb. One cross, no. 31, has short bar terminals; eight crosses (Cross-slabs 2B, 9B, 9C, 20B, 21B, 23, 27 and 64B) have forked or curving terminals.

The simplest linear crosses, of which Cross-slabs 30, 33, 36 and 40 are representative, are commonly found carved on slabs and pillars at Insular sites in a multiplicity of contexts – architectural features, focal-points, boundary and burial markers and bullauns. For example, two pillars incised with linear Greek crosses set among a group of four mark St Patrick’s Chair, Marown, on the Isle of Man. Although the site may have originally been used for pagan burials, no bodies were found associated with the crosscarved pillars, suggesting either that they were exorcising and Christianising a formerly pagan sacred site, or that they marked a Christian focal-point or boundary. Patrick Kermode considered these pillars to date to between the 6th and 8th centuries (Kermode 1907, 102–3, plate VI, 5 and 6). However, the same style of cross is also found on a slab marking a Viking burial at Killoran Bay, Colonsay (RCAHMS 1984, 150, fig. 298), in the Inner Hebrides of Scotland, which effectively dates it to the 9th century. Lionard dated Insular examples of simple linear Greek crosses to the 6th to early 7th centuries, as did the Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS) in respect of some Ionan slabs, although others (Iona 1–7) were considered to be considerably later (RCAHMS 1982, 16). In south-west Ireland, some 5th- or early 6th-century ogam stones that may commemorate Christians are decorated with simple linear crosses, suggesting that this form was present in Ireland from the conversion period onwards (Swift 1997, 99). Comparison with Welsh incised linear crosses supports a potentially wide date range for this type. In south-east Wales and along the English border, a simple linear Greek cross on side A of the Trecastle or Llywel Stone, Breconshire, has been dated to the early 6th century; a similar incised cross at Llanmadog, Glamorgan, is suggested to date to the 7th to 9th centuries; while a third such cross, at Pyle, Glamorgan, has been tentatively dated to the 11th to 12th centuries (Redknap and Lewis 2007, 235–40, fig. B42a; 359–60, fig. G56; 491–2, fig. G115). At Llanfihangel-y-Creuddyn, Cardigan, a rectangular slab with a roughly rounded top that bears some resemblance to the incomplete Cross-slab 36 has been dated to the 7th to 9th century (Edwards 2007, 163–4, fig. CD18). At St Nicholas in Tremarchog, Pembrokeshire, a Roman-letterinscribed stone with a simple linear Greek cross can be dated to the 6th century from epigraphic evidence, while a slab at Penally, Pembrokeshire, incised only with a large linear Greek cross, may date to the 7th to 9th centuries (Edwards 2007, 491–2, figs P133.1–2, fig. P85; 422). A similar incised slab at Newchurch, Carmarthenshire, has also been dated to the 7th to 9th centuries (Edwards 2007, 274–5; fig. CM38). Northumbrian slabs marked with simple linear crosses appear to fall mainly into the 7th- to 9th-century bracket.

X X

19B

Cruciform

19A

Group 3 Expansional

Subgroup (c) Armpitted

Subgroup (b) Ribbon

Subgroup (a) Outline

Group 2 Outline crosses

X X

Catalogue Group 1 Linear crosses number and Subgroup (a) Subgroup (b) Subgroup (c) face Linear Ringed centre In circle

Table 7.1 Cross-referencing of slabs and typological groups of cross designs (socket-stones, jambs and fragments classified as ‘Miscellanea’).

Miscellanea

7th to 12th century

8th to 11th century

7th to 12th century

9th to 11h century

9th to 11th century

7th to 12th century

9th to 11th century

7th to 12th century

9th to 11th century

7th to 12th century

10th to 11th century

9th to 11th century

7th to 12th century

Date

the cross-slabs 179

Subgroup (c) Armpitted

39B

Subgroup (b) Ribbon

Group 3 Expansional

Subgroup (a) Outline

Group 2 Outline crosses

39A

28B

28A

22C

22B

22A

21B

21A

20B

20A

Catalogue Group 1 Linear crosses number and Subgroup (a) Subgroup (b) Subgroup (c) face Linear Ringed centre In circle

Cruciform

Continued Table 7.1 Cross-referencing of slabs and typological groups of cross designs (socket-stones, jambs and fragments classified as ‘Miscellanea’).

Miscellanea

7th to 12th century

9th to 11th century

10th to 11th century

7th to 12th century

9th to 11th century

7th to 12th century

9th to 11th century

7th to 12th century

9th to 10th century

7th to 12th century

9th to 11th century

10th to 11th century

7th to 12th century

9th to 11th century

7th to 12th century

9th to 11th century

Date

180 high island: excavation of an early medieval monastery

64B

64A

58 X

Miscellanea

Subgroup (c) Armpitted

Cruciform

Subgroup (b) Ribbon

Group 3 Expansional

Subgroup (a) Outline

Group 2 Outline crosses

Catalogue Group 1 Linear crosses number and Subgroup (a) Subgroup (b) Subgroup (c) face Linear Ringed centre In circle

Continued Table 7.1 Cross-referencing of slabs and typological groups of cross designs (socket-stones, jambs and fragments classified as ‘Miscellanea’).

7th to 12th century

9th to 11th century

7th to 8th century

7th to 12th century

9th to 11th century

7th to 12th century

10th to 11th century

Date

the cross-slabs 181

182 high island: excavation of an early medieval monastery

A slab marked with a simple linear cross extending to the edges of the slab, from Coquet Island off the coast of present-day Northumberland, has been dated from historical evidence to the last quarter of the 7th to the first quarter of the 8th century (Cramp 1984, 170, pl. 164, 864–6). At Addingham, Cumbria, a simple linear cross on a slab that may originally have been an upright gravemarker has been dated to the 6th to 8th centuries (Bailey and Cramp 1988, 47–8, ill. 15–16). A similar cross incised on an irregular slab at Whithorn, Wigtownshire, is datable by stratification in a late 7th- to early 8th-century context (Craig 1997, 433). At Hoddom in Dumfriesshire (an early monastic site whose medieval carved slabs show clear influence of Northumbrian exemplars), a small subcircular stone marked with a cross with limbs of unequal length, but whose crossing is at the centre of the stone, has been dated to the 9th century (Craig 2006, 131–2, plate 5.11). These comparanda suggest that High Island slabs carved with plain linear Greek crosses can only be dated to the 7th to 12th centuries. Linear crosses with crosslets on each limb of the type found on Cross-slabs 18 and 22C are also found elsewhere in Ireland: for example, in outlined form on a slab at Inishmurray, Co. Sligo (O’Sullivan and Ó Carragáin 2008, 81, fig. 21, 001.12), that is tentatively dated to the 9th century or later; and on an inscribed but undated gravemarker at Ballineanig, Co. Cork (O’Sullivan and Sheehan 1996, 366, fig. 213). Cross-slab 18, which was reused as the innermost lintel along the entrance passage to one of the beehive huts in antiquity and is over 2m long in its current state, was almost certainly originally set upright as a pillar stone, where it may have functioned as a termon. A similar fate was meted out to Cross-slab 16, a pillar with notched ‘arms’ and a linear cross-in-circle that may originally have been a grave-marker. Linear crosses with crosslets also occur in south-east and south-west Wales. Of the Welsh examples, Llangyndeyrn 1 and Llandysul 2, Carmarthenshire, and Nevern 5, Pembrokeshire, may date to the 7th to 9th centuries; Capel Colman, Pembrokeshire (face C), may date to the 7th to 8th centuries; while Llandewibrefi 6, Cardiganshire, has been dated epigraphically to the 9th century; and Llanmadog 3, Glamorgan, may also date to the 9th century (Edwards 2007, 157–9, fig. CD13; 161–2, fig. CD15; 257–8, fig. CM29; 301–2, fig. P8.2–3; 401–2, fig. P74; Redknap and Lewis 2007, 361, figs G57a–b). Another rectangular slab incised with a cross with crosslets but of unknown Welsh provenance was dated by Nash-Williams to the 7th to 9th centuries (Nash-Williams 1950, 222, plate XXV). Llandewibrefi 6 is particularly interesting since, like Cross-slab 18, it is a pillar stone. At 2.49m high and 26.5cm wide, with a linear cross at the apex of the pillar and an inscription running vertically down the shaft, it

gives an idea of the possible original appearance of Crossslab 18. The comparanda suggest that linear crosses with crosslets at High Island (18 and 22C) may also date to the 7th to 9th centuries. Widely forked or curved terminals (Cross-slabs 2B, 9B and 9C, 20B, 21B, 23, 27 and 64B), also found on ringed and cross-in-circle designs, are the most common variants of the linear cross and are a distinctive feature of High Island linear crosses, relating to the form of the expansional crosses in Group 3 that may have been a development or refinement of them. This is particularly suggested by the cross incised on Cross-slab 64B, where the lowest limb terminates in a D shape identical to the four incised on the terminals of the cross on the other face of the slab. The fact that this is a completed design, not an unfinished one, seems proved by an almost identical design carved on a slab at Ballylosky, Co. Donegal (Lacy 1983, 243, fig. 123). In contrast, only one slab in the subgroup is incised with one of the most ubiquitous types of linear cross found on both manuscripts and monuments in the early medieval Christian world: a small Greek cross whose limbs terminate in short bars or cusps (Cross 31). Lionard notes that this type of cross originated in Rome towards the middle of the 5th century and was being used in Insular manuscripts by the 7th century; he also cites an example on a Welsh monument dated by inscription to AD 652 (Lionard 1961, 101). The type was extremely long-lived, Lionard pointing to an example at Clonmacnoise that seemed likely to date to the early 12th century (Lionard 1961, 102, fig. 24, 13). In these circumstances it is impossible to date Cross 31 other than very generally – in other words, from the 7th to the 12th centuries. Subgroup 1(b) Linear crosses with a ring around the crossing are somewhat less common than those without such a ring. The subgroup consists of five slabs: Cross-slabs 10, 12, 18, 21A and 22B. Typically at High Island, this group has short limbs ending in widely forked or cusped terminals outside an encircled crossing. One (Cross-slab 10) has an antler-like, bifurcated upper terminal; three others (Crossslabs 12, 13 and 22B) have double, slightly ovoid rings. The last of these also has crosslets towards the end of the limbs and resembles a cross incised on a slab at Kilcashel, Co. Donegal, generically dated by Lionard to the 9th century (Lionard 1961, 118, fig. 13, 3). Cross-slab 12 is particularly interesting in that it also has four incised circles in the quadrants, formed by the edges of the slab and the centrally placed cross, the upper two of which contain linear Greek crosses and the lower two of which were originally similarly incised (see description of Cross-slab 12). This motif may bear some relationship to an altar-slab, a stone slab marked with five crosses inserted into a cavity to effect consecration

the cross-slabs 183

(Lionard 1961, 136). Possible examples of altar-slabs are known from Downpatrick, Co. Down, Clonmacnoise, Co. Offaly, Gallen Priory, Co. Offaly, Kilpecan, Co. Tipperary, and Ardmore, Co. Waterford (Lionard 1961, 137). Another established format utilising the five linear cross motif is the portable altar, which was designed to provide a consecrated stand on which priests could place the sacred elements when celebrating Mass, although no definite examples have been identified in Ireland. Priests might also be buried with their personal altar, as was the case, for example, with St Cuthbert (Radford 1956, 326–35, fig. 1), and may have been the case with the possibly 7th-century slab, roughly incised with five crosses disposed at corners and centre, found in the grave of an elderly male at Ardwall Isle in Kirkcudbright, Scotland (Thomas 1967, 161–2). Given that the lower circles incised on the High Island slab contained linear Greek crosses that have since flaked away, it is possible that it may originally have marked the burial of a priest. However, since slabs marked with five crosses continued to be used both for burials and on altarstones up to the 12th century, this feature in itself is not enough to date Cross-slab 12. The ringed motif, however, seems likely to represent the same ideas as outline and free-standing ringed crosses, that is, of the theology of the crucified Christ as Saviour and God for eternity, which first appeared in Insular areas in the 9th century and continued until the 11th. Subgroup 1(c) Linear crosses-in-circle are usually of Greek form, with a variety of terminals: forked, bar and curved. This motif is particularly favoured at High Island to fill central roundels on cruciform slabs. Not all are simple linear crosses, however: in one case, Cross-slab 5B, a linear cross with barred terminals is superimposed on a cross-of-arcs; in another, Cross-slab 7, a geometric motif emphasises three horizontal lines crossing three vertical lines at right-angles, the lines curving inwards above and below the crossing to follow the shape of the roundel. The simple Greek linear cross-in-circle motif is ubiquitous in sculpture-rich areas of Ireland, for example: at Gallarus, Innisfallen, Kilfountan, and on a lost stone at Kilmalkedar, all in Co. Kerry; at Castleconnell, Co. Limerick, and Cloghinch, Co. Tipperary (Okasha and Forsyth 2001, 147–50; 157–60; 161–5; 169–71; 185–8; 191–3); on a lintel over a door at Fore, Westmeath (Henry 1947, pl. 23); at five sites in Galway (Higgins 1987, Pt. i, 59–63; Pt. ii, fig. 71–77); at Kilmore Moy, Co. Mayo (Henry 1937, 267–8, fig. 1), and on cursing stones at Inishmurray, Co. Sligo (O’Sullivan and Ó Carragáin 2008, 81, fig. 22). Most of these have not been authoritatively dated, although the Cloghinch example has been argued to date to the 7th century on contextual grounds (Kelly 1988, 98), and a slab with a linear cross-in-circle on a long shaft at Dunmore,

Co. Galway, has been tentatively dated to the 7th to 10th centuries (Higgins 1987, Pt. ii, 361, fig. 79 and pl. 28A). Higgins noted that this type of cross was in common use in Ireland by the 7th century, but still being carved in the 12th century. However, he considered most to date to the 7th to 10th centuries (Higgins 1987, Pt. i, 172). In the Kingdom of Dyfed, one of the areas of Wales colonised by Irish settlers (Charles-Edwards 2000, 158–72; Edwards 2007, 5), the cross-in-circle incised slab is common. Among 21 examples, the memorial to Voteporigis at Castell Dwyran, Carmarthen, has been dated to the late 5th or early 6th century, demonstrating the antiquity of this type in Insular Christian art (Ó Cróinín 1995, 18–19; Edwards 2007, 202–6, fig. CM3, 1–3). At St Non’s Chapel in St Davids, Pembrokeshire, an unshaped pillar incised with a linear cross-in-circle with a short stem has been dated by Edwards on the grounds that its simplicity lends itself to an early date, that is, to the 7th or 8th century (Edwards 2007, 449–50, fig. P100). At Mathry, also in Pembrokeshire, another unshaped pillar carved with a similar cross has been dated by Edwards to the 7th to 9th centuries (Edwards 2007, 384, fig. P62). A cross-incircle that originally had an extended shaft with bifurcated terminal is incised on a rectangular-section pillar at Llanfihangel Ystrad, Cardiganshire. This pillar has been dated by inscription to the late 8th or early 9th century (Edwards 2007, 166–9, fig. CD20, 1–3). At Llanychaer, Pembrokeshire, simple linear crosses-in-circle are incorporated into designs on three faces of a shaped pillar dating to the late 8th or 9th century (Edwards 2007, 362–6, fig. P49, 2–4). At Fishguard South, Pembrokeshire, Face A of a roughly quadrangular pillar was re-carved in the 9th to 11th centuries to include a linear equal-armed ringed cross (Edwards 2007, 322–3, fig. P16), while at St Dogwells, Pembrokeshire, a linear Greek cross in a rectangular frame, similar to Cross-slab 40, has been dated to the 7th to 9th centuries (Edwards 2007, 474–5, fig. P120). Examples of the simple cross-in-circle have also survived at Iona (RCAHMS 1982, 180–81, figs. 5–6), but remain undated. A more complex design for a ringed cross is to divide it into segments by eight incised or moulded lines radiating from a central boss to the circumference of the circle, so forming two interlocking Greek crosses with straight limbs. This design is found on Cross-slab 37, on an undated slab at Faha, Co. Kerry (Cuppage 1986, fig. 169), and on two 11th-century slabs at Merthyr Mawr, Glamorgan (Redknap and Lewis 2007, 474–7, figs. G101a–b and G102a–b). In outline, rather than linear cross-in circle design form, the same motif is found on the Ilquici and Ilci cross-carved slabs at Margam, Glamorgan (Redknap and Lewis 2007, 427–33, figs G84a–h; 434–7, figs G85a– g), the former of which has been dated by inscription to the 10th to 11th centuries, the latter to the late 10th to

184 high island: excavation of an early medieval monastery

11th centuries. This suggests that the motif emerged in the 10th century and continued into the 11th. Cross-slab 37 may have been carved at the same period. In summary, the wide range of dates suggested for simple linear crosses-in-circle in the Insular world and the lack of contextual evidence make it impossible to date most Cross-slab Subgroup 1(c) slabs other than very generally, that is, to the 7th to 12th centuries. The exception is Cross-slab 37, which probably dates to the 10th to 11th centuries.

Group 2: Outline crosses Outline crosses are those delimited by incised or carved lines, either in relief or false relief, to give the impression of being raised above the surface of the slab. This group is divided into three: plain outline, ribbon and armpitted. Subgroup 2(a) consists of only one example, Crossslab 60, an upright slab recorded and sketched by Wakeman in 1839. It was carved, apparently in relief, with a plain Latin cross. It is not clear from this sketch whether the cross shaft terminated in a straight rectangular edge, similar to those terminating the other limbs, and was partially covered by heaped earth, or whether the cross itself was depicted with a mound-shaped base. The latter type is most likely to represent the cross upon which Christ was crucified at Golgotha, popularised by early Insular accounts of pilgrimages to Jerusalem and fictitious representations in mosaics, such as that at S. Pudenziana in Rome (which dates to the 5th century), representations of which with stepped or expanded bases became popular in the 7th to 9th centuries (see Maddern 2013, 214, 242–3). However, given Wakeman’s clear intention to depict the neglected state of the site, it seems more likely that the mound represents actual rather than carved earth, and that this was a very plain Latin cross of a type also seen on Face B of a slab at Margam in Glamorgan, dated by the inscription on Face A to the late 9th century (Redknap and Lewis 2007, 408–11, fig. G78b), and on three slabs at Llanwda in Pembrokeshire that date to the 9th to 11th centuries (Edwards 2007, 350–51, 354, figs. P37–38, P51). Thus, Cross-slab 60 may also have dated from the late 9th to the 11th centuries. Subgroup 2(b) Five crosses fall within this category: Cross-slabs 4A, 4B, 22A, 38 and 41, each of different design. The two faces of Cross-slab 4 each bear a ribbon cross with widely splayed terminals, that on Face A being of Latin form, the base divided into two horizontal bars, each ending in a trilobate terminal. Underneath this base a rectangular space appears to have been reserved. This unusual format suggests that the motif may have been intended to represent a cross on a stand, in other words, to be a skeuomorph of a wood or metal

model. A carved cross of otherwise very different form at Inishmurray also appears to be a skeuomorphic cross on a stand (O’Sullivan and Ó Carragáin 2008, 123–4, fig. 001.49A). Face B shows a Greek cross with four small pellets at the crossing; there are also bosses with hollowed centres in the quadrants formed by the edges of the slab and the limbs of the cross. In both crosses the shape of the terminals and general style of carving have parallels in south-east Wales. A slab in SS Peter and Iltyd’s Church at Llanhamlach, Breconshire, includes a Latin ribbon cross with terminals similar to those carved on High Island 4B. The Welsh cross has been dated by the accompanying inscription to the 10th/11th century (Redknap and Lewis 2007, 210–13, figs B32a–c). A similar motif to that carved on Cross-slab 4B appears on one face of a cross-carved slab at Neath Abbey in Wales, dated by inscription to the 9th to 10th centuries. Redknap and Lewis point out that this motif also appears on a possibly 10th-century hogback from Castledermot and on a monument from Knockbarrow, Co. Offaly (Redknap and Lewis 2007, 274–7, fig. G9a, citing Corlett 1999, 11; O’Brien 1994, 16; O’Brien and Sweetman 1997, 103). A similar double-band cross with splayed forked terminals with rounded edges is found on a pillar at Pen-Arthur Farm, St Davids, Pembrokeshire. It has been dated from the 9th to the 11th centuries (Edwards 2007, 457, fig. P106). These exemplars suggest that the motif originated in Ireland, rather than Wales, and that it was in use between the 9th and 11th centuries. Cross-slab 4B is unusual in that it has four small flattened pellets at the crossing. This cross, like that on the other face of this slab, may be a skeuomorph; the pellets may represent pegs or nails securing the horizontal to the vertical limbs. Robert Stevenson has noted, in a different context, that pellets are characteristic of the 10th- to 11th-century Anglo-Scandinavian sculpture of Northumbria (including Galloway, which was annexed by Northumbria for part of this period), in addition to being ‘scattered among the ornament in earlier manuscripts and metalwork’ (Stevenson 1956, 94–5). Although not conclusive, taken together with the form of the cross this evidence tends to support a 10th- to 11th-century date for Cross-slabs 4A and 4B. Another interesting feature of this slab is that it has a carved tenon, indicating that it may at some point have been inserted into a leacht or dry-stone altar; at all events, it was clearly intended to stand upright in a position where both sides could be seen. It is also interesting in that it is carved with a number of Christological symbols, which are discussed in the Iconography section below. Similarly, Cross-slab 22A, which has twisted ribbon decoration forming the limbs, the uppermost of which contains a triquetra, may date to the 10th to 11th centuries.

the cross-slabs 185

Higgins’ drawing of this slab, based on photographs and drawings by Macalister (1896) and Westropp (1905) and others, shows the ribbons forming a triquetra in each lateral limb, and a detached triquetra in the upper and lower limbs (Higgins 1987, Pt. ii, 346, fig. 67), but these details were not visible by the time of the excavations of the 1990s. Higgins did not assign a type – and thus a date – to this slab, but because triquetrae at the ends of cross-limbs most frequently occur in Anglo-Scandinavian crosses, it seems likely to date to the 10th to 11th centuries (Stevenson 1956, 94). A Northumbrian example of a cross-arm ending in triquetrae may be seen at Brigham in Cumbria, tentatively dated by Richard Bailey and Rosemary Cramp to the 10th or 11th century (Bailey and Cramp 1988, 77–8, ill. 157). Triquetrae in cross terminals are also comparatively common on Welsh cross-carved slabs, and are of the same date as their Irish and Northumbrian counterparts. Triquetrae carved on slabs at Llandyfaelog Fach and Llanfrynach, Breconshire (Redknap and Lewis 2007, 185–90, figs B16b–e; 200–202, figs B26a–b), have been dated to the late 10th century and 10th to 11th centuries, respectively; similar motifs have been found carved on slabs and freestanding crosses at six sites in Glamorgan, at Laugharne in Carmarthenshire, and at Nevern in Pembrokeshire (Redknap and Lewis 2007, 288–92, figs G16a–d, f; 310–12, figs G31a–d; 340–42, figs G44a, d; 369–73, figs G63a–b; 411– 20, figs. G79b–c; 466–72, figs G99c–f, h; Edwards 2007, 221–2, fig, CM10; 395–96, fig. P72). All these Welsh slabs and crosses have been dated to the 10th to 11th centuries. A similarly shaped cross to Cross-slab 22 at Kill, on the mainland a short distance from High Island, was assigned by Higgins to his Type IV, in other words, to between the second half of the 7th century and AD 750 (Higgins 1987, Pt. ii, 345, fig. 66); however, this example lacks both the ribbon interlace and the distinctive triquetrae. On Iona, a similarly shaped cross (without triquetrae or interlace), carved on a slab formerly at the Nunnery but now lost, has not been dated (RCAHMS 1982, 183–4, fig. 29). Two further slabs resembling the Kill cross at the early monastic site on Caher Island, Co. Mayo, were dated by Françoise Henry to the 7th century on the grounds that it was the period of greatest development of the monastery (Henry 1947, 38). In the case of Cross-slab 22A, however, the presence of both ribbon interlace and triquetrae argue for a much later date. Cross-slab 38 is a cruciform stone of which only a fragment of the head survives. The ribbon moulding forms four ‘petals’, each of which contains a small boss. A similar design is found on an undated fragment of a cross-slab, now missing, from Inishkeel, Co. Donegal (Lacy 1983, 271–2, pl. 45). The use of ribbon moulding on Cross-slab 38 suggests a date of the 10th to 11th century, but there is insufficient detail to date the slab definitively. Similarly,

Cross-slab 41, a cruciform slab incised with a double-ribbon Latin cross with slightly splayed terminals, may date to the 10th to 11th centuries. Subgroup 2(c) Only three High Island cross-carved slabs are armpitted: Cross-slabs 17, 29, and 63. The first of these appears to be of unframed Latin form, with straight-sided limbs ending in straight terminals. It has a shaft only slightly longer than the surviving complete arm, suggesting that it may originally have been planned as a Greek cross, which in execution became a Latin cross. Both Greek and Latin armpitted crosses are common in Insular stone sculpture from the 7th century onwards, but the shapes of arms and armpits vary considerably. Irish, Scottish and early Northumbrian crosses influenced by Iona have straight arms and terminals, with circular or ovoid armpits, whereas later Scottish and Northumbrian examples have concave arms and terminals and widely splayed armpits. Irish, Scottish and Manx armpitted crosses are frequently also ringed. The stepped cross incised on the wooden inner lid of St Cuthbert’s coffin, which has been shown by Richard Bailey to be contemporaneous with the figural carving on the coffin and to date to around the time of St Cuthbert’s death, AD 698, is of the early Northumbrian armpitted type (Bailey 1989, 231–43, fig. 14). Early armpitted cross-heads at Whitby, North Yorkshire, which are generally plain and date to the late 7th to 9th centuries, have splayed rather than armpitted limbs (Lang 2001, 237–50, ill. 897–1011). At Hexham, Northumberland, an armpitted cross-head decorated with a rosette has been dated by Rosemary Cramp to between AD 725 and AD 800 (Cramp 1984, Pt. i, 179–80, Pt. ii, plates 910–913). However, this cross-head has shorter limbs and wider armpits than Cross-slab 17, so that the form is not directly comparable. Northumbrian armpitted cross-heads of the late type were still being carved in the 11th century and are identifiable by the style of figural carving on each face (see, for example, Cramp 1984, Pt. i, 68–73, Pt. ii, plates 43–48). The very simple outline cross incised on Cross-slab 17 therefore belongs to a different tradition and seems indicative of either an ascetic monastic philosophy, where all ornament was eschewed in order to concentrate and meditate solely on the Cross itself, and/or an early date. Since recent excavations have confirmed the presence of two structures earlier than the church on High Island, it is possible that Cross-slab 17 was originally associated with one of them. Outside Northumbria, eighteen examples of comparatively simple outline Irish armpitted Latin crosses on recumbent slabs are cited by Lionard, who notes that they are ‘almost exclusively confined to Inis Cealtra [Co. Clare], (eighteen examples) and Nendrum [Co. Down], (seven examples)’ (Lionard 1961, 114–17, fig. 11); he found

186 high island: excavation of an early medieval monastery

that the earliest dated example of a framed Greek version of the armpitted cross on a recumbent slab was dedicated to Abbot Muirgal of Clonmacnoise, Co. Offaly, who died in AD 789 (Lionard 1961, 117). However, armpitted Latin crosses carved on upright and recumbent slabs have a wider distribution in the Insular world than Lionard’s work suggests. They have also been found in Iona, in the monastic cemetery Reilig Odhráin (RCAHMS 1982, 183, 187, figs 25, 53b; Fisher 2001, 128–35, figs. 34A, 70, 94), and on the Isle of Inchmarnock (Bute), on Isle Maree at Dòid Mhàiri (Port Ellen, Islay), A’Chill (Muck), Kilkerran and Kilmartin (Fisher 2001, 128–35, figs 34A, 70, 94:77, fig.6;11, fig.92;117;149). Only the Dòid Mhàiri slab (which is also ringed, and is carved in the distinctive Scandinavian Ringerike style) has been dated, in this case to the second half of the 11th century. The other slabs are carved in a variety of Insular styles, suggesting a date range of the 7th to early 10th centuries. In Wales, armpitted cross-carved slabs are found at coastal sites in Cardiganshire and Pembrokeshire, suggesting Irish rather than Northumbrian influence. An inscribed slab bearing an armpitted Latin cross at Llanarth, Cardiganshire, has been dated to the 9th or early 10th century by Nancy Edwards. In this she cites Victor Nash-Williams, who noted that such a date would be compatible with the epigraphic evidence; furthermore, ‘there are no indications of Viking-Age stylistic influence on the monument’ (Edwards 2007, 176–8, fig. CD25, citing Nash-Williams 1950, no. 110). Two further examples of outline Latin crosses with straight limbs and circular armpits are found at St Ismael’s Church, St Ismael’s, and at St Brynach’s Church, Pontfaen, both in Pembrokeshire. The former has been dated by Edwards to the 9th or early 10th century, the latter to the 9th to 11th centuries (Edwards 2007, 485–6, fig. P129; 424, fig. P87). On the Isle of Man, armpitted and ringed crosses are the most common type, both on slabs and as free-standing crosses. The motif is found on slabs at most ecclesiastical sites, particularly Maughold (Kermode 1907, plates XI–XIII, XV, XVIII–XXIV, XXVI–XXXIV, XXXVI, XXXVIII, XL–XLI, XLIII–V, XLVII, L–LIX, LXI–LXII). Just under half of these ringed and armpitted crosses are filled with interlace, geometric ornament and figural sculpture of Scandinavian type, and date to the 10th and 11th centuries. Only one slab, at Andreas, is carved with a Latin cross with armpits, but no ring. However, vestiges of ring-chain decoration on the shaft and a runic inscription suggest it may date to the 11th century (Kermode 1907, 78–9, 164–5, plate XXXIX). The wide range of dates suggested for comparable slabs makes it extremely difficult to arrive at a probable date for Cross-slab 17, which could have been carved at any point between the 8th and 11th centuries, although, given the

shape of the armpits, the earlier part of this range seems most likely. Cross-slab 29 is slightly easier to date since, in addition to being armpitted, it has a ringed head, a feature that was most common in the 8th to 10th centuries in Ireland. Lionard noted that the ringed cross was so common in Ireland that it became known, misleadingly, as the ‘Celtic cross’. He distinguished four main types and six variants, of which Cross-slab 29 most nearly approximates to ‘Ringed cross hollowed’. This was the most popular version of the ringed cross to be incised on recumbent slabs, of which Lionard noted 45 examples. Of these, he dated two examples at Clonmacnoise, Co. Offaly, to the first quarter of the 8th century, and another to the first quarter of the 9th century, from their dedication inscriptions. Another inscribed slab at Iona was dated by Lionard to the end of the 9th century (Lionard 1961, 117–27). More recent scholarship has cast doubt on Lionard’s dating because it was largely based on assumptions that names which appear in the annals were in fact those incised on recumbent slabs at Clonmacnoise – as above – and elsewhere (Swift 1997, 248; Ó Floinn 1995, 252–3). However, comparable research by RCAHMS on the 32 Ionan slabs carved with ring-headed and armpitted crosses suggests that Lionard was not far out. Three slabs were dated by RCAHMS, using epigraphic rather than annalistic techniques, to the 8th and 9th centuries (RCAHMS 1982, 185–91, figs. 34–65, 70). In Wales, two armpitted ring-headed crosses comparable to Cross-slab 29 are found in Pembrokeshire. At St David’s Cathedral, a fragmentary cross-carved stone of similar format to Cross-slab 29, but carved in deep relief, has been dated by Edwards to the 10th to 11th centuries; a more detailed and crisply carved version found on a slab at St Edrin’s Church, Llanedrin, has been dated to the 9th or early 10th century (Edwards 2007, 438–9, fig. P96; 479–80, fig. P124). Taken together, these comparanda suggest that Insular armpitted and ring-headed cross-carved slabs first appeared in the 8th century and reached a peak of popularity in the 9th and 10th centuries, before declining to a small number of 11th-century examples and Scandinavian adaptations; high crosses in this format are of similar dates. Cross-slab 29 therefore seems most likely to date to the 9th to 10th centuries. The last of the three slabs in this subgroup, Crossslab 63, resembles Cross-slab 22 in outline, including the slightly projecting lateral arms that render both slabs cruciform, but differs from it in that it lacks interlaced bands and triquetrae and has pronounced circular armpits. It more nearly resembles the slab at Kill, Co. Galway, which as previously mentioned is situated on the mainland a short distance from High Island, and slabs at Caher Island, Co. Mayo. Higgins thought that the Kill slab dated to between

the cross-slabs 187

the second half of the 7th century and AD 750 (Higgins 1987, Pt. ii, 345, fig. 66); Henry considered that the Caher Island slabs dated to the 7th century (Henry 1947, 38). Since Cross-slab 63 does not include later features, such as interlace or triquetrae, and is of the same geological type as that used to form Caher Island cross-slabs (a geological type not found on High Island), it may well have originated on Caher Island and be of early date, that is, the 7th to 8th centuries.

Group 3: Expansional crosses Expansional crosses rival linear crosses with forked terminals as the most popular motifs carved on slabs at High Island, there being fourteen examples carved on thirteen stones: Cross-slabs 3, 6, 8 9A, 11, 14, 15, 20A, 32, 34, 39A, 39B, 61 and 64A. It could be said that crosses with expanded D-shaped terminals form the nearest to a ‘house style’ for High Island slabs, being ‘the most distinctive and elaborate cross designs, that are not found on any other Atlantic Island site’ (White Marshall and Rourke 2000, 168). Although Lionard considered that the ultimate inspiration for the expansional cross was the Byzantine cross with splayed terminals (a design that was later taken up in Merovingian metalwork), he was clear that all the recumbent slabs he studied at Clonmacnoise carved with crosses that included interlace and fret in their expansions dated, with only one exception, to the 10th century. White Marshall and Rourke thought that recumbent stones with internally decorated expansional cross designs appeared in the mid-9th century in Ireland and began to replace what had formerly been the most popular motif of that century: the ring-headed cross. Expansional crosses appear to have been even more popular in the 10th to 11th centuries, when the ringed cross motif was abandoned (White Marshall and Rourke 2000, 166–7). The nearest comparanda to the High Island expansional crosses are found at Inchagoill and Clonfert, both in Co. Galway. The former most closely resembles Cross-slab 8, although it lacks the quadrant bosses and peltae which feature prominently on that slab. Higgins assigned both Cross-slab 8 and the Inchagoill slab to his Type III – that is, the 8th and 9th centuries – as the main period in which these slabs were likely to have been carved, continuing into the 9th and 10th centuries (Higgins 1987, Pt. ii, 342–3, figs. 63–4). Expansional crosses are also found at Inis Cealtra, one with D-shaped upper terminals being grouped under Macalister’s ‘tenth-century type’ (Macalister 1916, 152–4). The largest numbers of expansional crosses on recumbent slabs, however, are found at Clonmacnoise, Co. Offaly, which Macalister grouped and dated in the early years of the 20th century. Recognising the dangers inherent in attempting to identify slabs from single names, he would only date slabs with inscriptions that included epithets dis-

tinctive enough to positively identify individuals recorded in the Annals of Clonmacnoise. Eight such slabs are incised with expansional crosses, three of which he considered to date to the late 9th century and five to the 10th century (Macalister 1909, 97–100). This research was partially confirmed by Raghnall Ó Floinn, whose study of Clonmacnoise slabs resulted in the identification of three major types, which he labelled A to C, Type B of which covered expansional crosses with semi-circular terminals, without central rings. He found that three Type B slabs could be dated independently on the basis of their inscriptions, citing three of the eight inscribed slabs noted by Macalister in 1909 dating to the late 9th and mid-10th century (Ó Floinn 1995, 254). Therefore, although the Clonmacnoise slabs are much more refined than those at High Island, and the lack of inscriptions on the High Island slabs makes it impossible to link them with historic personages as at Clonmacnoise, it seems reasonable to assume that the expansional cross motif was adopted there at roughly the same time as at Clonmacnoise and other ecclesiastical sites in Ireland, that is, the late 9th to early 11th centuries.

Cruciform

Thirty-eight carved cross motifs at High Island have been incised on 29 slabs or pillars that are themselves roughly cruciform. The pillars demonstrate a tendency to apply minimum effort in achieving the requisite format, with notches defining (or suggesting) the arms (nos. 16, 43–45, 50). However, most cruciform crosses (Cross-slabs 1–2, 5, 7, 20, 22, 27–29, 33, 38, 41, 48–49, 54–55, 59, 62–63) have four extended, although short, limbs, carved out of slabs using the natural shape and grain of the stone wherever possible, with the result that some are very irregular. Only twelve cruciform slabs (nos. 43–50, 54–55, 59, 62) are not also incised with a cross motif on at least one face of the slab. Otherwise undecorated cruciform slabs may have been boundary or focal-point markers, turned out in quantity according to need. On the other hand, uninscribed grave-markers or votives were not uncommon at early ecclesiastic sites. For example, of the surviving corpus of two free-standing crosses and 72 small upright noncruciform cross-slabs at St Berrihert’s Kyle, Co. Tipperary, only one includes the name of the person commemorated (Okasha and Forsyth 2001, 223). Pádraig Ó hÉailidhe cites parallels for these uninscribed slabs that appear to date them to the 6th to the early 9th centuries (Ó hÉailidhe 1967, 122), while Fanning argues for a 9th-century date (Fanning 1976, 36). Whatever the date of the uninscribed slabs at St Berrihert’s Kyle, it seems that in early medieval monastic communities, memory and oral history, or possibly liber vitae of various kinds, were considered sufficient for the commemoration of the dead; uninscribed crosses may simply have prevented the intercutting of graves,

188 high island: excavation of an early medieval monastery

although different cross motifs may have been used to identify individuals (see ‘Functions’, below). Comparanda for the High Island cruciform slabs are found elsewhere in Ireland, for example, at Garranebane, Glanleam and Skellig Michael, all in Co. Kerry (O’Sullivan and Sheehan 1996, 277, figs. 176–177; 280–81, figs. 180–181) and in Irish-influenced areas of Scotland and Wales. At the monastic site at Ardwall Isle, Galloway, a site considered by Charles Thomas to ‘raise a strong presumption of Irish influence’ (Thomas 1967, 177), fragments have been found of uninscribed cruciform slabs with widely splayed arms that Thomas dates to the earlier part of his ‘Phase III’, that is, the 8th and 9th centuries (ibid., 155–8). On Iona, seven slabs are cruciform, of which three are also incised with cross motifs (RCAHMS 1982, 180–81, 189, 191–2, figs. 4, 15, 65, 72–76), two linear and one outline ring-headed. None of the Ionan cruciform slabs has been dated, although the general chronology of cross motifs suggests that the ring-headed Latin cross carved in relief on Iona 65 dates it to the 8th to 9th centuries (see above). Elsewhere in the western Highlands and Islands, twelve cruciform slabs are carved with a variety of cross motifs (Fisher 2001, 29, figs F, Z, AA, BB, CC; 32, figs. N and Q; 39, figs. L, S and X; 42, fig. 5; 45, fig. F), some resembling High Island slabs. None of these has been dated. However, a cruciform slab at Llanglydwen, Carmarthenshire, incised with an outline Latin ring-cross has been dated to the 9th to 10th centuries (Edwards 2007, 254–5, fig. CM26), while an elongated cruciform slab carved with a linear Greek cross at Llanmadog, Glamorgan, has been tentatively dated to the 7th to 9th centuries (Redknap and Lewis 2007, 359–60, fig. G56). It will be evident from these comparanda that it is only possible to date cruciform slabs from the type of cross incised or carved on them, which means that thirteen High Island slabs (nos. 19, 43–50, 54–55, 59 and 62) cannot be dated other than to the 7th to 12th centuries.

Conclusions

Study of cross-carved slabs of comparable contexts and incised motifs to those on High Island suggests that there were two main periods of sculptural activity, the first beginning in the 7th century, the second in the 10th century. This agrees with the results of the most recent archaeological excavations, which suggest a date range towards the end of this period for the church and graves at the east end of it (see Section 5). It has been noted that dating the High Island slabs has been made more difficult by the lack of inscriptions, a not uncommon scenario at Insular monasteries of the early medieval period. Dedications carved on grave-markers, votives or memorials in Roman script were not universal in Ireland until the later Middle Ages, while ogam inscriptions, which were easier to carve because they are

composed of straight lines, do not appear to have been used in Galway (see Charles-Edwards 2000, 174 (Map 8) for distribution map of carved ogam inscriptions), and in any case seem to have been a feature of early Irish literacy that had been supplanted by Roman script by the 9th century. Ogam inscriptions were most commonly used on boundary-markers, where, to quote Jane Stevenson, ‘their advantage as an objective record detached from the vagaries of human memory and prejudice came to be appreciated’ (Stevenson 1989, 147). This use seems unlikely to have been required on a very small island hermitage where the only boundary markers required were termons, which marked the enclosure or sanctified area that provided sanctuary for miscreants, when required. In addition, uninscribed grave-markers may reflect an ascetic theology and/or the difficulties and greater expenditure of resource involved in carving Roman text on extremely hard and uneven surfaces, since the High Island slabs, unlike those at Clonmacnoise, were not finished with smooth faces. At all events, it should not be assumed that the High Island community were illiterate, either in ogam or Roman script. The figure holding two books aloft on Cross-slab 7 reminds us that monastic tradition required literacy of its holy men and women, since Christianity is ‘the religion of the book’, using gospels and psalters, liturgical works and hymnals to meditate, preach and praise. Cuminius, a 7th-century abbot of Iona, directed that a priest who was unable to read the prayers and lections of the liturgy was not permitted to offer the sacraments during Mass (Warren 1881, xv, 98). The lack of inscribed dedications may be part of a continuing oral vernacular culture and a belief that the dead were known to God by name. Although the ecclesiastic community may have had a form of liber vitae or other written record of its holy dead, it is likely that community memories were considered sufficient to remember where the graves of members of the hermitage were located (see Maddern 2013, 51, 54). The lack of inscriptions means that it is even more necessary to ‘read’ the symbolism embodied in the shapes of the carved crosses to reach their underlying messages. The following section considers the symbolism of the various motifs embodied in the carvings, and in the figural sculpture. It also considers the major theological developments of this period that may be reflected in the carved details of the crosses.

Iconography

In Christian iconography the cross has a number of interrelated meanings, chiefly those of the Passion and of the Son of Man, who was prophesied in both biblical and apocryphal books to appear at the end of time, thus heralding the coming of Christ the final Judge. To quote George Henderson, ‘it is reasonable to suppose that

the cross-slabs 189

monastic art in the 7th and 8th century consciously aimed at [a complex] level of symbolism and allegory, of biblical cross reference and depths of meaning’ (Henderson 1980, 16); there is no reason to suppose that the symbolism of the cross became any simpler in the 9th to 12th centuries. One of the most important aspects of the sign of the cross in early Insular Christian culture was that it was considered to offer protection to all created things, both living and dead, that it marked. So it was used to protect churches, grave-markers, manuscripts, everyday tools and utensils, animals and humans, either in material or immaterial form (the latter through genuflection (making the sign of the cross on the body) and making the sign of the cross in blessing over others), many examples of which are given in Adomnán’s Life of Columba, which was written between AD 688 and AD 692 (Anderson and Anderson 1991, 5). This emphasis on the protective power of the cross is demonstrated by a peculiarly Irish innovation, the lorica, or breastplate prayer, a widespread form of apotropaic prayer in litany form that invokes the protection of the cross for every part of the body, versions of which survive in manuscripts dating from the 9th to the 14th centuries (Gougaud 1911). The protective function of the cross was often complemented by the use of another Irish motif closely identified with La Tène metalwork, the pelta, which features in both Irish manuscript miniatures and metalwork and in Irish and Scottish stone sculpture of at least the 7th to 12th centuries (Allen and Anderson 1903, 374–5; 377–80; Wright 1982, 99). A particularly vivid example of the complementary function of the cross and pelta motifs is seen on the 8th- or early 9th-century Athlone Cross, where three large peltae are arranged on Christ’s chest, suggesting a breastplate or shield. Here, the peltae symbolise Christ’s righteousness (Isaiah 59:17; Ephesians 6:14), faith and love (1 Thessalonians 5:8) (Harbison 1999, pl. 89). In this context the use of the pelta, whose primary meaning in Insular Latin is a light shield, appears particularly appropriate, recalling the loricae that invoke the Trinity and Christ’s protection (Howlett 2006, 2172). Significantly, the word pelta was used by both the 7th-century Anglian bishop Aldhelm and the 9th-century Irish ecclesiastic Laidcenn Mac Báith (to whom a lorica is attributed) in contexts requesting divine protection (Herren 1987, 24). The use of the pelta motif on Cross-slabs 1A and B, and possibly 2A, 8 and 9A, likewise suggests that, used in conjunction with the cruciform shape of the first two slabs and the expansional crosses that form the main features of the two latter slabs, they were designed to invoke the apotropaic properties listed in contemporary loricae, and accord well with their apparent function as headstones or grave-markers. Other symbols used in Insular Christian art are derived from those used in Late Antiquity. The earliest Christians,

persecuted by the Roman emperor Diocletian and others for their faith, resorted to the use of symbols to send subtle messages to other believers. The cross was avoided by the earliest Christians (for whom it represented a particularly undignified and humiliating form of execution); where it did appear, it was disguised as an anchor. At the same time the Greek letters χ ρ ι (chi-rho-iota) and Α and Ω (alpha and omega) became common symbols for Christ. Chi-rho-iota are the first three letters of the name ‘Christ’ in Greek (ΧΡΙΣΤΟΣ), while alpha is the first letter of the Greek word arche (‘beginning’) and in scripture was used to express the primeval beginning (Becker 2000, 13); omega is the last letter in the Greek alphabet and also means ‘last’. Because alpha and omega ‘encompass’ all other characters of the alphabet, they symbolise totality, God, and especially Christ as the beginning and end of all things. They also allude to the words attributed to Christ in the Apocalypse of St John, where he describes himself four times as ‘the first and the last, the beginning and the end’ (Apocalypse 1: 8 (Weber 1994, 1882); Apocalypse 1: 17–18 (Weber 1994, 1883); Apocalypse 21: 6 (Weber 1994, 1903); Apocalypse 22: 13 (Weber 1994, 1905)). The alpha and omega symbols thus became not only a secondary motif of Christ’s monogram but also apocalyptic symbols for Christ (Becker 2000, 15). The chi-rho-iota and alpha and omega symbols are found, singly or together, carved on stone slabs dating to the 6th to 9th centuries at Insular sites, such as that dedicated to the mid-8th-century Northumbrian scholar and saint, Berectuine, at Tullylease, Co. Cork, which also includes a prayer formula and cross design clearly linking it to the book arts (Henderson and Okasha 1992, 32; Okasha and Forsyth 2001, 119–23; Maddern 2013, 181–2). Occasionally the chi-rho was incorporated in the form of a hook attached to the upper limb of a cross rather than incised initials; the ‘hooked cross’ is seen on, for example, the Echoidus stone at Iona (RCAHMS 1982, 182, fig. 22B–D), and on nine slabs in Ireland, mainly at ecclesiastical sites on the western fringes, some of which Herity considered to date to the mid-7th century (Herity 1990b, 214), and which are certainly no later than the 9th century. At High Island, it is possible that the circular spiral or boss attached to the upper left limb of the ribbon cross on Face A of Cross-slab 22 may be a misunderstood representation of a chi-rho symbol (which is normally a hook attached to the upper right limb of a cross), but this must be considered doubtful. This lacuna at High Island is shared with cross-carved slabs in south-west Wales (Edwards 2007, 89) and tends to support the conclusion reached above that the main period of sculptural activity at High Island was comparatively late. By the 7th century, when St Féichín of Fore was reputedly establishing monasteries along the western fringe of

190 high island: excavation of an early medieval monastery

Co. Galway, including that on High Island (Stokes 1892, 4–5; Moran, this volume, p.16), the ‘unhooked’ cross had become the primary symbol for proclaiming allegiance to Christ and claiming his protection, and it was normally rendered in linear form. However, eleven or twelve of the High Island linear crosses are contained in a circle (see Table 7.1). This form was considered by Nash-Williams to be a development and simplification of the chi-rho motif enclosed in a wreath or circle, which had developed in the 5th century (Nash-Williams 1950, 16) and is therefore likely to be amongst the earliest motifs to have been carved on the High Island slabs. A further development was the extension of all four limbs of the cross outside a ring surrounding the crossing. Lionard cites Ó Ríordáin in suggesting that this was due to the imitation of wooden prototypes in stone carving (Lionard 1961, 125), but it may, in fact, have been a reference to resurrection and everlasting life through the cross, since the ring had been the symbol of eternity from earliest antiquity (Ladner 1995, 99–100, 110). One of the most distinctive forms of linear cross at High Island has forked terminals. Some are widely curved, others have short, straight terminals. These are generally later than plain linear crosses and may be based on metal prototypes; the bifurcation of the ends of the limbs imitates the effect achieved when iron bars are split and the ends separated. However, many of the High Island linear terminals are too widely curved to be skeuomorphs of metal crosses. In these cases, one of the intentions in emphasising the terminals may have been to refer to the symbolism attached to the numbers four and eight, respectively. Number symbolism was taken very seriously in the early Middle Ages, as it had been in antiquity. The Late Antique author Martianus argued that the number four pervades the world and human life, citing the seasons, compass points, elements, ages of man and cardinal virtues; in the Bible, the number four represented the four rivers of Paradise and the four evangelists (Weber 1994, 6; Ladner 1995, 105; Werner 1994, 456–7). By splitting the terminals into two, however, eight elements result. The number eight and the octagon represented both eternity and resurrection in the early Middle Ages, this being the reason that baptismal fonts and baptisteries, which enabled believers to enter into eternal life through Christ, were octagonal (Ladner 1995, 106). It is therefore possible that the forked terminals at High Island refer both to the baptism, which enabled believers to be ‘reborn’, and to eternal life, making them particularly appropriate for grave-markers. Number symbolism may also provide the clue to the meaning of the circles or bosses that appear in varied numbers on some High Island slabs. Cross-slab 4B, for example, has a boss with hollowed centre in each of the

quadrants formed by the limbs of the Greek cross. In this context their most likely meaning is to refer to the four evangelists, to each of whom is attributed an account of the central historical event in Christian belief, the crucifixion. If this is correct, then the bosses in the upper corners of the slab – one whole but possibly originally hollowed, and one hollowed – may be intended to represent the sun and moon, or the eclipse that is described in the synoptic gospels as taking place at the historical Crucifixion from the sixth to the ninth hour (Weber 1994, 1572, 1603, 1655). Similarly placed large bosses on either side of the upper limb of a ringed outline cross at Dòid Mhàiri, on Islay, were considered by Ian Fisher to represent the sun and moon and to date to the second half of the 11th century (Fisher 2001, 136, no. 351). The sun and moon also appear on stylised crucifixion scenes on Irish High crosses, such as that on the head of the south face of the 9th- or 10thcentury Muiredach’s Cross at Monasterboice, Co. Louth (Harbison 1994, 89). This interpretation is made even clearer by examples of manuscript illuminations that are approximately contemporary with these carved slabs, for example folio 232r of the late 10th-century Gospels of Otto III, where two rounded personifications of the sun and the moon are placed at either side of the head of a cross bearing the crucified Christ (Aachen, Domkapitel, Liuthar or Aachen Gospels; Nees 2002, 13, fig. 4). A Latin cross with four circles in the quadrants formed by the limbs of the cross and a shield-shaped surround surmounted by two small circles over the upper limb of the cross at Llys-y-Frân, Pembrokeshire, was suggested by Nancy Edwards to be a representation of a crux gemmata or jewelled cross (Edwards 2007, 375–6, fig. P57). This was a motif that first appeared in the early 4th century, after the discovery of the True Cross by the Roman Emperor Constantine or his mother, Helena. The ‘True Cross’ was encased in gold and jewels by Theodosius in AD 417 (Harbus 2002, 44) and widely reported by pilgrims, including the Gaulish bishop Arculf. His account of the holy places of the Near East, De Locis Sanctis, as narrated to Abbot Adomnán of Iona, began to circulate around AD 690. This was a widely circulated and much-studied work in Insular ecclesiastic circles, since the Northumbrian scholar Bede not only produced a synopsis in his Historia ecclesiastica but also an independent version. O’Loughlin has demonstrated that manuscripts of Bede’s version were copied or paraphrased extensively from the 9th to the 13th centuries, including a Middle Irish translation in the Leabhar Breac; the number of surviving manuscripts suggests that the work was growing in popularity in the 12th to 13th centuries (O’Loughlin 2000, 101). However, the circles on the Llys-y-Frân slab are not placed on the shaft and limbs of the cross because it is linear, thus weakening Edwards’ suggestion. More persua-

the cross-slabs 191

sive is her suggestion that parts of the design may have been inspired by ampullae, or oil-flasks, brought back as souvenirs from the Holy Land by pilgrims. Helen Roe has explained how these ampullae were decorated with scenes from the life of Christ, based on models found in a group of artworks dating to the latter half of the 6th century. In particular, the Rabbula Gospels Crucifixion miniature of around AD 586–87 is considered to have influenced all later versions of the Crucifixion. This miniature depicts Christ fully clothed and nailed to the cross, surrounded by the crucified thieves, the sponge-bearer Stephaton and lance-carrier Longinus, the sun and moon, the Virgin Mary and St John, and the soldiers casting lots (Roe 1960, 200–201). Victor Nash-Williams (in the context of a slab at Llanychaer, Pembrokeshire, that does not include the circles but portrays a garbed Christ on the cross), likewise drew attention to Coptic plaques and textiles of the 6th and 7th centuries that show a simplified Crucifixion scene showing Christ on the cross, fully clothed in a longsleeved colobium or tunic, flanked by the two thieves on their crosses and a hollowed circle on either side of Christ’s head (Nash-Williams 1950, 191). It therefore seems likely that the persons who commissioned the carvings on High Island 4B, Llys-y-Frân and Dòid Mhàiri were following a convention that, rather than creating a complex figural scene, used symbols to represent the sun and moon as ‘shorthand’ for the full Crucifixion scene that their audience would have read about (or have had read to them) in the four gospels, and possibly have seen in manuscript miniatures or ampullae of the early medieval period. The ‘shorthand’ type of symbolism may also be the explanation for the unusual image on the reverse face of the slab (Cross-slab 4A), which shows a Latin cross on a ‘stand’ sited on top of a tall rectangular shape that may represent an altar, but might equally well be compared to crosses ‘growing’ out of coffins or sarcophagi that symbolise the Resurrection. These motifs are found in several Irish hermitages and monasteries, such as Inishmurray, Co. Sligo, and Inishcaltra, Co. Clare (O’Sullivan and Ó Carragáin 2008, 90, fig, 21; 123, fig. 30; Okasha and Forsyth 2001, 78–83). The connection between tomb, cross and resurrection is fundamental in early Christian theology. The Gospel of Peter, an apocryphal fragment dating to around AD 150 that was discovered in AD 1884 in a tomb at Akhimin in Egypt, bound with other Apocalyptic writings such as the Book of Enoch, sets out a variant version of the resurrection, where Christ is brought out of the tomb by two shining beings, followed by a cross. A voice from heaven then asked the cross whether it had preached to ‘them that sleep’, the answer being in the affirmative (James 1924, 92). This is an early use of prosopopoeia, the attribution of human emotions and speech to an inanimate object,

the best-known early medieval example of which is the Old English poem The Dream of the Rood (see Fleming 1966; Ó Carragáin 2005). Although there is no evidence that Irish scholars of the early medieval period had access to the Gospel of Peter, it is clear that variant apocryphal accounts of the resurrection were known in Irish monastic circles. Martin McNamara describes a ‘rich Irish tradition in apocryphal literature’, 39 examples of which, dating between AD 650 and 800, were itemised by Bernard Bischoff (Herbert and McNamara 1989, xiii; Bischoff 1954). It is intriguing, therefore, that Cross-slab 4A shows a large cross, apparently on pedestal feet, balanced on top of a rectangular base. These two parts may be intended to illustrate the connection between the tomb and the Resurrection of Christ, a motif also used on Late Antique sarcophagi as part of a stereotypical programme of motifs designed to reflect resurrection theology. For example, a marble sarcophagus dating to the 6th century in the basilica of St Vitale, in Ravenna, includes in its decoration a sculpted depiction of an orant Daniel between two lions at one end of the base, over which, on the lid, is a straightarmed Latin cross with straight edges to the limbs, on a rectangular stand. This arrangement reflects the custom of using the Old Testament to prefigure the New, in this case the story of Daniel’s ordeal in the Lions’ Den as an antecedent of Christ’s death on the cross and subsequent resurrection from the tomb. This gives the key to the symbolism of the High Island cross-on-rectangle, which appears to refer to the expectation of bodily resurrection from the dead. A further layer of symbolism is suggested by the tricuspid terminals on the two ‘feet’ of the cross, which may refer to the Trinity or to the three days between Christ’s death on the cross and his resurrection. If this interpretation is correct, then the carving on both faces of Cross-slab 4 follows a well-established tradition of showing sequential episodes from the gospels on separate panels, as is seen in many Insular High Crosses from the 8th century (for example, the Ruthwell Cross in Northumbria) and the 9th century (such as Muiredach’s Cross at Monasterboice, Co. Louth (Harbison 1994, 13, 85–91)). Despite the tendency to use symbols as shorthand for more graphic images, figural sculpture is not entirely absent at High Island. A slab sketched by Wakeman in 1839 but now missing, Cross-slab 61, appears to have been carved with a human figure. The figure had the almost triangular head typical of the 9th to 11th centuries and an abbreviated torso. The drawing is too indistinct to determine whether the figure was carrying any object, but the arms were apparently not raised. This contrasts with Crossslab 7, where the triangular head has sketchy features, the body is virtually full-length and both arms are raised in the orans position, which was the standard method of showing

192 high island: excavation of an early medieval monastery

a person praying from Antiquity to about the 13th century. The figure on Cross-slab 7 is interesting in two respects. First, the figure is apparently not praying but holding aloft two rectangles that are almost certainly intended to represent books. White Marshall and Rourke state that ‘a cross within a circle clearly decorates one of these books and it is possible in good light to see another cross on a somewhat larger book in his left hand’. These decorations were not visible in 2010, unfortunately, but such signs would have been appropriate for ecclesiastical books, book satchels or chrismals (boxes containing the reserved Host or communion wafers). They also noted that a figure carrying a book in each hand is very rare in Insular art and cited the sole example known in Insular manuscript art: the abbreviated figure of St Matthew painted above the frame of the canon tables on folio 4v of the Book of Kells (White Marshall and Rourke 2000, 149). In the Book of Kells the meaning of the two open books may be to refer to the canon tables themselves, since their purpose was to provide cross-references between two or more gospels. This suggests that the figure carved on Cross-slab 7 may indeed be carrying two books, and have been intended to represent an academic or scribe. It seems less likely that the objects held by the figure standing in what was probably a boat were satchels for books, or that they were chrismals, special warnings against the loss of which on boats were given in chapter 8 of Cuminius’ de Mensura Poenitentiarum (Warren 1881, 138). An additional layer of symbolism may have been intended; Carol Farr has explained that the palm-outwards and raised arm pose was not only that of prayer but also of the Crucifixion (Farr 1997, 106). Since Cross-slab 7 was used as a footstone and this image faced towards the body, it could have supplemented the cruciform shape of the slab in claiming Christ’s protection for the buried person. Orant figures without books are found carved on slabs elsewhere in the Insular world in the 9th to 11th centuries, for example: on the ‘Adam and Eve’ slab at Llanhamlach, Breconshire, that dates to the 10th to 11th century (Redknap and Lewis 2007, 210–13, fig. B32a); on a 9thcentury slab at Llangyfelach, Glamorgan (Redknap and Lewis, 352–5, figs. G52b–e); and on a 10th-century slab at Cadoxton-Juxta-Neath, Glamorgan (Redknap and Lewis, 277–8, figs. G10a–b). All of these figures were originally associated with carved crosses, but none of them holds anything in their hands, which are extended palm outwards. However, the Llangyfelach orant figure, with head, hands and feet protruding from what appears to be a carpet-page from an illuminated gospel book, has affinities with the image of St Matthew in the early 8th-century Book of Durrow (Dublin, Trinity College MS A.IV.5 (57), folio 21v). The St Matthew evangelist page shows a head and small sideways-pointing feet emerging from a densely

patterned body resembling an inlaid jewelled buckle or part of a carpet-page (Meehan 1996, 34–5). The Llangyfelach figure hints at the close relationship between prayer and study of biblical texts at this period. Likewise, Crossslab 7 suggests that the person for whom the stone was carved, or re-carved, in the 10th to 11th century was both learned and prayerful. However, the fact that the torso was not rendered as a carpet-page as at Llangyfelach may indicate that the focus of the person depicted was on prayer or evangelism rather than scribing. The second interesting aspect of the figure carved in the lower half of Cross-slab 7 is that it appears to be standing in a boat with a sharply pointed prow and rounded stern. The prow is represented by a vertical line between the figure’s gown, where the legs could be expected to be, and by oblique lines running from the sides of the frame to the outline of the figure’s gown. This may have been a representation of a local curragh or hide-covered boat, the designs of which appear to have remained constant for many centuries. James Hornell, who made a comprehensive survey of curragh types in the 1930s, shows that the type of curragh used in Connemara and the Inishbofin Islands had a sharply pointed prow and rounded bottom; although the stern itself was apparently formed from a straight piece of timber, the intention of the sculptor of the figure on Cross-slab 7 seems to have been to indicate the type of craft by emphasising one of the main features of its construction: stretching hides over a frame of withies to form a semi-circular profile (Hornell 1938, 13–15, 26, plate I, fig. 2, plan 3). Hornell also considered the early history of the curragh. He noted that it featured in early Irish pagan mythology, particularly in imramha or tales of voyages in search of an earthly Paradise. Irish fascination with these tales continued into the time of the early Christian Church, resulting in accounts of the travels of saints and monks either ‘in search of a quiet home on some uninhabited island or with a view to carry the torch of their faith to the benighted people of other lands’. These travellers included St Brendan and St Columba (Hornell 1937, 78–80). The motif of a man in a boat may also have referred to St Peter, the fisher of souls, or even to Christ himself. The motif of a man, possibly in a boat, with arms raised in prayer, greeting or admonition, as used on a footstone that was part of the monumental sculpture assemblage at the east end of the church (in other words, the most sanctified spot in a hermitage or monastic enclosure for burials in the early middle ages in Ireland), raises the question of the identity of the late 10th- or early 11th-century body buried there. White Marshall and Rourke speculate that this personage may have been Gormgal, who was the anmchara, or soul-friend (confessor), of all Ireland at the end of the 10th to the beginning of the 11th century, and

the cross-slabs 193

who died in either 1017 or 1018 according to the Annals of Ulster (White Marshall and Rourke 2000, 224; Gwynn and Hadcock 1970, 386). George Petrie elaborates the bare facts set out in the Annals with the following account of some of the 11thcentury monastic inhabitants of ‘Ard Olen’, compiled mainly from documents assembled by John Colgan in the 17th century and incorporated in an account of ‘The Territory of West Connaught’ in AD 1684 by ‘the learned O’Flaherty’: It is … celebrated for the eremitical retirement of St Gormgall, a very spiritual person and of renowned sanctity, who died on the 5th of August 1017, and was there interred, together with diverse other holy hermits that lived with him. Ten of them are named by Father Colgan… . (Petrie 1845, 427) It should not be inferred from this that the community on High Island in the 10th to 11th centuries was composed of a few hermits. Thomas Charles-Edwards, in the context of Columbanus and his monasteries, notes that a very few monks became hermits, ‘the monastic form of the other pattern of the religious life in which the individual sought God by himself, not in a community’, and that some Irish monasteries had places set aside for hermits (CharlesEdwards 2000, 383). This may have been the case at High Island, since the archaeological evidence suggests that the monastery was a flourishing and busy place in the 10th and 11th centuries rather than a refuge for recluses. In support of their suggestion that Gormgal (or one of his companions, ten of whom were reported by a contemporary, Corcrán, to have shared his retreat to an ascetic life on the island and to have been buried with him) was buried at the east end of the church on High Island, White Marshall and Rourke cite the position of Cross-slab 7 and the date of the remains within the grave it marks (White Marshall and Rourke 2000, 222–3). Although there are difficulties with Petrie’s sources of information about Gormgal (since his main source, John Colgan, was a hagiographer dependent on somewhat exiguous literature of the 11th century), it seems likely that Gormgal was resident on the island in the late 10th or early 10th century. If he did indeed ‘retire’ to High Island to pursue a her-

mit’s life, he would probably not have been the first such inhabitant on the Island. The Annals of Ulster show that the number of anchorites who died at sites such as Clonmacnoise increased from three in the 7th century to a peak of 32 in AD 800–850; at the time of Gormgal’s death, this had declined to twelve (Dwyer 1981, 10). However, High Island is not listed as one of the monasteries that had a hermit or hermits at any time between the 7th and 12th centuries, nor is it mentioned in the context of the ascetic reform movement, members of which were known as Célí Dé (companions of God) (Dwyer 1981, 10). Gormgal’s role on High Island is therefore uncertain (for a wider discussion on this topic, see Section 2). The suggestion that the graves at the east end of the church were those of the holiest inhabitants of the island, who died in the late 9th to late 10th century at the earliest and the mid-12th century at the latest, leads to consideration of the symbolism of the motifs carved on the slabs covering them, or acting as head- or footstones. The footstone for Grave 1 and the headstones for graves 2 and 3 have been considered under Typology subgroups 1(a) and 1(c) above. The recumbent stones on Graves 2, 3 and 8 are incised with double-ended outline versions of the expansional cross that appears to be unique to High Island. The first of these recumbent cross-slabs, Cross-slab 3, like the footstone for the same grave, Crossslab 4 (discussed above), is carved from garnet mica-schist indigenous to the island. It is roughly trapezoidal in shape and includes two expansional crosses joined end-to-end by a long shaft. The two central roundels and three of the six D-shaped terminals are filled with three-line geometric motifs; it seems likely that the three empty terminals were originally carved with the same geometric infill, now flaked away. Cross-slab 6, like the footstone for the same grave, Cross-slab 7 (discussed above), is carved on shelly fossiliferous limestone, probably from the Aran Islands, Co. Galway. Since these were apparently the only two slabs imported to High Island (a difficult procedure given the weight of the stone and the general inaccessibility of the monastery) and that Cross-slab 6 has been carved in a more refined style than the other expansional cross-slabs on the island, the person who was originally interred

Table 7.2 Cross-referencing of grave-markers (by catalogue no.) to Graves (1–8) outside the east wall of the church.

Grave

Headstone

Recumbent

8 14

Sidestone Footstone

194 high island: excavation of an early medieval monastery

in the grave may have had significant connections with the Aran Islands, since, as previously noted, the slab that marks this grave is carved from stone acquired with what must have been considerable difficulty from those islands, and is one of only two slabs carved from such stone on High Island. This certainly suggests the person for whom the slab was carved was particularly revered. Cross-slab 14, of which only approximately two-thirds are visible (since the stone is partially sealed by the church wall, the remaining third may be intact), was probably also carved with a double-ended expansional cross with geometric infills, now mostly flaked away. It differs from the two previous recumbent slabs in that the design, when complete, appears to have featured two Latin outline expansional crosses that overlapped for the lower third of their shafts rather than being joined together by one continuous shaft. Two other slabs with the more usual form of expansional cross, Cross-slab 9A and 11, were used respectively as a headstone and footstone for a late 9th- to late 10th-century burial. Of the other expansional crosses, Cross-slab 15 was used as a side-slab for the altar in the church, while Cross-slab 20 was possibly a grave-marker but was moved from its find-spot in the 19th century, making it difficult to allocate a specific function. Cross-slabs 34 and 39, which were found in monastic rubble, may also have been grave-markers but, equally, could have been incorporated in the church or its altar for liturgical use, particularly the former, which includes the crosses of the two thieves at Christ’s Crucifixion (see below). Cross-slab 64, which is thought to have been imported from the Women’s Cemetery on Omey Island, was probably a grave-marker. The key to the meaning of the expansional crosses carved on all these slabs is, first, that they may all be associated with graves (with one exception that was part of the church altar, and possibly two further exceptions that may be associated with the church liturgy or the altar), and secondly, that they all date to the 9th to 11th centuries, in other words, in the century leading up to, and just after, the year AD 1000. It has been argued elsewhere that Northumbrian name stones dating to the late 7th to early 9th centuries that are carved with expansional crosses reflect a complex matrix of ideas centering on the second coming of Christ at the Last Judgement. These ideas were pressing because early ecclesiastical computations had suggested that this event would take place in or around the year AD 800 (Maddern 2013, 169–72). The second wave of millennial anxiety centered round the year AD 1000 (Landes 2000; Fried 2003), since it was believed that the ‘sixth age’, which equated to the days of Creation as described in Genesis and to 6,000 years after Creation, would then end, ushering in the Apocalypse; Ireland was considered by Fried to be ‘one of the main

influences of millennial expectations’ (Fried 2003, 43). Visionary or prophetic references to an Apocalypse and Last Judgement occur in both the Old and New Testaments and in the extensive body of apocryphal and vision literature. Ireland has a rich tradition of such literature (Herbert and McNamara 1989, xiii), which, it has been argued, was widely known in the early middle ages in Ireland (Dumville 1973, 308). In canonical scripture the main references to the last days are found in Daniel, the synoptic gospels, the Pauline letters, I Corinthians and I and II Thessalonians, and the Apocalypse. It was believed that there would be a general Judgement on the Last Day, when Christ appeared with his angels in flaming fire to punish unbelievers and those who disobeyed the Gospel. Their punishment would be eternal destruction and exclusion from God and His glory (II Thessalonians 1: 7–11). In response to this threat the ‘name stone model’, with local variations, appears to have been continued in Ireland to make the same points as the earlier carved slabs; that is, to proclaim the allegiance of the person commemorated by the stone and to claim their place at Christ’s right hand on the Last Day. The expanded centre and terminals of Northumbrian name stones were normally left blank, although some contained motifs such as crosses-of-arcs in their central expansions. An early medieval monastic accustomed to the art of lectio divina – or, rather, visio divina – would have been able to supply images or references to fill the empty spaces or, in the case of those with infills at High Island and Clonmacnoise, to have interpreted them appropriately. To quote Robert Stevenson: ‘the recognition of variously shaped crosses would be as natural for people trained or brought up to the relatively restricted repertoire of early Christian art, as for us to recognise an important word spoken in different accents or heard indistinctly’ (Stevenson 1981–2, 2–3). In the mind’s eye of an educated ecclesiastical viewer around the turn of the 10th to 11th century in Ireland, large central expansions incised on grave stones would have recalled images of Christ glorified in majesty or as Agnus Dei, images of which – painted in manuscript miniatures, or set in mosaics – the viewer had himself seen, or had read or heard about from visitors to Gaul, Rome or Jerusalem. These would certainly have included two basilicas in Rome repaired by Pope Sergius towards the end of the 7th century, Old St Peter’s (which displayed mosaics of Christ both in Majesty and as Agnus Dei worshipped by the Evangelist beasts) and SS. Cosmas and Damian (which also displayed Christ in Majesty and as Agnus Dei surrounded by elders and evangelists in the apse mosaic) (Ó Carragáin 2005, 248–55). The alternative image would have been of the crucified Christ, but images of the glorified Christ as man or Lamb of God appear more likely

the cross-slabs 195

since they are the Apocalyptic partners of the four Evangelist beasts that the educated ecclesiastic viewer was most likely to have ‘seen’ in the four expanded terminals. The combination of the glorified Cross and the Evangelist beast symbols was important from the late Antique Period onwards, for example depicted in mosaic at S. Pudenziana at Rome, in the 5th-century clipeus mosaic of a bishop in the Catacomb of S. Gennaro at Naples, and in vault mosaics at the 5th-century Mausoleum of Galla Placidia and the 6th-century Cappella Arcivescovile (where a chi-rho rather than Latin Cross is depicted), both at Ravenna (Nees 1978, 4–5, citing Fasola 1975, 133 ff. and fig. 87; Lowden 1997, figs. 63, 66). Closer to home, beginning with the (probably) early 8th-century the Book of Durrow, the arrangement of the four Evangelist beasts around the Cross was common in Insular manuscripts, either as a frontispiece to the codex, as in the Book of Durrow (Dublin, Trinity College MS A.IV.5 (57)) and the Macdurnan Gospels (London, Lambeth Palace MS 1370), or preceding individual gospels, as in the Lichfield Gospels (Lichfield, Cathedral Library, MS s. n.) and the Books of Armagh (Dublin, Trinity College MS 52) and the Book of Kells (Dublin, Trinity College MS A.I.6 (58)). In the Insularinfluenced Trier Gospels (Trier, Cathedral Treasury MS 61), the Evangelist beasts are grouped around a portrait bust of Christ holding a book, thus identifying the figure as the Maiestas Domini (Lord in majesty). To quote O’Reilly: The Insular adaptation of the Early Christian convention of reading the exalted Cross as an image of the glorified Christ enables the four-symbols page to be read as an evocation of the apocalyptic vision of the four living creatures around the throne of Christ’s majesty. (O’Reilly 1998, 87) Cross-slabs 3 and 6, both recumbent slabs used to cover graves, have multiple transoms and D-shaped terminals. Michael Herity commented that crosses with multiple transoms are rare in early Christian art; none of the comparanda he noted (at Ardmoneel and Teampall Geal, Co. Kerry, and at Inishmurray, Co. Sligo) bears any resemblance to Cross-slabs 3 or 6. Herity noted, however, that a cross with double bars and with square expansions at the ends of the terminals is depicted in a carpet-page on folio 1v of the Book of Durrow (Dublin, Trinity College MS A.IV.5 (57)) (Herity 1990b, 213–14, fig. 1). It faces a fourevangelist carpet-page, folio 2r, where the evangelists are set in the quadrants around a cross that has an expanded centre and terminals that extend to the border of the page (Meehan 1996, 15). A similar carpet-page showing a cross with double bars and round expansions is depicted on folio 33r of the Book of Kells (Dublin, Trinity College MS A.I.6 (58)). The Kells carpet-page faces a very elaborate chi-

rho page and therefore could be expected to have much the same symbolical meaning as that of Durrow. Martin Werner considered that the eight-circle cross on folio 33r of the Book of Kells represented a Patriarchal cross, the double-barred emblem of the True Cross or Cross of the Crucifixion, which had been described by Abbot Adomnán of Iona in De Locis Sanctis (see above). Folio 33r could, Werner argues, be a highly abstract representation of fragments and the titulus of the True Cross (the horizontal piece of wood on which Pilate inscribed Christ’s title in three languages) that was displayed on the altar of the church of Golgotha in the Holy Sepulchre complex at Jerusalem for the Good Friday Adoratio Crucis (adoration of the Cross) ceremony (Werner 1994, 455–6). However, early medieval symbolism was multivalent, so we should not suppose this was the end of the interpretations that could have been made of the High Island multiple-transomed crosses by a 10th- or 11thcentury ecclesiastic. There are a total of eight expansions on the crosses depicted on both folio 1 of the Book of Durrow and folio 33r of the Book of Kells. Cross-slab 3 includes six D-shaped terminals and two round central expansions, Cross-slab 6 has eight D-shaped terminals and two round central expansions, Cross-slab 14, when complete, may have had six D-shaped terminals and two round central expansions. This suggests that, in addition to referring to the coming Day of Judgement and the True Cross, the three High Island recumbent slabs made use of number symbolism. The number six was considered to be perfect in Late Antique philosophy because it was the sum and product of its parts; for St Gregory the Great it represented the perfection of the Creation in six days (Ladner 1995, 105–6). Likewise, since the earliest days of Christianity the number eight was considered to symbolise resurrection and eternity (ibid., 106) and was therefore an appropriate number for both Christian grave stones and carpet-pages in illuminated gospel-books. Cross-slab 34, a kite-shaped slab carved with an outline expansional cross, is unusual in that the terminals are forked, rather than D-shaped, and even more unusual in being incised with two small linear crosses on either side of the shaft of the main cross. The ‘Christ between two thieves’ motif is found on a surprising number of crosscarved slabs at early Irish eremitical and turas settlements, given that it is largely absent from Anglo-Saxon sculpture of the early medieval period. Apart from a group of three linear crosses carved on a pillar at Whithorn (a monastic settlement on the coast in present-day Dumfries and Galloway, part of the Anglo-Saxon kingdom of Northumbria in the 8th century) and four slabs at three ecclesiastic sites near or in the city of Lincoln that fall stylistically into two groups dating respectively to the 10th and later 10th or early 11th century (Everson and Stocker 1999, 172–5,

196 high island: excavation of an early medieval monastery

197–8, 203, 209; pl. 187–189, 249, 256), this motif is conspicuous by its absence. This may reflect an Anglican preference for the alternative scenario of Christ between Mary and John, which found its fulfilment in the art of the 12th century following the teachings of St Anselm and Bernard of Cîteaux on the Virgin Mary. In Ireland and Irish-influenced areas, however, the ‘Christ between two thieves’ motif occurs repeatedly. Higgins noted seven examples of non-figural three-cross groups, mostly linear, in Co. Galway, all of which have a large central cross with smaller ones either resting on its lateral limbs or positioned on either side of it. Ecclesiastic visio divina skills were needed to interpret some of these linear crosses, since the three separate crosses are not immediately apparent in every case. Other slabs carved with linear or outline interpretations of this theme have been found at Gallen (Co. Offaly) and Reask (Co. Kerry), as well as the figural example at Inishkea North (Co. Mayo). However, the greatest number and particularly interesting selection of slabs carved with variations on this theme at a single site are found on Inishmurray slabs 001.53, 001.57 and 001.66, although the style in which it is depicted differs radically between the three slabs (O’Sullivan and Ó Carragáin 2008, 124, fig. 24; 127, fig. 25; 131, fig. 26). Theologically, the theme of the two thieves is important in emphasising the concept of forgiveness for the truly penitent, since the early medieval period was a time of considerable anxiety about an impending Last Judgement. The importance of the motif is reflected in the Good Friday liturgy set out in the 8th-century Bobbio Missal (MS Paris Lat. 13246), probably written at Pavia at an establishment with links to the Columban monastery at nearby Bobbio (Vogel 1986, 323). Folios 202r–v of the Missal set out the Lictio Sancti Aeuangeli secundum Iohannem, the reading from the gospel according to St John the Evangelist, one of three readings for ad sexta parasceve, that is, Good Friday. St John’s gospel was particularly revered in early Insular monastic communities because it was considered to include a first-hand account of both the Crucifixion and Resurrection, as written by a disciple close to Christ. Thus, St Cuthbert of Lindisfarne, who had anecdotally cured the sick by laying a copy of the gospel of St John on them, was buried with a copy of that gospel; Bede was engaged in a translation of the gospel from Latin to the vernacular shortly before his death (Mynors 1956, 358–62, fn 1; Colgrave and Mynors 1969, 582–3). The Bobbio Good Friday reading is not, however, taken from the gospel of John. Verse 19: 18 of that gospel merely states that ‘two others’ were crucified with Christ, as does Mark 15: 27–32. The Bobbio lection records the episode where the two thieves speak at the Crucifixion, and is a composite of Matthew 27: 38–44 and Luke 23:

39–43. These two passages emphasise the instant access to heaven promised to the repentant thief, setting the tone for the Good Friday office. This may have been the main reason for the popularity of the ‘Christ between two thieves’ motif in Ireland. The Irish Church embraced the concept of penance and the penitential from an early date, as witnessed by the large number and variety of surviving penitential texts (Bieler 1963). In addition, the good thief featured in at least one apocryphal text translated into Gaelic, for example the Gospel of Nicodemus (Herbert and McNamara 1989, 86–7); as mentioned previously, such apocryphal literature was widely disseminated and studied in early medieval Ireland. As well as liturgical and literary references, sculptured references to penitence were extremely appropriate for communities such as High Island. As Thomas Charles-Edwards has demonstrated, penitential prayer for members of the monastic community was an integral part of the role of monks in 7th- to 8th-century Ireland (Charles-Edwards 2000, 117–18). The presence of so many slabs carved with the ‘Christ between two thieves’ motif on the western fringes of Ireland suggests that emphasis on this vocation continued in Irish eremitical and monastic communities for at least four further centuries. The repentant thief was a timely reminder of the possibility of forgiveness. Another High Island expansional cross (number 11) seems likely to have been multivalent in its symbolism. Very unusually, it is anthropomorphic, with details of face, hands and feet very sketchily incised. Carved on a roughly triangular slab that was used as the footstone to Grave 5, in addition to referring to the second coming of Christ in glory, it almost certainly represents the crucified Christ. A parallel may be seen on folio 291v of the Book of Kells, the St John portrait page, where a nimbed head (now partly trimmed away) is depicted outside the border of the page, hands on each side of the page and feet at the base of the page (Meehan 1994, 39). Since the feet have red circles in the heels representing the wounds made during the crucifixion, and St John is depicted with the scribal tools (pen, inkwell and book) he would have needed to write his gospel, the disembodied person must be Christ, the subject of St John’s writing. The similar iconographic scheme of Cross-slab 11 indicates that both Christ’s Passion, resulting in man’s salvation, and Christ’s second coming on the Last Day are invoked by this stone. Here, the use of an anthropomorphic figure contrasts with the comparatively common custom in early Christian Ireland of depicting a roughly triangular face in the appropriate part of a cross-slab, which is seen at, for example, Clogher, Co. Tyrone, Kilbroney, Co. Down, and at Caher Island and Knappaghmanagh, Co. Mayo (Roe 1960, figs. 1–2, plate VI). Helen Roe thought that these cross-slabs all dated to the mid-7th century, but it is now generally agreed that

the cross-slabs 197

Robert Stevenson’s dating of the 10th century is more likely to be correct (Stevenson 1956, 94–6). This, in turn, suggests that Cross-slab 11, as part of an Irish movement towards depicting Christ on the Cross in sculpture, also dates to the 9th to 11th centuries. There were surprisingly few outline crosses on High Island, given that this style of interpreting the cross was more common after the 9th century than before it in the Insular world. In particular, only five ribbon crosses (nos. 4A, 4B, 22A, 38 and 41) survive. Redknap and Lewis commented that: ‘it has been argued that the complex ornament on the pages of gospel-books, with its many visual ambiguities, may have lent itself to spiritual meditation, and the complex decoration on crosses…may well, in addition to their other functions, have invoked a similar response (Redknap and Lewis 2007, 94). This certainly seems to be true of the High Island outline crosses. The complex symbolism on Face A and Face B of Cross-slab 4 has already been noted. In addition to the symbolism imbedded in the cross depicted on Face B, this face of the slab includes a motif that was almost certainly intended to be a triquetra underneath the cross, symbolising the three persons of the Holy Trinity: Father, Son and Holy Spirit. This motif is also found on Crossslab 22A, only one of a probable four triquetrae carved in the terminals having survived the depredations of time and Atlantic weather. Cross-slab 22A is also interesting in that the ribbon moulding interlaces to form a lozenge around a central roundel of now indecipherable geometric ornament. As Jennifer O’Reilly has pointed out, the lozenge appears in many Insular and Carolingian manuscript miniatures, for example on folio 290v of the Book of Kells (Dublin, Trinity College MS A.I.6 (58)) that dates to around AD 800, on folio 1v of the 9th-century Macdurnan Gospels (London, Lambeth Palace MS 1370), and on folio 329v of the Vivian Bible (Paris, Bibliothèque Nationale, lat. 1), where it frames, or in itself represents, the cosmological or apocalyptic Christ (O’Reilly 1998, 77–8, pl. 3; 87–9, pl. 5). It was also used to illustrate 11th-century computistical manuals. One such was compiled by Byrhtferth, a monk at Ramsey, Cambridgeshire, around the year AD 1011. It includes a diagram: … that sets out the harmony of the macrocosm and the microcosm through correlating various quaternities of time, space, and matter: the four seasons, together with their related months and parts of the Zodiac; the four cardinal directions and winds; the four elements and their properties; the four humours and ages of man. (O’Reilly 1998, 83, pl. 4; Oxford, St John’s College Library MS 17 f. 7v)

It can be seen, therefore, that the lozenge, in referring to quaternities in time, space and matter, ultimately refers back to the cross, thus providing a focus for ecclesiastical meditation on Christ as the centre of Creation, and that the triquetrae in the terminals of the cross emphasise this by referring to the three persons of the Trinity, in other words, that Christ was the Son of God who was present at the beginning of time, was on Earth in the flesh and in the Spirit in the past and present time, and would come again on the Last Day. Only three armpitted crosses survive on High Island and of these only one, Cross-slab 29, is complemented by the ring or wheel around the head that is so common in other parts of Ireland and Pictland, it became known as the ‘Celtic Cross’ (see ‘Typology’, above). Peter Harbison, noting that Irish High Crosses frequently showed the Crucifixion with a figural Christ within the ring, thought the ring might represent the Cosmos, ‘the most central event in the history of which early Christians saw as being the Crucifixion’ (Harbison 1994, 12). However, it has been argued above that the lozenge, rather than the ring, represented the Cosmos in early medieval symbolism. An alternative hypothesis might be that the ring represents Christ’s halo, as depicted in Late Antique and early medieval Christian art since the 2nd century AD; it was reserved for Christ and Christian saints from the 4th century (Gannon 2003, 34, note 86). In support of this idea, it is notable that Northumbrian figural crosses and manuscript miniatures that depict the Crucifixion show Christ with a halo, but do not ring the heads of sculptured crosses; conversely, Irish figural crosses that are ringed do not show Christ with a halo, although some Irish manuscript miniatures of the Crucifixion show a haloed Christ (Allen and Anderson 1992, 144). Gertrud Schiller suggests that a cross within, or extending beyond, a halo is used to represent the persons of the Holy Trinity, especially Jesus (Schiller 1972, 135, figs. 150–153, 346–354). The halo itself may have arisen out of the conception of God as being the source of light. One of the main Christological debates of the 4th century, Arianism, centred around whether Christ was of one person with God the Father, or, as Son of God, had been created by God and was both distinct from and inferior to God. An argument used by those opposed to Arianism was that the relation of a ray of light to the source of that light showed that distinct forms could emanate from a common substance; thus, Christ could also be God (Schapiro 1980, 117).

Functions

Some High Island slabs discussed above have shapes that announce their function, for example the long recumbent slabs (nos. 3, 6 and 14) that were found covering graves at the east end of the church. In the excavations of the

198 high island: excavation of an early medieval monastery

1990s, triangular slabs were found used both as headstones (Cross-slab 9) and as footstones (Cross-slabs 4, 11, 12 and 13). Other headstones were cruciform (Cross-slabs 1–2, 5) or rectangular (Cross-slab 8). Only one sidestone dividing cist graves was found and it is, perforce, rectangular (Cross-slab 10). Round stones, however, seem likely to have had several ritual functions. Three water-worn round stones at High Island (nos. 24, 31 and 35) are incised with Greek crosses. A Greek cross incised in a circle is traditionally understood to symbolise the Host, and was one of the designs impressed on consecrated Mass wafers from Late Antiquity in the Roman Church (Martene 1788, vol. 1, 117). Careful examination of the minute details of folios 34r and 48r of the Book of Kells reveals mice nibbling at similarly marked wafers, showing that it was still the practice of the Irish Church in the 9th century to mark the Host with a cross (Meehan 1994, 45; Lewis 1980, 146–8). However, although the shape and marking of the three High Island cross-marked pebbles would have recalled the Eucharist to an ecclesiastical viewer, the likely function of these stones was not so much to recall the Host as to be used for ritual blessing, cursing and oath-taking. At Inishmurray, Co. Sligo, another early medieval island pilgrimage destination, a large group of rounded waterworn stones incised with Greek crosses are so-called ‘cursing stones’, part of whose purpose may have been to perform ritual oath-taking. Tomás Ó Carragáin notes that on mainland Europe, particularly in the later 10th and 11th centuries, formal liturgical cursing was common; texts from this period include formulae for liturgical clamours or appeals to God for justice (O’Sullivan and Ó Carragáin 2008, 335, citing Little 1993, 19, 44 and 149). The only Irish early medieval literary reference to a Christian cursing stone, that in the Betha Maedoc Ferna, does not describe it as marked in any way. Saint Maedoc, who flourished from the late 6th to early 7th century, is described as having made ‘a binding covenant and union of body and soul’ with Molaise of Devenish, the likely patron saint of Inishmurray (O’Sullivan and Ó Carragáin 2008, 30–35). Under this covenant, any person reproved by either saint would be cursed and excommunicated by both. Maedoc’s cursing stone was placed in or adjacent to the church he built at Killybeg (present-day Killybegs), Co. Donegal, where the ‘erenaghs or tenants of this church’ had the right to turn it anti-clockwise three times to avenge any wrong or injustice done to them. The culprit was expected to die within a year (Plummer 1922, vol. 1, 247–8; vol. 2, 239–41; O’Sullivan and Ó Carragáin 2008, 338–9). Ó Carragáin argues that the Clocha Breaca (speckled or ‘cursing stones’) on Inishmurray provide material evidence that liturgical cursing was becoming formalised at some Irish sites in the 10th and 11th centuries, and that cursing rituals were a necessary clerical response to the increasingly vio-

lent and exploitative nature of the upper hierarchy of Irish society. He cites Giraldus Cambrensis’ conclusion that the perceived vindictiveness of Irish saints was a reflection of their role in keeping the peace in the absence of other provisions for administering social justice in the early middle ages (O’Sullivan and Ó Carragáin 2008, 339, citing Brewer and others 1861–77; Smith 1990, 340). Although the very few ‘cursing stones’ found on High Island would appear to argue against the importance of this function at that site, the stated position of the Killybeg cursing stone is interesting given that no. 24 (and two other similar type stones without cross decoration, see Section 6, Finds Catalogue 4:1, 4:5) was found in accumulated soil over the paving encircling the church. However, this was not the only function of the ‘cursing stone’. As previously explained, the cross was considered to have apotropaic powers; this would have enabled the stones to be used to ward off curses, as prayer stones or, according to Whitley Stokes, to effect miraculous healings, as well as being used for making ritualised oaths and maledictions to protect the ecclesiastical community (Wakeman 1885, 233, citing Stokes 1868, 295). White stones, in particular, were thought to have healing powers; in his Life of Columba, Adomnán describes the saint picking up a white pebble from the river and telling his companions that through the pebble the Lord would bring about the healing of many. Accordingly, cures were effected when the sick drank water into which the stone had been dipped (Anderson and Anderson 1991, 398–405). Thus, nos. 24, 31 and 35 may have served several ritual and medicinal functions. Less likely is the function of marking graves attributed to five cross-carved pebbles from the monastic enclosure or graveyard at Iona, all of which, it should be noted, are significantly larger than nos. 24, 31 and 35 at High Island (RCAHMS 1982, 16, 188–9; Sharpe 1991, 228, 374–5, note 411). Apart from the slabs excavated at the east end of the church in the 1990s (Cross-slabs 1–14) and the slab that formed the south side of the altar (Cross-slab 15), it is not possible to distinguish definitively between slabs that were originally designed to mark boundaries, buildings or burials (although it has been suggested above that Cross-slab 18 was originally a termon or boundary marker). The apotropaic cross, in whatever shape it was incised or carved, was clearly considered equally appropriate for all three functions. It seems probable, however, that the majority of slabs were grave-markers. Even though these slabs were apparently not incised with the names of the departed, remembrance of the dead was important to surviving members, who valued the witness of their predecessors and expected their prayers on their behalf in Paradise; equally, members of communities wished to be remembered in prayer after death (see Maddern 2013, 201–6).

the cross-slabs 199

This is demonstrated very clearly in De Abbatibus, an early 9th-century Northumbrian eulogic poem that records the lives of prominent members of the community at Crayke, in North Yorkshire, a cell of Lindisfarne, including the Irish scribe Ultan. The poem indicates that the presence and influence of former members was still felt and valued after their deaths; the dead Eadfrith, speaking to his living former pupil Æthelwulf in a dream, confirmed the purpose and value of the community, to which he clearly still considered himself committed (Campbell 1967, 62–3, lines 789–94). This view of the community as consisting of both living and dead meant that memory, and therefore mnemonic devices, were very important. Part of the purpose of grave stones that were displayed where the community could see them regularly was to provide a tangible record of members of the ecclesiastic community who had lived a ‘good’ life and died a ‘good’ death, in ascetic terms. Given the variety of forms of cross-carved slabs at High Island, it is possible that each slightly different form was dedicated to a particular individual by which the community could identify those within living memory. Thus, even though the names of those buried were not incised on the stones, the stones themselves recognised and rewarded those who had fulfilled the expectations of the community, members of which were able to recognise individual graves from the motifs carved on the slab. Prayers for the dead would have been prompted even by uninscribed markers since, in a predominantly oral culture, memory was more important in commemoration than the written word (Thomas 1994, 22). They are also likely to have provided a sense of belonging and continuing tradition which perhaps, in some cases, compensated for lack of offspring. To quote Katy Cubitt, death was ‘one of the great mnemonic moments in monastic life’ and ‘physical objects [were] some of the chief vehicles for the transmission of monastic memory’ (Cubitt 2000, 271). Most importantly, however, the cross-carved slabs enabled meditation on death, burial and spiritual and bodily resurrection for judgement on the Last Day.

Conclusions

The High Island cross-carved slabs, even though uninscribed, tell us a great deal about the ecclesiastics who lived there, particularly those of the 9th to early 11th centuries. The first wave of inhabitants appear to have been monks

settled by St Féichín of Fore in the 7th century (Stokes 1892, 4–5; Gwynn and Hadcock 1970, 386). Cross-slabs 1–2, 5, 7 and 16, which are cruciform, and slabs carved with simple linear crosses and which may have been reused, nos. 26–27, 30, 33, 36 and 40, may belong to this period. The 9th- to 11th-century ecclesiastic inhabitants of the island shared the theological concerns of the day, their preference for grave-markers with D-shaped terminals and central expansions suggesting some anxiety about a possibly imminent Last Day. Their holiest dead were celebrated with grave assemblages of considerable theological complexity, showing that the monks responsible for the design of the cross-carved slabs were both scholarly and able to keep abreast of developments at other ecclesiastic communities. It is not clear whether the same community as that supposedly established by St Féichín in the 7th century was still functioning by 1017 or 1018, when Gormgal died. However, according to George Stokes, there continued to be a monastic community on High Island until the 17th century, when, as previously noted, the hagiographer John Colgan acquired the residual documentation of the community (Stokes 1892, 5). White Marshall and Rourke have discussed the difficulties surrounding this documentation insofar as it relates to Gormgal, pointing out that for his facts about the putative saint Colgan relied on a poem written by Gormgal’s contemporary, Corcrán, which has now disappeared (White Marshall and Rourke 2000, 222– 4). However, it is clear that there was an ecclesiastic named Gormgal who was associated with High Island in the 10th or 11th century. It seems likely that, although Gormgal and his companions had apparently retreated from the world, the world that valued their spirituality visited the island as pilgrims and continued to do so long after Gormgal’s death. The grave assemblages at the east end of the church may have formed the chief focus for such pilgrims. Coins found during the 1990s excavations, an 11th- or early 12thcentury penny of possible Scandinavian provenance and an early 13th-century halfpenny of King John (see Section 6.7), bear witness to the presence of overseas visitors so that, although this phase of the island’s history is not well documented, it seems likely that the monastery on High Island was an important destination for pilgrims in the 11th to 13th centuries.

200 high island: excavation of an early medieval monastery

7.1 Catalogue of cross-slabs Ian Fisher (revised by Georgina Scally and Christine Maddern)

Introduction Georgina Scally

Sixty cross-slabs (and fragments thereof) have been found at the early medieval monastery on High Island together with a further six items, such as cross-bases and possible parts of cross-shafts. Of the 60 recorded crosses from High Island, twenty were known prior to the excavation, while 40 were uncovered as a result of the excavation; 45 are decorated, fifteen are undecorated; 47 are complete or largely complete, while thirteen are comprised of fragments. Of the total number of 45 decorated cross-slabs, eleven are decorated on two or more faces, 31 on one face and three are indeterminate. Of the 47 complete or largely complete crosses, 21 are cruciform, of which thirteen are decorated. Of the stones that were uncovered as a result of the excavation, nearly half (nineteen out of 40) were found in the vicinity of the graves east of the church, reinforcing this as the central focus of the ecclesiastical settlement. Combined, the entire corpus comprises one of the largest and best preserved collections of early medieval carved and cruciform slabs from an excavated monastic site in Ireland.

Numbering system Each cross-slab or associated item was allocated a unique Catalogue Number ranging from no. 1 to no. 66. Fiftysix are referred to in the catalogue, in the main text and on the drawings as Cross-slabs 1, 2, 3, etc. while four are described as Crosses (24, 31, 35, 46) and six associated items are referred to by their catalogue number (51, 52, 53, 56, 57, 58). Dimensions: all measurements quoted refer to maximum identifiable size.

Numbering of graves Nearly half (nineteen out of 40) of the cross-slabs uncovered were found in the vicinity of the graves outside the east wall of the church. Eight graves were identified in this location. The graves have been numbered 1 to 8 in order of layout, south to north. Due to the fact that the graves belong to at least two, and possibly three, phases of activity, this layout does not reflect a chronological sequence. For a detailed description of each grave, see Section 5.2.5, and for discussion of the suggested chronology, see Section 10.

Glossary Arms: the horizontal limbs of a cross. Background: the recessed area of a carved stone, which has been cut back to a lower level than the rest of the stone. Band: two incised parallel lines forming a simple or more complex shape, as in knotwork or interlace. (The descriptive term ‘band’ is synonymous with ‘ribbon’; ‘band’ is used throughout this report in favour of ‘ribbon’.) Boss: a circular area of stone raised or standing out from the rest of the stone surface, usually standing on its own or with a number of other bosses. Chevron: V-shaped incision, which may be single or nested. (Nested means incised inside one or more other chevrons.) Crosslet: small horizontal or vertical lines intersecting the limbs of a cross near the terminal to form four smaller crosses on the larger cross. Cruciform: when a stone has been shaped into a cross. Double band: four incised parallel lines forming an outline, sometimes splitting into two single bands to form a feature. (The term ‘double band’ is synonymous with ‘double ribbon’; ‘double band’ is used throughout this report in favour of ‘double-ribbon’.) Flared: when an arm or shaft of a cross expands progressively outwards for its entire length. Fork: a single line divided into two at the end; commonly used as a terminal to the limb of a linear cross. Fretwork: decorative carved work comprising intersecting straight lines in a frame. Greek cross: a cross whose four limbs are of equal length. Incised: a line forming a groove in the surface of the stone, usually made by cutting or, less frequently, by friction. Latin cross: a cross whose shaft is of greater length than the arms. Linear cross: a cross formed when a single line or groove forms the shaft and a single line or groove forms the arms. Nested forks: two or more ‘V’ shapes at the end of a limb of a linear cross. Outline cross: a cross formed by an area of stone recessed or cut back (incised) or raised (in relief) from the surrounding area of stone. Outlined cross: an outline cross (see above) enclosed within an incised linear outline. Outlined Linear cross: a linear cross (see above) set within another outline. Orans: a human figure with arms extended in a gesture of prayer. Pellets/pelleted: small, ball-shaped motifs with rounded or flattened heads. Pelta/peltae: decorative motif based on a small Roman shield, common in Roman and Celtic art, consisting of a ‘C’ shape terminating in two inward-facing spirals, sometimes joined by a V-shaped ligature.

the cross-slabs 201

Quadrant: one of the four spaces formed between the limbs of a cross and a border or edge of a slab. Roundel: a circle or round shape. Shaft: the vertical limb of a cross. Termon: ecclesiastical boundary-marker delineating area of sanctuary and exempt from secular charges. Tricuspid: having three cusps or points. Trilobate: having three lobes or rounded shapes on a shaft or stem. Triquetra: a knot consisting of three interlacing members. Wedged terminals: incised triangular shapes appended to the ends of the limbs of a cross. (Note: some crosses are described in height (H), width (W) and thickness (T), while others are described by length (L). Where it is thought that the stone was upright, H has been used; where it is thought that the stone was laid horizontally, L has been used.)

Cross-slab 1 (Fig. 7.1) Provenance: footstone for Grave 1. Dimensions (max.): H: 0.79m x W: 0.31m x T: 20–30mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces. Archaeological context: Grave 1 contained remains of a burial (F41) dated by radiocarbon analysis to cal. AD 1033–1169 (UB-4156, 913±19 BP). The remains were interred after the church had been constructed. Probably reused. Description: pillar of garnet mica-schist, slightly tapered to a rounded base and expanded with convex armpits in the upper part to form a minimal cross with squarish head. There is considerable weathering, especially in the shaft, both faces of which appear to have been undecorated. The surface of the upper part of Face A (east-facing) bears irregular notches which define the head of an outline Latin cross. Its centre is heavily flaked, but two concentric circles define a roundel 160mm in diameter, which encloses an equal-armed linear cross 60mm in width. The three upper limbs terminate in forks. In the upper armpits of the Latin cross there are irregularly placed bosses, the right one possibly paired and the left one apparently larger, but all heavily flaked, as is the top arm of the cross. On the concealed Face B (west-facing) a similar linear cross (W: 65mm) is set slightly obliquely on the summit of a three-level boss. The boss measures 260mm in diameter at the base and 140mm at the apex, and is set on a semicircular platform extending into the shaft. Once again there appears to be a pair of small bosses (or possibly a pelta) in the upper left angle, backing on to those on Face A, while the bosses (or pelta) in the right angle

were probably similar, but are now too damaged to be identifiable.

Cross-slab 2 (Fig. 7.1) Provenance: headstone for Grave 2 found in situ within a recess (F51) in the church wall. Dimensions (max.): H: 0.79m x W: 0.28m x T: 30mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces. Archaeological context: no human remains were uncovered in this grave and it is probable that the headstone marked a pseudo grave. Probably reused. Description: cruciform stone of very weathered garnet mica-schist with swollen shaft and short side-arms. Face A (east-facing) is closely related in its ornament to Crossslab 1, although there is no outline cross. At the centre of the cross-head two concentric circles define a worn roundel 170mm in diameter, within which are traces of an equal-armed cross with a combination of straight and slightly forked terminals. In the top left quadrant there are remains of a linked pair of bosses, a motif that was probably repeated in the damaged top right quadrant. The only ornament on the concealed Face B (west-facing) is a linear cross in the cross-head, 110mm x 130mm in width with forked terminals up to 50mm long.

Cross-slab 3 (Fig. 7.1) Provenance: recumbent on Grave 2. Dimensions (max.): L: 1.19m x W: 0.59m (incomplete) x T: 40mm. Geological ID: garnet mica-schist. Remarks: fractured. Archaeological context: no human remains were uncovered in Grave 2 and it is likely that the recumbent slab marked a pseudo grave. Description: irregular and roughly tapered slab of garnet mica-schist, broken into three pieces and lacking the upper right edge. Vertically at the centre in low relief there is a double-band, 80mm in combined width. Above and below it splits to enclose central roundels, each linked by peripheral bands to three large, D-shaped terminals. There is an incised arc at the top left angle. The upper roundel, 140mm in diameter, is filled with two back-toback Ds with similar infilling, linked by two horizontal bands. The upper terminal, which is angled to match the oblique upper edge of the slab, has a fretwork filling. The right terminal, which is incomplete and worn, was probably similar, but the ornament of the left terminal is effaced except for the enclosing ribbon. The lower roundel appears

202 high island: excavation of an early medieval monastery

250mm

6 5A

Fig. 7.1 Cross-slabs 1–6.

Fig. 7.1

the cross-slabs 203

to have been similar to the upper one, but due to the narrowness of the lower half of the slab, its terminals, which preserve no ornament, are set tangentially to it.

Cross-slab 4 (Fig. 7.1) Provenance: footstone for Grave 2. Dimensions (max.): H: 0.67m x W: 0.39m x T: 80mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces. Archaeological context: no human remains were uncovered in Grave 2 and it is probable that the footstone marked a pseudo grave. It is suggested that this grave-marker may formerly have been the footstone for the pre-church phase of Grave 3 and when the church was built and the graves restructured, the slab was moved a short distance to the south to function as the footstone for pseudo Grave 2 (see Fig. 10.3 (a)–(d)). Reused. Description: tapered slab of garnet mica-schist with a small base-tenon. The tenon measures H: 110mm, W: 130mm, T: 40mm, being equally recessed from both faces. Face A (east-facing) bears a Latin cross 0.29m high formed by a continuous band in low-relief. It has large forks on both upper and side limbs, that on the right bearing a crosslet. The shaft rises from a slightly off-horizontal bar with crosslet terminals outlined so as to form a tricuspid outline. Face B (west-facing) is incised with a Greek cross, the limbs being 0.25m to the point where forks form terminals. As on Face A, the outline is formed by a continuous band. In the armpits there are four 50–70mm bosses with hollowed centres, and at the upper angles of the slab there are damaged bosses, the one on the right also having a hollowed centre. In relief below the cross there is a T-shaped outline, H: 0.14m, with swollen shaft and oval head, possibly a poor attempt at a triquetra.

Cross-slab 5 (Fig. 7.1) Provenance: headstone for Grave 3 Dimensions (max.): H: 0.94m x W: 0.39m. Geological ID: garnet mica-schist. Remarks: decorated on both faces. Archaeological context: Cross-slab 5 may originally have marked the pre-church phase of Grave 3, but when the church was built it is suggested that the headstone was moved a short distance to the north, so that it was centered at the west end of the recumbent slab (Cross-slab 6) for Grave 3. Grave 3 contained human remains (F70) radiocarbon dated to cal. AD 980–1025 (UB-4266, 1023±23 BP). Description: short-armed cruciform slab of garnet micaschist, damaged in the top arm and right-hand side edge

of the cross-head on Face A (east-facing). It measures H: 0.94m x W: 0.39m, the shaft being slightly tapered. The armpits defining the cruciform outline are slightly notched, and curving arcs enclose the lower ones on Face A. In the cross-head of this face, three incised circles form a double roundel, 0.22m in overall diameter. The slab is too weathered to determine whether the roundel originally enclosed a cross, as seems probable. In the head of the concealed Face B (west-facing), shallow wedge-shaped recesses define a low-relief cross with expanded arms and sunken armpits. These cut into the outer ring of a double roundel, 0.16m in overall diameter, enclosing a linear Greek cross superimposed on an outline Greek cross with widely flaring arms.

Cross-slab 6 (Fig. 7.1) Provenance: recumbent on Grave 3. Dimensions (max.): L: 1.62m x W: 0.37m x T: 100mm. Geological ID: shelly fossiliferous limestone, probably from the Aran Islands, Co. Galway. Remarks: one of two limestone cross-slabs from High Island, the other one being the footstone (Cross-slab 7) for the same grave. Archaeological context: Grave 3 was a dual-phased grave, with a pre-church phase and a restructured phase contemporary with the church. Cross-slab 6 is suggested to be contemporary with the restructured phase. The grave contained human remains (F70) dated by radiocarbon analysis to cal. AD 980–1025 (UB-4266, 1023±23 BP) Description: rectangular slab of shelly, fossiliferous limestone, much weathered. An incised border 30mm from the edge of the slab is preserved in places. A double band in low relief, which runs the length of the slab, is attached at the top to a central roundel and three D-shaped terminals. The roundel and the top terminal retain traces of fret-ornament, and the latter preserves part of an external boss with a central hollow at the right angle. Below the central roundel, in the upper half of the slab, there is a further pair of terminals, apparently without a central feature, possibly as a result of flaking. In the lower half of the stone the pattern of a central roundel and D-shaped terminals appears to have been repeated in mirror-image, but in this area only faint traces of carving are discernible.

Cross-slab 7 (Fig. 7.2) Provenance: footstone for Grave 3. Dimensions (max.): H: 0.89m x W: 0.32m x T: 90mm. Geological ID: shelly fossiliferous limestone, probably from the Aran Islands, Co. Galway.

204 high island: excavation of an early medieval monastery

Remarks: one of two limestone crosses from High Island, the other one being the recumbent slab for the same grave (Cross-slab 6). This was the only footstone decorated on its west face only. All but the upper c. 0.35m of this stone was concealed within the grave. Broken in two pieces. Archaeological context: Grave 3 contained human remains (F70) dated by radiocarbon analysis to cal. AD 980–1024 (UB-4266, 1023±23 BP). Description: cruciform stone of fossiliferous limestone, similar to Cross-slab 6. The head has broken off, probably when the east gable of the church collapsed, and the base was rough, suggesting that it may have been incomplete. On Face A (west-facing) the head is carved in low relief and the shaft bears an incised figure, probably secondary. The lower armpits of the cross-head are notched; the upper ones are open to flank the top arm. The cross-head has a border roughly 25–30mm wide, flaked off in places. This encloses a low-relief cross with a double roundel, the outer being 110mm in diameter within concentric outer arcs. The much-weathered roundel encloses fretwork with multiple horizontal bars. Despite the break at the top of the shaft, the lower part of the broad hollow outlining the cross-head can be traced in this area. At the sides it merges with what were probably secondary incised borders, W: 25–40mm, whose defining grooves return as the base and horseshoe-shaped head of the frame enclosing a standing figure 0.36m high. This figure, with oversized triangularshaped head and traces of simply-incised features, wears a long robe and stands in a curved framework that may represent a boat with a raised prow or some other structural feature (a small incision at the central lower level of the figure may represent the frontal view of a bow). The figure stands frontally in an orans position with U-bent arms, holding in each hand a rectangular object, possibly a book. White Marshall and Rourke (2000, 149) suggest a cross, one of them encircled, can be seen on each of these objects, but these possible motifs are unclear. The motif on the left rectangle appears to show four fingers of the figure’s open right hand. On the other rectangle there are two horizontal bars, but no visible vertical one. Face B (east-facing) is very roughly dressed and bears no discernible decoration.

Cross-slab 8 (Figs 7.2, 7.12) Provenance: headstone for Grave 4. Dimensions (max.): H: 0.91m x W: 0.73m x T: 140mm. Geological ID: blue grey mica-schist. Remarks: this was the only headstone decorated on one (east) face only. The stone was drawn by William Wakeman in 1839 (Slab 12 RIA MS 12T9, 43).

Archaeological context: Grave 4 was a dual-phased grave, with a pre-church phase and a restructured phase, contemporary with the church; Cross-slab 8 is most probably contemporary with the latter phase. The grave contained human remains (F40) dated by radiocarbon analysis to cal. AD 980–1023 (UB-4155, 1027±19 BP). Description: almost rectangular slab of blue-grey micaschist with quartz veins, flaked in the upper part and damaged along the right edge. Almost the entire area of the stone is filled with an expansional cross in relief up to 15mm high, whose terminals extend to the edges of the slab except at the base, where there is a 0.15m margin. This indicates that it could not have stood without support. A bold raised band, divided in the short arms, frames the central roundel and large D-shaped terminals. The roundel is filled with fretwork. The D-shaped terminals are also filled with fretwork, the most salient features of which are central vertical divisions. There are four bosses in the quadrants of the cross, and in the upper angles there are pairs of bosses linked by C-scrolls or peltae, now much damaged, which extended to the edges of the slab.

Cross-slab 9 (Fig. 7.2) Provenance: headstone for Grave 5. Dimensions (max.): H: 0.68m x W: 0.52m x T: 150mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces and on top. Archaeological context: Grave 5 was a dual-phased grave, with a pre-church phase and a restructured phase, contemporary with the church. The headstone was contemporary with the earliest phase. The grave contained the earliest excavated human remains (F37) from the site, dated by radiocarbon analysis to cal. AD 881–977 (UB-3999 1126 ±22 BP). Description: triangular slab of garnet mica-schist, heavily flaked in places. On Face A (east-facing) a Latin cross in relief has a plain tapering shaft, dividing into a broad band that frames a central roundel. D-shaped terminals extend to the edges of the slab. The roundel contains a Greek outline cross with expanded limbs terminating in convex curves, over which is incised a linear Greek cross with forked terminals. The terminals contain worn infill. The upper angles of the quadrants were originally filled with C-scrolls or peltae with bossed terminals, similar to those on the adjacent headstone (no. 8). Strips of meander fretwork of varying widths flank the shaft, conforming to the taper of the slab. The reverse side (Face B, west-facing) and concealed top edge are incised with linear crosses with widely forked terminals, the latter compressed on one axis to match the thickness of the slab.

the cross-slabs 205

8 7

250mm

Fig. 7.2 Cross-slabs 7–11 (for no. 8, see Fig. 7.12).

Fig. 7.2

206 high island: excavation of an early medieval monastery

Cross-slab 10 (Fig. 7.2) Provenance: sidestone between Grave 5 and pseudo Grave 6. Dimensions (max.): H: 0.54m x W: 0.76m x T: 90mm. Geological ID: mica-schist. Remarks: incomplete. Archaeological context: decorated side facing north and entirely concealed. Reused. Description: in the upper half of the slab there is a large, linear Latin cross centered in a linear circle beyond which the arms of the cross project for various lengths before ending in wide, curving, forked terminals. The original upper terminal once resembled the other three, but an acute-angled double forked terminal has been superimposed on it. This is probably a secondary feature.

Cross-slab 11 (Fig. 7.2) Provenance: footstone for Grave 5. Dimensions (max.): H: 0.62m x W: 0.39m x T: 60mm. Geological ID: garnet mica-schist. Remarks: decoration on east face only. Archaeological context: found in situ at the east end of Grave 5. The grave contained the earliest excavated human remains (F37) from the site, dated by radiocarbon analysis to cal. AD 881–977 (UB-3999 1126 ±22 BP). Probably reused. Description: tapered slab of garnet mica-schist, chipped at the top and broken obliquely at the foot but intact at the sides. Face A (east-facing) bears an expansional cross, executed in relief against a sunken, tooled background. It is cruder in design than other examples described above, and only the shaft is more than a few millimetres long. The central roundel and roughly D-shaped terminals are partly framed by bold double bands, but these are not continued at the edges of the slab and the foot of the lowest terminal. A raised central roundel, 75mm in diameter within a 20mm band, is notched at the angles to form an equal-armed cross with expanded arms on which a linear cross is superimposed. Each of the side-terminals contains four horizontal finger-like bars, while the lower and damaged upper terminals are divided vertically by single grooves. The upper terminal is filled with a roughly triangular shape bisected by a curved central incision that falls short of both the upper and lower edges of the triangle, giving the impression that it was intended to represent a nose in a very roughly sketched face. The lower D-shaped terminal is also roughly triangular and is divided by a vertical incision, giving the impression of crudely sketched feet placed side by side. This anthropomorphism is also

reflected in the D-shaped side-terminals, which are incised with horizontal lines to give a rough impression of hands. This motif is discussed in Section 7, ‘Iconography’.

Cross-slab 12 (Fig. 7.3) Provenance: footstone for Grave 6. Dimensions (max.): H: 0.66m x W: 0.36m x T: 90mm. Geological ID: garnet mica-schist. Remarks: decoration on east face only. Archaeological context: found in situ at the east end of pseudo Grave 6. Despite the grave being adorned by grave-markers (originally there would have been a headstone within the niche (F63) above the grave within the east wall of the church), the grave had never been dug for burial and no human remains were recovered. Probably reused. Description: tapered slab of garnet mica-schist, broken at the top right corner and flaked in places. On Face A (eastfacing), the lower part of the right edge appears to have been chamfered to enable insertion into a socket. Filling the upper part there is a boldly incised linear cross, H: 235mm x 210mm across, contained within a centrally placed, ovoid shape defined by two incised grooves 30mm apart. At the junctions with the outer ring the cross-arms fork as oblique terminals, the upper one largely flaked off. In the upper quadrants there are 70mm incised roundels filled with linear Greek crosses. In the lower quadrants 40mm incised roundels, the one to the left bearing vestiges of a linear Greek cross, flank the lower terminal.

Cross-slab 13 (Fig. 7.3) Provenance: footstone for Grave 7. Dimensions (max.): H: 0.59m x W: 0.32m x T: 70mm. Geological ID: garnet mica-schist. Remarks: decoration on east face only. Archaeological context: found in situ at east end of pseudo Grave 7. The area of the grave had never been dug in preparation for burial and no human remains were recovered. Probably reused. Description: tapered slab of garnet mica-schist, broken and flaked at the foot. Face A (east-facing) bears a double-ringed linear cross very similar to that on Cross-slab 12, but without the incised roundels in the quadrants. The cross measures H: 220mm x W: 185mm to the junctions with the outer of the ovoid shapes, where there are short, forked terminals that extend nearly to the edges of the slab at the head and sides.

the cross-slabs 207

13 12A

250mm

Fig. 7.3 Cross-slabs 12–18 (for no. 17, see Fig. 7.12).

Fig. 7.3

208â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Cross-slab 14

Cross-slab 16

(Fig. 7.3) Provenance: recumbent on Grave 8. Dimensions (max.): visible for L: 1.18m x W: 0.63m x T: 60mm. Geological ID: mica-schist. Remarks: west end of stone is sealed by the church wall. Archaeological context: approximately one-third of the cross-slab was sealed by the church wall; it remains unknown whether the grave contains human remains. Description: rectangular slab of schist, heavily flaked. The right edge is roughly bevelled. A central shaft is formed by a double band 140mm in overall width, which splits to enclose a 175mm diameter roundel at its base. A broad band forms a D-shaped terminal abutting the roundel at the top left edge. There are faint traces of carving to the right and above, suggesting the pattern may have been repeated, but the slab is too heavily worn to identify any such carving. It has been suggested that this stone may be inverted with its head to the east and its foot concealed by the church wall (I. Fisher, pers. comm.).

(Fig. 7.3) Provenance: reused as inner lintel of church doorway. Dimensions (max.): L: 0.93m x W: 0.44m x T: 140mm. Geological ID: garnet mica-schist. Remarks: traces of plaster adhere to fractured edge in church interior, suggesting stone had been damaged prior to reuse in this location. Archaeological context: probably inserted as a lintel when the roof of the church was raised. Reused. Description: cruciform stone of garnet mica-schist, much weathered and partly concealed by the sides of the doorway. Boldly cut notches define the lower armpits of a cross-head, which is of the same width as the shaft, and a damaged rounded top arm or finial is partly visible. (Note: there is no evidence of the upper notches shown in Herity 1990a, fig.39b.) There are traces of an irregularly incised margin in the shaft and enclosing the armpits. Two incised, slightly irregular circles are visible in the cross-head and the top of the shaft, the former possibly bearing a linear cross with widely forked terminals.

Cross-slab 15

Cross-slab 17

(Fig. 7.3) Provenance: built into south side of altar (F209) within the church. Dimensions (max.): H: 0.76m x W: 0.82m x T: 120mm. Geological ID: garnet mica-schist. Remarks: traces of plaster adhere to its exterior (decorated) face. Archaeological context: probable late addition to the altar. Reused. Description: irregular slab of garnet mica-schist, damaged in the upper part and the right edge. Small areas of plaster adhere to the surface, suggesting that the slab was once concealed. The surface is filled with an expansional cross whose broad framing band is continued along the foot and the oblique right edge of the slab, enclosing tooled fields from which the cross rises in relief. At the centre of the frame, within a raised band, there is a relief roundel 110mm in diameter, notched to form a cross with expanded and roundended arms, on which is obliquely superimposed a linear cross with bar terminals. The lower and right terminals at the edge of the slab are filled with symmetrical, D-shaped fretwork with central divisions. A flange of plain stone projects beyond the right terminal. The top and left terminals are too damaged for any ornament to be identifiable.

(Figs 7.3, 7.12) Provenance: set into inner east face of enclosure wall. Dimensions (max.): H: 0.62m (incomplete) x W: 0.56m x T: 75mm. Geological ID: garnet mica-schist with quartz veins. Remarks: drawn by Wakeman 1839 (Slab 13 RIA MS 12T9, 43). Archaeological context: probably inserted during rebuilding of the church enclosure when the north-east passage (F414/F8040) was built. Probably reused. Description: incomplete cross-slab of garnet mica-schist with quartz veins, lacking the upper right quadrant. A substantial piece of the right edge is detached but remains in position. Wakeman recorded its height in 1839 as 0.76m, and this is verified by measurement of identifiable features on his sketch, but even this may not be the original height since he appears to show a broken top edge. The slab bears a boldly incised outline Latin cross with deep, roughly circular armpits up to 60mm wide. The lateral limbs originally extended to the edges of the slab. The slab now lacks the right and top arms, but Wakemanâ&#x20AC;&#x2122;s sketch suggests that the latter was only a little shorter than the shaft, which measures 180mm in height.

the cross-slabs 209

19A

19B

20A

250mm

20B

21A

21B

Fig. 7.4 Cross-slabs 19–21 (for nos. 20, 21, see also Fig. 7.12).

Fig. 7.4

210 high island: excavation of an early medieval monastery

Cross-slab 18

Archaeological context: unknown.

(Fig. 7.3)

Description: much worn stone on all sides and broken in a curve along right-hand side of Face A (south-facing). Face A bears a short-armed expansional cross in low relief, framed by a flat band up to 30mm in width. The central roundel, 140mm in diameter, shows traces of an inner circle, and the dividing groove of the top arm continues down as a vertical division in the upper half of the roundel. The lower and left terminals enclose D-shaped bands in which Macalister identified fretwork, but little detail remains. The left terminal extends slightly beyond the edge of the slab and the lower junction is marked by a slight offset and the upper one by a notch, both continued as grooves across the edge. The right terminal is broken off and the top one damaged, no detail being visible. In the upper left quadrant there is a 50mm boss; Wakeman’s sketch of 1839 shows a similar one in the right quadrant, although the terminal was already lost. A curved raised area survives in the upper left angle, possibly with a boss in the lower part. Face B (north-facing) bears a much-worn linear cross, H: 250mm x W: 215mm, with forked terminals of varying length.

Provenance: inner lintel of Cell B. Dimensions (max.): L: 2.1m x W: 0.34m x T: 270mm. Geological ID: garnet mica-schist. Remarks: cracked in situ. Archaeological context: may originally have been an upright pillar stone marking the location of the early ecclesiastical settlement. Reused. Description: pillar of garnet mica-schist, sheared across near one end and reused as the innermost lintel in the entrance-passage of Cell B and extending along the east wall of the cell at both ends. The exposed area of the lower face bears an incised linear cross with short crosslets at the ends of each limb, 0.26m in width and over 0.46m in length, since one end of the shaft is concealed by adjacent masonry.

Cross-slab 19 (Fig. 7.4) Provenance: in situ at leacht on north-east side of lake, where it was placed by G.H. Kinahan in 1869 (Kinahan 1868–9, 554–5). Dimensions (max.): H: 1.05m x W: 0.52m (incomplete) x T: 140mm. Geological ID: garnet mica-schist. Remarks: one of two stones found in debris (and placed upright) by G.H. Kinahan in 1869. Archaeological context: unknown. Description: leaning against leacht at north-east shore of lake, about 20m south-east of entrance of enclosure, where it was placed by Kinahan in 1869. Irregular cruciform stone of garnet mica-schist, with much weathered surface. The right arm on Face A (east-facing) is much sheared. The cross-head is formed by deep notches with inner hollows at the lower N and upper S angles, and surface grooves and hollows in the other armpits.

Cross-slab 20 (Figs 7.4, 7.12) Provenance: in situ 6.5m south of church, where it had been placed by G.H. Kinahan in 1869 (Kinahan 1868–9, 554–5). Dimensions (max.): H: 1.1m x W: 0.39m x T: 100mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces. One of two stones found in debris by G.H. Kinahan in 1869. Drawn by W.F. Wakeman in 1839 (RIA MS 12T9, 44). The slab was also drawn by R.A.S. Macalister in 1896 (Macalister 1896, 199. Cross D).

Cross-slab 21 (Figs 7.4, 7.12) Provenance: in situ at Brian Boru’s well. Dimensions (max.): H: 0.73m x W: 0.54m x T: 70mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces. Archaeological context: the stone was drawn in 1839 by W.F. Wakeman, who wrote beside it ‘At the church on High Island’. Thirty years later G.H. Kinahan also drew the stone and referred to it being at the well (Kinahan 1868–9, 348). The stone must therefore have been placed at the well some time between 1839 and 1869. One face (Face B) was also drawn by R.A.S. Macalister in 1896 (Macalister 1896, 206. Cross B). Description: tapered slab of garnet mica-schist, obliquely broken at the foot. The ornament on Face A (north-facing) is developed around a linear cross, H: 250mm x W: 240mm, with an irregular ring up to 150mm across and with irregular forked terminals. These extend to the sides and top of the slab. The lower terminal is also open. The outlines of the ring and terminals are repeated by firm grooves forming bands 35–40mm wide, giving the appearance of an outline cross. The upper left angle is defined by a curving groove enclosing a hollow or weathered boss, and the same motif was probably carved in the right angle. On Face B (south-facing) there is a linear cross 180mm in height and width, with wide, curving, forked terminals, the lower one having straighter and smaller forks than the others.

the cross-slabs 211

24A

22A

22B

22C 24B

50mm

23 25 0

250mm

27 29 0

250mm

28A

Fig. 7.5 Cross-slabs 22–30.

28B

Fig. 7.5

212â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Cross-slab 22 (Fig. 7.5) Provenance: in situ leaning against boulder near the south landing. Dimensions (max.): H: 0.72m x W: 0.53m x T: 70mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces and one side. Drawn by G.H. Kinahan in 1869 (Kinahan 1869, 554) and one side (Face A) was also drawn by R.A.S. Macalister in 1896 (Macalister 1986, 206. Cross A). Archaeological context: unknown; probably associated with pilgrim station located close to the (south) landing place. Description: short-armed cruciform cross-slab of garnet mica-schist. The lowest part is reduced in thickness and notched at one edge, presumably to fit a socket-stone. Face A (exposed side) bears a Latin cross in high relief against a pocked background, which continues round the edges above and below the short arm-projections. The better-preserved left arm is defined by an offset below and a notch above. The cross is framed by a flat band up to 40mm in width, enclosing a narrower banded cross. Although now too flaked to see, it is clear from Macalisterâ&#x20AC;&#x2122;s drawing of 1896 that the bands originally crossed at four points to form a central lozenge enclosing an incised design (Macalister 1896, 206). The top terminal contains a triquetra, but any ornament that may have been carved in the other terminals and the lozenge is indistinct. In the upper left quadrant there is a boss or raised area, attached by a spiral to the angle of the upper limb of the cross. Face B (inner south-facing side) bears a much-worn linear cross with a double roundel (as on Cross-slab 13), but with crosslets rather than forked terminals. The cross measures 0.3m x 0.29m and the roundels are 0.13m and 0.22m in diameter. The flat end-surface of the surviving left arm, unworked in contrast to the tooled surfaces above and below, bears a bold linear cross with crosslet terminals (H: 145mm x W: 100mm).

Cross-slab 23 (Fig. 7.5) Provenance: basal stone of altar (F209) within church. Dimensions (max.): H: 0.67m x W: 0.53m x T: 90mm. Geological ID: garnet mica-schist. Remarks: decoration is deeply incised and well preserved. Archaeological context: found face-down within lowest course of stone in the altar (F209). It is possible that the stone may originally have stood upright within one of the two pre-church slotted features. Reused. Description: rectilinear slab of garnet mica-schist, broken

across obliquely at the left foot and heavily flaked in places, including the top and lower right. The whole extent of one face is filled with a linear cross resembling that on Face B of Cross-slab 21. The linear cross is defined by bold, round-bottomed grooves up to 30mm in width and 15mm in depth. The cross measures H: 0.34m x W: 0.22m, with the arms 20mm above mid-height, within wide, curved, forked terminals. The upper three of these extend to the edges of the slab, however the lower one is so damaged by flaking it is not possible to say if it was ever carved.

Cross no. 24 (Fig. 7.5) Provenance: found within soil and stone deposit (F4) over paved surface (F29) around the church. Dimensions (max.): D: 0.82m x T: 0.43mm. Geological ID: ironstone. Remarks: one of three pebbles of similar type, the others being 95E124:4:1 and 95E124:4:5. For discussion on these stones, see Maddern, this volume, Section 7, and Cotter, this volume, Section 6, Fig. 6.4. Archaeological context: placed at the foot of Grave 3, probably by pilgrims visiting the monastery in the period post-abandonment. Description: rounded pebble worn at the perimeter. On both domed faces there are central hollows, a series of concentric circles and numerous scratches. On Face A the inner circle is filled with four nests of three or four chevrons in the armpits of an equal-armed cross with swollen limbs.

Cross-slab 25 (Fig. 7.5) Provenance: from rubble within excavated area (north side) of monastic enclosure wall. Dimensions: H: 0.52m x W: 0.24m x T: 70mm. Geological ID: garnet mica-schist. Remarks: uneven thickness. Archaeological context: unknown, found in rubble. Description: surface flaking on both sides with one edge acutely but probably naturally bevelled. Shallow, curved indentations, no more than 20mm deep, on both sides (the indentation on one side being considerably more pronounced than the other), suggesting the stumpy arms of a Greek cross.

Cross-slab 26 (Fig. 7.5) Provenance: from rubble north-west of Cell B. Dimensions: H: 0.23m (incomplete) x W: 0.2m (incomplete) x T: 45mm.

the cross-slabs â&#x20AC;&#x192; 213

Geological ID: garnet mica-schist. Remarks: fragment. Archaeological context: unknown, found in rubble. Description: fragment with two undamaged edges. It bears an incomplete but firmly incised groove, L: 120mm, which extends to a broken edge of the slab. The groove is 20mm wide and 15mm deep and may have been part of a linear cross, although the slab is too incomplete to be certain. A rounded notch on one damaged edge is probably not artificial.

Cross-slab 27 (Fig. 7.5) Provenance: within a soil and stone deposit (F4) that had accumulated over the paved surface (F29) around the church. Dimensions (max.): H: 0.27m (incomplete) x W: 0.16m (incomplete) x T: 250mm. Geological ID: garnet mica-schist. Remarks: incomplete. Archaeological context: found within a deposit (F4) that had probably accumulated in the period post-abandonment. Description: cruciform stone heavily worn throughout. The short side-arms, incomplete at the ends, have L-shaped or rounded armpits. The wider of the vertical arms appears to have been the upper one, with a projection at the top left. On one face shallow grooves up to 15mm wide define a linear cross with short forked terminals at the foot and side, and a wide semicircular upper terminal. The lower part of the shaft is incomplete.

Cross-slab 28 (Fig. 7.5) Provenance: within a soil and stone deposit (F4) that had accumulated over the paved surface (F29) around the church. Dimensions (max.): H: 0.25m (incomplete) x W: 0.11m (incomplete) x T: 150mm. Geological ID: garnet mica-schist. Remarks: fragment; decoration on both faces. Archaeological context: from post-abandonment deposit. Probably redeposited. Description: irregular pentagonal fragment of garnet mica-schist, broken on most edges but preserving one arm of a cruciform stone. The one partly preserved cross-arm has been tapered, having an obtuse armpit. On Face A there is an equal-armed 40mm cross with forked terminals within a double roundel 140mm in overall diameter, while Face B displays a similar cross within a triple roundel.

Cross-slab 29 (Fig. 7.5) Provenance: stray find (199:7) from rubble south of church enclosure wall. Dimensions (max.): H: 0.55m (incomplete) x W: 0.29m (incomplete). Geological ID: garnet mica-schist. Remarks: incomplete. Archaeological context: unknown, redeposited. Description: disc-headed cross of garnet mica-schist, heavily worn, much broken at the edges and lacking most of upper part of left quadrant. An outline or low-relief Latin cross is superimposed on the disc. In contrast to its muchworn shaft, the head is boldly defined by four D-shaped or crescentic hollows, up to 45mm long and 20mm deep. The lowest curves of the disc-head are preserved at either side of the shaft, but the remainder is so damaged that it is uncertain whether the side-arms of the cross projected beyond the disc, although that seems unlikely. The back of the stone is rounded in section and undecorated.

Cross-slab 30 (Fig. 7.5) Provenance: from hearth floor (F326) within Cell B. Dimensions (max.): L: 0.36m (incomplete) x W: 0.44m (incomplete) x T: 60mm. Geological ID: garnet mica-schist. Remarks: incomplete. Archaeological context: almost centrally placed with decorated face up within hearth floor (F326). Reused. Description: probably rectangular, but broken across for reuse in the hearth and damaged at the edges. On the exposed face there is a linear cross, 145mm high in incomplete height by 125mm in span. Since the shaft was truncated by the transverse break, it is uncertain whether the cross was of Latin or Greek form.

Cross 31 (Fig. 7.6) Provenance: stray find from rubble east of Cell A. Dimensions (max.): D: 135mm x T: 750mm. Geological ID: mica-schist. Remarks: much worn, poorly inscribed cross. Archaeological context: unknown, redeposited. Description: ovoid, water-rounded pebble. One surface bears a worn linear Latin cross (H: 95mm x W: 60mm), defined by very shallow grooves up to 12mm wide and with a curved vertical stem. There are nearly vertical, but very slightly wedged terminals on all four arms of the cross.

214 high island: excavation of an early medieval monastery

Cross-slab 32

Cross-slab 34

(Fig. 7.6) Provenance: from paved floor (F510) within Cell A. Dimensions (max.): L: 0.56m x W: 0.31m x T: 100mm. Geological ID: garnet mica-schist. Remarks: complete, with incised decoration quite well preserved. Archaeological context: one of two decorated stones found within paved floor of Cell A; decorated face found face-up. Reused. Description: triangular slab flaked on the left surface and notched in places at the right edge. One face is filled above the butt with an expansional Latin cross framed by a flat band 20–25mm in width, and continuous except where the upper terminal extends to the top edge of the slab. The split band rises from the lower terminal to diverge round an ovoid centre and into the much worn D-shaped side-terminals on either side, which are tangential to the central oval. The dividing groove of the shaft rises continuously through the ovoid centre and the top arm and terminal to the edge of the slab, and intersects the lateral arms of the cross in the oval. This merges with the surround of an inner D-shaped band in the right terminal, which may have resembled the better-preserved lower terminal. This shows a D-shaped inner band with a central division in the horizontal base. The upper terminal shows only the vertical division described above. The left terminal is fragmentary.

(Fig. 7.6) Provenance: from rubble (F443) within stepped (northwest) passage (F437) through the monastic enclosure wall. Dimensions (max.): H: 0.51m (incomplete) x W: 0.31m (incomplete) x T: 35mm. Geological ID: garnet mica-schist. Remarks: small part missing on shaft and left side. Archaeological context: unknown, redeposited. Description: kite-shaped cross-slab of garnet mica-schist, broken at the foot and left edge, naturally bevelled at the lower right edge, and flaked in places. It bears a Latin cross (H: 0.27m, W: 19m) executed partly in outline and partly in low relief, built around a linear cross with a central roundel. The grooves of the linear cross, which divide the bands framing the outline, and those enclosing the bands themselves, are unusually broad and flat-bottomed. The central roundel, 80mm in diameter, encloses an incised linear Greek cross. The side-arms have D-shaped terminals and the upper and lower ones are forked, but all appear to be open-ended. In the lower quadrants there are flat-bottomed ‘sunken’ linear Latin crosses. That to the left measures H: 105mm x W: 65mm and the right one, set obliquely to the shaft of the main cross to secure more space, is H: 125mm x W: 70mm. These crosses probably represent the crosses of the thieves at the Crucifixion, as on cross-carved slabs at Inishmurray, Co. Sligo (see Maddern, this volume, Section 7).

Cross-slab 33

Cross no. 35

(Fig. 7.6) Provenance: found within paved floor of Cell A. Dimensions (max.): L: 0.54m (incomplete) x W: 0.24m (incomplete) x T: 40mm. Geological ID: garnet mica-schist. Remarks: incomplete. Archaeological context: one of two decorated stones found within paved floor of Cell A; decorated face found face-down. Reused. Description: cruciform stone broken obliquely at the foot and damaged at the right edge as well as being surfaceflaked. The shaft curves slightly and the top upper left is expanded and peaked. The surviving left arm projects 35mm and is 110mm high, and the damaged right arm projects only a few millimeters. In the cross-head, and rising into the top arm, there is a weathered linear Greek cross, 145mm high.

(Fig. 7.6) Provenance: from rubble layer (F474) within north-west area of monastery. Dimensions (max.): D: 1.40m x T: 45mm. Geological ID: quartzite. Remarks: discussed also by Cotter, this volume, Section 6. Archaeological context: unknown, redeposited. Description: circular pebble of quartzite, flattened in section and with slightly rounded faces; flaked in places. It measures D: 130–140mm x T: 45mm. It is naturally shaped except for a peripheral pocked ledge on one face, which frames a raised central area 110mm in diameter. This bears an almost equal-armed cross, 55mm in span, with straight forked terminals formed by round-bottomed grooves.

Cross-slab 36 (Fig. 7.6) Provenance: from rubble (F802) south of Cell B. Dimensions (max.): H: 0.39m (incomplete) x W: 0.27m x T: 45mm.

the cross-slabs 215

100mm

250mm

38 36

Fig. 7.6 Cross-slabs 31–38.

Fig. 7.6

216 high island: excavation of an early medieval monastery

39A

39B

250mm 0

100mm

43 0

Fig. 7.7 Cross-slabs 39–45.

250mm

Fig. 7.7

the cross-slabs 217

Geological ID: garnet mica-schist. Remarks: incomplete. Archaeological context: unknown, redeposited. Description: naturally peaked slab curved in the characteristic bowed section of mica-schist. It is broken across at the foot to leave a thin and irregular flange, and much worn. It bears a linear Latin cross defined by much-weathered shallow grooves up to 15mm wide, the top and left arms being almost untraceable. It measured about 150mm in height, or more if the foot is incomplete at the break, by 90mm wide.

Cross-slab 37 (Fig. 7.6) Provenance: from deposit of silt and grit (F897) over leacht (F887) outside Cell B. Dimensions (max.): H: 0.62m (incomplete) x W: 0.47m x T: 90mm. Geological ID: garnet mica-schist. Remarks: found face-up. Archaeological context: found within a deposit that had formed after the water-collection and drainage channel (F8025) was no longer in use. The stone may originally have stood upright on or close to the leacht (F887) near where it was found. Redeposited. Description: irregular cross-slab with tapered foot, broken at the top left angle and much worn, especially at the foot and right edge. The decorated face bears a circle divided into eight segments by bands forming two interconnected Greek crosses. In the centre is a plain incised linear Greek cross with limbs extending to the borders of the central roundel.

Cross-slab 38 (Fig. 7.6) Provenance: discovered in 1972 by R. Murphy during digging of the drain west of Brian Boru’s well. Dimensions (max.): H: 0.32m (incomplete) x W: 0.32m (incomplete) x T: 60mm. Geological ID: mica-schist. Remarks: incomplete, only cross-head survives. Archaeological context: recovery of this stone in proximity to Brian Boru’s well suggests it may originally have marked the location of the well, but this is not verifiable. Description: head of cruciform stone of mica-schist, broken obliquely at the junction with the shaft and lacking the end of the left arm. The slightly tapered arms, up to 40mm in projection, have asymmetrical L-shaped armpits. The background is cut away to define a low-relief band-cross, whose centre is incised with a 65mm roundel bisected by a linear Greek cross whose limbs project

beyond the roundel. The bands forming the outer cross are three-quarter circles, divided at their ends to form forked terminals extending to the edges of the slab. Each of the sunken quadrants contains a boss measuring about 35mm in diameter.

Cross-slab 39 (Fig. 7.7) Provenance: stray find from rubble (F398) removed from Cell B. Dimensions (max.): H: 0.53m x W: 0.53m x T: 95mm. Geological ID: garnet mica-schist. Remarks: decorated on both faces. Archaeological context: unknown. Description: tapered in the lower part and much worn throughout; sheared obliquely at the right and lower edges of Face B. The side and top edges are notched in line with the angles of the cross-terminals to give a minimal cruciform appearance, although without any surviving projection. Face A is filled with an expansional cross in relief whose D-shaped terminals appear to have adjoined the centre without intervening arms, although no roundel is now discernible. The best-preserved terminal, although incomplete, is the lower one, which is outlined by a double band 55mm wide and may have enclosed a triquetra. The side-terminals, whose angles are rounded at the notched edges of the slab, are heavily worn and the top one is effaced. In the upper quadrants there are bosses, the largest one being 55mm in diameter, and at the top angles there are pairs of 60mm bosses linked by C-scrolls or peltae. The expansional cross on Face B, which preserves only the D-shaped left and top terminals, has straight sections linking the terminals to give a lozenge-shaped central plate. There are some traces of a broad margin, but it is very faint. At the upper angles there are raised panels, curved on the inside, which form lozenge-shaped upper quadrants. There are indications that this pattern was repeated in the lower quadrants, but these have been almost completely sheared off.

Cross-slab 40 (Fig. 7.7) Provenance: unknown, located lying flat on the southwestern side of the church (White Marshall and Rourke 2000, 144). Dimensions (max.): H: 0.51m (incomplete) x W: 0.31m x T: 70mm. Geological ID: garnet mica-schist. Remarks: incomplete, stone is missing. Description is based on White Marshall and Rourke (2000, 144). Archaeological context: unknown. Description: triangular slab, broken or worn at the top.

218â&#x20AC;&#x192; high island: excavation of an early medieval monastery

In the upper part of one face there is a roughly incised equalarmed linear cross within a sub-rectangular frame, open or incomplete at the top.

Cross-slab 41 (Fig. 7.7) Provenance: unknown, located by White Marshall and Rourke lying face-down on the south side of Cell B (White Marshall and Rourke 2000, 144). Dimensions (max.): H: 1.08m x W: 0.43m (incomplete) x T: 120mm. Geological ID: garnet mica-schist. Remarks: incomplete, stone is missing. Description is based on White Marshall and Rourke (2000, 144). Archaeological context: unknown. Description: cruciform slab of garnet mica-schist, flaked in places and with the right arm broken off. It has slightly rounded armpits and the top arm is tapered while the left arm, about 60mm in projection and 0.25m high, is less regular. In the upper part there is an outline cross about 0.55m high, with the arms set just above mid-height. This cross has rounded armpits and wedged terminals. A linear cross with forked terminals is superimposed on the outline cross. In the lower part there is an outer band up to 40mm wide, but this is not recorded as continuing above the ends of the side-arms.

Cross-slab 42 (Fig. 7.7) Provenance: from soil and stone deposit (F4) around foot of Grave 5. Dimensions (max.): H: 0.123m (incomplete) x W: 0.145m (incomplete) x T: 30mm. Geological ID: mica-schist. Remarks: incomplete, fragment. Archaeological context: possibly left by pilgrims visiting the monastery in the period post-abandonment. Probably redeposited. Description: triangular fragment of greenish schist, worked on two edges, both with a distinct bevel. These edges meet at a slightly obtuse angle, and near the broken edge there is part of a right-angled groove that appears to form a broad margin. A long, irregular groove may be secondary or accidental; too little survives for the pattern to be reconstructed.

Cross-slab 43 (Fig. 7.7) Provenance: from rubble (F2) east of church. Dimensions (max.): H: 0.41m x W: 0.16m x T: 35mm.

Geological ID: garnet mica-schist. Remarks: undecorated. Archaeological context: unknown, redeposited. Description: minimal cruciform stone of garnet micaschist, slightly tapered and with a pointed foot which appears to result from natural breaks. Pairs of rounded notches, 0.14m and 0.33m from the top, run along the edges to define a minimal cross on each face.

Cross-slab 44 (Fig. 7.7) Provenance: from rubble (F7) on top of south wall of church enclosure. Dimensions (max.): H: 0.28m (incomplete) x W: 0.18m (incomplete) x T: 22mm. Geological ID: garnet mica-schist. Remarks: incomplete fragment, undecorated. Archaeological context: unknown, redeposited. Description: irregular slab of variable thickness. Surviving edge is tapered with well formed, slightly off-vertical sides. A deep notch, 50mm wide, suggests the stone was a cruciform type cross, but it is too incomplete to reconstruct its form.

Cross-slab 45 (Fig. 7.7) Provenance: from a soil deposit (F215) within the church. Dimensions (max.): L: 0.137m (incomplete) x W: 94mm (incomplete) x T: 28mm. Geological ID: garnet mica-schist. Remarks: incomplete fragment, undecorated. Archaeological context: found with soil (F215) beneath mortar floor (F202) within the church. Redeposited. Description: irregular, notched slab broken across at both ends. The edges bear two notches, not aligned horizontally and one deeper than the other. This appears to form armpits of an incomplete minimal cross.

Cross 46 (Fig. 7.8) Provenance: found within rubble (F1) at north-east corner of church enclosure wall. Dimensions (max.): L: 0.142m x W: 50mm (incomplete) x T: 14mm. Geological ID: garnet mica-schist. Remarks: incomplete, undecorated. Archaeological context: unknown, redeposited. Description: irregular, notched slab broken on several edges. The cross has a constriction formed by a semicircular

the cross-slabs 219

47 46 0 0

250mm 48

100mm

250mm

52 51

Fig. 7.8 Cross-slabs 46–52.

220â&#x20AC;&#x192; high island: excavation of an early medieval monastery

notch in one edge and a wider but damaged one in the other, possibly part of a cruciform or disc-headed stone.

Cross-slab 47 (Fig. 7.8) Provenance: stray find (199:9) from rubble excavated in vicinity of the church. Dimensions (max.): L: 0.4m (incomplete) x W: 0.24m x T: 55mm. Geological ID: garnet mica-schist. Remarks: incomplete and undecorated. Archaeological context: unknown, redeposited. Description: broken at both ends and flaked on both faces. At one edge there is a curved notch 55mm wide and 15mm deep, but the corresponding rebate on the opposite edge appears to be a natural break and it is uncertain whether the slab was designed to be cruciform.

Cross-slab 48 (Fig. 7.8) Provenance: from rubble (F802) south-west side of Cell B. Dimensions (max.): H: 0.32m (incomplete) x W: 0.16m (incomplete) x T: 40mm. Geological ID: garnet mica-schist. Remarks: incomplete and undecorated. Archaeological context: unknown, redeposited. Description: incomplete cruciform stone of coarse garnet mica-schist, split obliquely down the middle and broken at top and foot. The surviving but damaged side-arm projects 60mm on the underside and is at least 135mm high. Below it there are two curves at the edge, formed by worked notches. It has been suggested (I. Fisher, pers. comm.) that the rounded tenon of Catalogue no. 52 (which was found in the socket-slab of Catalogue no. 51) is the missing foot of the slab. If this is correct, the surviving height is 0.4m.

Cross-slab 49 (Fig. 7.8) Provenance: from rubble (F802) south-west of Cell B. Dimensions (max.): H: 0.41m (incomplete) x W: 0.25m (incomplete) x T: 35mm. Geological ID: garnet mica-schist. Remarks: incomplete and undecorated. Archaeological context: unknown, redeposited. Description: incomplete cruciform slab broken down one edge and at both ends. The surviving side-arm projects from 55mm to 70mm and is 150mm high, and its armpits are respectively right-angled and rounded. At the broken edge there is a possible slightly rounded upper armpit, which would make the top of the cross 0.16m in width.

Cross-slab 50 (Fig. 7.8) Provenance: from rubble (F805) south side of Cell B. Dimensions (max.): H: 0.49m (incomplete) x W: 0.12m x T: 45mm. Geological ID: grey-green schist. Remarks: incomplete and undecorated. Archaeological context: unknown, redeposited. Description: cruciform slab with naturally bevelled edges, heavily laminated and damaged, especially in the lower part and in the head. The lower notches of the head are well preserved, the left one being semicircular and the other U-shaped; the upper right notch retains part of a semicircular curve, but flaking of the head has removed most of the worked curve of the fourth notch. Allowing for damage, the top arms appear to have been narrower than the lower arms.

Catalogue no. 51 (Fig. 7.8) Provenance: found in partly dismantled section of southeast corner of church enclosure wall. Dimensions (max.): L: 0.58m x W: 0.47m x T: 70mm. Geological ID: garnet mica-schist. Remarks: cross-slab base-tenon (for Catalogue no. 52) found in situ. Archaeological context: building stone within church enclosure wall possibly inserted when the wall was rebuilt contemporary with construction of Cell B. Reused. Description: socket-slab slightly damaged in places. A socket measuring 150mm long by 40mm wide pierces the slab 120mm from one long edge. On the same central axis and 20mm from the socket the slab is pierced by a circular tapering hole, 110mm in maximum diameter and narrowing to 75mm. It has been suggested (by Scally) that this may have been the upper pivot-hole for a doorway of the type seen in buildings at Bray Head, Valentia Island (Co. Kerry), and at Gallarus Oratory, where the hole is very similar in diameter (A. Hayden, pers. comm.). If this is correct, it is likely that the hole pre-dates the socket as it more centrally placed.

Catalogue no. 52 (Fig. 7.8) Provenance: within east wall of church enclosure. Dimensions (max.): H: 100mm (incomplete) x W: 150mm (incomplete) x T: 25mm. Geological ID: garnet mica-schist. Remarks: fragment of tenon in situ within Catalogue no. 51. Archaeological context: part of tenon broken in situ. Reused.

the cross-slabs 221

54 0

250mm

Fig. 7.9 Cross-slabs 53–57.

222 high island: excavation of an early medieval monastery

Description: fragmentary base-tenon of coarse garnet micaschist, sheared obliquely at the top. It tapers slightly along its short surviving length to a slightly rounded foot. It is suggested (I. Fisher, pers. comm.) that it may be the lower part of the cruciform stone, Cross-slab no. 48.

Cross-slab 53 (Fig. 7.9) Provenance: upright within church, north of entrance. Dimensions (max.): H: 0.97m x W: 0.23m x T: 70mm. Geological ID: garnet mica-schist. Remarks: undecorated. Archaeological context: set upright within pit (F219) cut through paved floor of the church. Possible grave-marker associated with the pseudo Grave 10 (F214) in north-west angle of church. Description: undecorated triangular slab bevelled naturally at one edge and broken obliquely at the top. Tapers continuously along length to an acutely pointed base set in a socket in the paving of the church.

Cross-slab 54 (Fig. 7.9) Provenance: recovered by White Marshall and Rourke from south edge of lake (White Marshall and Rourke 2000, 157). Dimensions (max.): H: 0.2m (incomplete) x W: 83mm x T: 25mm. Geological ID: mica-schist. Remarks: undecorated. Missing. Archaeological context: unknown. Description: miniature cruciform stone or disc-headed cross broken obliquely at the head. The lower side of the head is defined by two armpits, one notched and the other wide and shallow, but no upper armpits are identifiable. Marshall and Rourke suggest that this was a ‘hand cross’ carried by a pilgrim.

Cross-slab 55 (Fig. 7.9) Provenance: found by White Marshall and Rourke lying on rubble outside the western façade of the church (White Marshall and Rourke 2000, 157). Dimensions (max.): H: 0.62m (incomplete) x W: 0.33m x T: 25mm. Geological ID: garnet mica-schist. Remarks: undecorated. Missing. Archaeological context: unknown. Description: L-shaped fragment, probably part of a cruciform stone. Surviving elements are head and upper part

of left arm. A well-rounded notch is visible where the arm intersects with the head. The fragment is reconstructed by White Marshall and Rourke (2000, 119) as belonging to the centre and top arm of a cruciform stone, and part of a side-arm about 0.18m long.

Catalogue no. 56 (Fig. 7.9) Provenance: beside leacht at the south side of the lower (small) lake in proximity to the south-east landing. Dimensions (max.): L: 0.79m x W: 0.78m x T: 100mm. Geological ID: garnet mica-schist. Remarks: undecorated. Archaeological context: cross-base associated with ‘station’ adjacent to the ‘monastery’ landing in the southeastern part of the island. Found in association with Cross fragment (Catalogue no. 57) found in situ. Description: irregular quadrangular socket-slab, damaged at two edges. The socket, set almost centrally, is rectangular with slightly curved sides and rounded angles, measuring 220mm x 110mm and T: 90mm. The incomplete slab (Catalogue no. 57) is now set in the socket. This was probably the ‘stone altar surmounted by a cross’ described in Petrie’s account of his visit in 1820 (Petrie 1845, 423).

Catalogue no. 57 (Fig. 7.9) Provenance: set into cross-base Catalogue no. 56. Dimensions (max.): H: 0.24m x W: 0.54m x T: 80mm. Geological ID: garnet mica-schist. Remarks: incomplete and undecorated. Archaeological context: cross fragment within cross-base (Catalogue no. 56) associated with ‘station’ adjacent to the ‘monastery’ landing-place in the south-eastern part of the island. Description: incomplete cross-slab or cruciform slab sheared at one end and probably split at one long edge. It is T-shaped, with a large transverse bar measuring 0.54m x H: 0.24m and a rounded central arm, tenon or head 0.14m in projection and W: 0.20m. It has been suggested by Herity (1990a, 92) that this was a vertically broken cruciform slab, and the rounded notch in one ‘armpit’ supports this. However, the ‘tenon’ fits the socket well and White Marshall and Rourke’s (2000, 157) identification of this as the lower fragment of a slab or cross is also possible.

the cross-slabs 223

250mm

Fig. 7.10 Cross-slabs 58–62 (nos. 59–62 by W.F. Wakeman, 1839).

224 high island: excavation of an early medieval monastery

Catalogue no. 58

1839 (Slab 14, RIA MS 12T9, 45).

(Fig. 7.10) Provenance: on altar (F209) inside the church. Dimensions (max.): L: 0.49m x W: 0.38m x T: 75mm. Geological ID: garnet mica-schist. Remarks: undecorated. Archaeological context: cross-base, probably that depicted by Wakeman into which a cross (Catalogue no. 59) is set, see Fig. 7.10, Catalogue no 59. Description: tapered quadrangular socket-slab flaked in places and damaged at one angle. The socket opens to the longest edge, a little off-centre, and measures L: 220mm x W: 65mm, tapering to 50mm at the rounded inner end. It has been suggested that this slab was used to hold a door-post (White Marshall and Rourke 2000, 83), but another suggestion by Scally is that it is the slab shown in Wakeman’s drawing of 1839 as the support for a cruciform stone, now missing (infra, no. 59). This identification is supported by the representation in the drawing (Fig. 7.10) of a damaged angle, and the same off-centre position of the socket.

Archaeological context: unknown.

Cross-slab 59 (Fig. 7.10) Provenance: unknown. Dimensions (max.): unknown (see below). Geological ID: unknown. Remarks: undecorated. Stone is missing. Drawing by W.F. Wakeman in 1839 (Slab 15, RIA MS 12T9, 45). Archaeological context: unknown. Description: stone known only from Wakeman’s drawing of 1839, which describes it as ‘south of the church’ (Herity 1990a, 93; White Marshall and Rourke 2000, 158). Cruciform stone whose open-ended socket-stone appears to be the surviving slab (Catalogue no. 58). By comparison with this slab, the visible height may be estimated as about 0.44m and the width as about 0.24m, and the foot of the shaft was not more than 60mm thick. The cross appears to be more carefully shaped than most of the cruciform stones recorded here. It is shown with slightly tapered sidearms and an expanded upper arm with a peaked top. The upper part of the shaft is straight, but most of it expands with a slight curve until the foot fills the base-socket.

Cross-slab 60 (Fig. 7.10) Provenance: unknown. Dimensions (max.): unknown (see below). Geological ID: unknown. Remarks: stone is missing. Drawing by W.F. Wakeman in

Description: stone known only from Wakeman’s drawing of 1839, which describes it as ‘south-east’ (Herity 1990a, 93) or ‘south of the church’ (White Marshall and Rourke 2000, 159). Upright slab with tapered sides and damaged top, annotated by Wakeman as 2ft 3ins (0.69m) in height. It bears a plain Latin cross, apparently in relief, whose shaft may rise from an expanded base.

Cross-slab 61 (Fig. 7.10) Provenance: unknown. Dimensions (max.): unknown (see below). Geological ID: unknown. Remarks: stone is missing. Drawing by W.F. Wakeman in 1839 (Slab 11, RIA MS 12T9, 43). Archaeological context: unknown. Description: stone known only from Wakeman’s drawing of 1839 on a sheet labelled ‘near the church’. Cross-slab, recorded by Wakeman as 2ft 6ins (0.76m) high, and probably about 0.29m in width. The foot of the slab, although damaged at one edge, appears to be bevelled to form a narrower tenon. In the upper part it bears in relief an expansional cross with large central roundel and lower terminal, and with no separate arms. The side and top terminals appear to project slightly beyond the outline of the slab. The right and top terminals appear to have bosses or scrolls at the outer angles, while the left terminal is oval. Below the cross there is a worn outline that may show the head and trunk of a human figure, with arms at its sides, separated by a damaged area from possible traces of carving below.

Cross-slab 62 (Fig. 7.10) Provenance: unknown. Dimensions (max.): unknown (see below). Geological ID: unknown. Remarks: stone is missing. Drawing by W.F. Wakeman in 1839 (Slab no. 16, RIA MS 12T9, 45). Archaeological context: unknown. Description: stone known only from Wakeman’s drawing of 1839. Cruciform stone, recorded by Wakeman as 3ft (0.91m) high, and probably about 0.45m in span. It has a slightly swollen shaft and arms created by deep curved notches below and angular L-shaped armpits above, with the upper edge of the upper limb and both arms stepped in outline. The arms are thus clearly defined, although they appear to project little, if at all. The head appears to

the cross-slabs 225

64a

250mm

64b

250mm

Fig. 7.11 Cross-slabs 63–66 (no. 64 by C. Elcock, 1883).

226 high island: excavation of an early medieval monastery

terminate in a narrow finial. The stone stands in an irregular socket-slab.

Cross-slab 63 (Fig. 7.11) Provenance: reputedly taken from High Island to the mainland as a grave-marker. Now located in Kill graveyard, Streamstown, Clifden, Co. Galway. Dimensions (max.): H: 1.01m x W: 0.34m x T: 70mm. Geological ID: mudstone. Remarks: the distinctive iconography and geological identification of this stone make it unlikely to have originated from High Island. Its geological composition and similar iconography suggest this stone originated from Caher Island, Co. Mayo. Archaeological context: unknown. Description: reputed to have been brought from High Island about the beginning of the 20th century, but probably from Caher Island, Co. Mayo. Cruciform slab of mudstone, a sedimentary rock found on Caher but not on High Island (White Marshall and Rourke 2000, 160); heavily coated with lichen. An outline Latin cross in bold relief is framed in a sunken background within a weathered margin that returns across the slab, defined on the lower side by a horizontal groove. The margin is interrupted by the side-arms of the cross, which project a few millimetres beyond the slab. A bold, flat band outlines a cross 0.5m high with expanded and slightly curved arms, the upper armpits being U-shaped and the lower ones semicircular. Set on the flat field enclosed by the outline there are a central 90mm boss and four smaller ones in the widest parts of the arms. On each edge of the top arm there is a small oval projection, a motif which is also found on Caher Island (Herity 1995b, 114–15, fig.29). In addition to the geological evidence, the probable provenance of this slab is strengthened by its resemblance to one on Caher Island, Co. Mayo, which has a cross of similar design although without the outer margin (Herity, infra; idem 1995b, pp.115, 119, fig.29).

Cross-slab 64 (Fig. 7.11) Provenance: unknown but reputedly from the Women’s Cemetery, Omey Island. Dimensions (max.): unknown. Geological ID: unknown. Remarks: missing. Known only from drawing by C. Elcock, 1883 (National Monuments Archive). It seems W.F. Wakeman (1891, Plate 1, fig. 32, p.354) also drew Face B of the cross and noted that the cross had already been sketched by C. Elcock (White Marshall and Rourke 2000, 160).

Archaeological context: unknown. Description: slightly cruciform slab with gently rounded armpits defining a high but shallow side-arm, the opposite and top edges apparently being damaged. The upper part of Face A bears an expansional Greek cross with very short arms, the outer arcs of the side and top terminals being curved, and the latter is angled to conform to the top edge of the slab. No ornament is shown in the terminals, but the drawing evidently represents a central roundel in relief, notched to form a cross with round-ended expanded arms, and a superimposed linear cross with bar-terminals. Face B bears an almost equal-armed linear cross, apparently deeply grooved, with wide curved terminals, the lower one being of D-shaped outline. Except for this last detail, this slab resembles many of those on High Island so closely that, as suggested by White Marshall and Rourke, it was probably either carved by a craftsman from the offshore site or brought from there to Omey at a later time.

Cross-slab 65 (Fig. 7.11) Provenance: inner orthostat north jamb Cell B. Dimensions: see below. Geological ID: garnet mica-schist. Remarks: not fully visible. Archaeological context: identified as a possible cross-slab by Ian Fisher. Probably reused. Description: possible disc-headed slab flaked in the upper part. It measures 0.81m in visible height by 0.33m in width in the shaft and it tapers in thickness from 0.2m at the foot to 90mm at the top of the shaft. The head measures 0.33m in height by 0.43m in width, and its left side shows only slight swelling beyond the vertical line of the shaft, whereas the top and right side are almost circular in profile, with a slightly rounded junction with the shaft. The head is badly flaked, but there appears to have been no ornament.

Cross-slab 66 (Fig. 7.11) Provenance: from rubble (F1) within church enclosure. Dimensions: L: 0.35m (incomplete) x W: 0.37m (incomplete) x T: 44mm. Geological ID: garnet mica-schist. Remarks: fragment. Archaeological context: unknown, redeposited. Description: incomplete slab sheared on two sides. One shallow, semicircular notch 80mm long survives along preserved edge. Circular flat boss in low relief c. 0.18m in diameter evident in upper right side. Stone is too incomplete to decipher its function.

the cross-slabs 227

20 17

Fig. 7.12 Cross-slabs 8, 17, 20, 21 by W.F. Wakeman, 1839.

228

8 the burial record 8.1 The human remains Barra O’Donnabhain

Introduction Six skeletons from three different burial locations were found during the excavations of the monastic complex at High Island. The remains of four individuals were recovered from the stone-built graves outside the eastern end of the church (Area 1). Eight stone-built graves were located in this part of the site and their position at the central location of the complex suggested that this was the most important cemetery, reserved for burial of elite members of the community in the monastery. The graves were numbered 1 to 8, south to north, which means the numbers do not reflect a stratigraphic sequence (Fig. 5.7). Three of the eight graves did not contain human remains, and it was not possible to excavate a fourth one. The extant church was just the latest in a number of building phases at the site and several of the graves pre-dated the structure, but were later modified following construction of the church. Of the two burials not located in the stone-built graves, one (Grave 11) was found outside the north-east passage through the church enclosure wall (Area 8), and the other (Grave 9) was found in a simple and unmarked grave inside the church (Area 2). Radiocarbon dating indicates that the remains were interred during various phases of burial activity that spanned a number of centuries: three of the four burials in the stone-built graves to the east of the church date from the late 9th/late 10th to early 11th century (Phase 1); the burial outside the entrance to the church enclosure and one burial from the stone-built graves east of the church date to Phase 2 (mid-11th century to late 12th/early 13th century). The skeleton interred inside the church dated from the mid-12th to early 13th century and is contemporary with the latter part of Phase 2.

Phase 1 Grave 5; Area 1 These remains (F37) were found in a stone-lined grave (No. 5) behind the east wall of the church (Fig. 5.16). This grave was altered when the extant church was built, so its earliest form is unknown. However, it is suggested by the excavator that the grave was originally a simple earth-dug grave, marked at its west end by a decorated headstone (Cross-slab 9) and at the east end by a decorated foot-

stone (Cross-slab 12) and possibly also by a recumbent slab. This suggestion is based on the fact that the skeletal remains were not contained by the sidestones of the grave, features that are suggested to have been added later, when the grave was remodelled at the time of the construction of the church in the 11th or 12th century. The grave contained the earliest dated human remains from the site – late 9th to late 10th century (cal. AD 881–977 (UB-3999, 1126±22 BP)), – and is suggested to have been associated with one or other of the earliest structures at the complex (see Section 5). When the later church was built, the sidestones and an upper layer of fill were added to secure the sidestones in place. The remains consisted of the mostly complete skeleton of an older adult (55+ years) whose head was placed between two pillow-stones. The only part of the skeleton not found were the bones of the hands and those of the thoracic and lumbar vertebrae. Most of the bones that were recovered were in a fragmentary condition, as well as being very poorly preserved. The cranial remains suggest that this may have been a male, but this is stated with diffidence due to the poor state of preservation. The dentition and the degree of fusion of the cranial sutures indicate that this was an older individual (at least 55 years old at the time of death). This is reinforced by the presence of an ossified thyroid cartilage, which is generally indicative of advanced age. It was not possible to estimate the living stature of this individual due to the fragmentary state of the bones. The following tooth positions were available for inspection (italics indicate teeth that were missing postmortem, while those indicated in bold script had been lost during life): 16 15 14 13 12 11 21 22 23 24 25 26 27 28 48 47 46 45 44 43 42 41 31 32 33 34 35 36 37 38

All of the teeth recovered had deposits of calculus (calcified plaque). These were of moderate size on the remaining anterior teeth, but they were large on the surviving premolars and molars (as per the grading system advocated by Buikstra and Ubelaker 1994). The calculus deposits covered the occlusal surfaces of the lower molars, indicating that all of the maxillary molars had been lost during life (the portion of bone containing the upper right second and third molars was not recovered). The upper right

the burial record â&#x20AC;&#x192; 229

maxillary molars may have been lost relatively early in life as the surviving opposing teeth had only mild to moderate wear, whereas that on the surviving anterior teeth was quite severe. A chronic dental abscess was found at the root of the upper right first molar. This was related to the high rate of dental attrition, which led to the exposure of the pulp cavity of the tooth. Dental caries was not noted in the teeth of this individual. Periodontal bone defects were noted at the following tooth positions: 14 48 47 46

33 34 35

The recession of periodontal bone may be related to the large deposits of dental calculus and to poor oral hygiene, although the identification of this disorder in archaeological material is problematic as the changes seen in ancient bone may reflect the continued eruption of teeth due to severe attrition, as occurred in the anterior dentition in this individual. Few of the joint surfaces of the bones were recovered. Mild degenerative changes (osteoarthritis) were noted in both shoulders, the right hip, right knee and left ankle. These degenerative changes are usually age-related. The left hip and both elbows were also available for inspection, but were normal. Degenerative changes were also noted in the bones of the neck and these changes were more severe on the left side. All of the cervical and the first thoracic vertebrae were affected; the remainder of the vertebrae were not recovered. The only other pathological change encountered was porotic hyperostosis on the parietals and the occipital of the skull vault. This condition, thought to be indicative of iron-deficiency anaemia, was mild on both parietals, but was severe on the occipital. The condition was active at the time of death.

Phase 1 Grave 3; Area 1 These remains (F70) were found in a stone-lined grave behind the east wall of the church, the best constructed grave from the site (Fig. 5.18). The earliest form of the interment was a stone-lined and stone-paved grave thought to have been marked at its west end by a decorated headstone (Cross-slab 5) and at its eastern end by a footstone (Crossslab 4). The grave was altered and realigned in Phase 2, when the church was built and an upper set of sidestones was added along with a recumbent slab (Cross-slab 6) and a new footstone (Cross-slab 7). The skeletal remains produced a date in the late 10th to early 11th century (cal. AD 980â&#x20AC;&#x201C;1025 (UB-4266, 1023Âą23 BP)), and are thought to be contemporary with the earliest use of the grave. Although

one of the paving slabs at the west end of the grave may have been broken at the time the grave was restructured in Phase 2, the burial was not disturbed during the later alterations to the grave and, in contrast to the other skeletons from the site, these remains were relatively well-preserved and the skeleton was complete. The morphology of the bones of the pelvis and cranium indicated that this was a male. The form of the pubic symphysis, the state of suture closure, the degree of dental attrition and the presence of ossified thyroid and costal cartilages all combine to suggest that this man was an older adult (55+ years) at the time of his death. Some of the long bones were recovered intact and the maximum lengths of these are listed in Table 8.1. Standard regression formulae used to estimate living stature (Trotter 1970) indicated that this man was about 162cm tall. Table 8.1 Grave 3, long bone maximum lengths (mm).

Bone

Left

Right

Humerus

295.5

303.3

Femur

417.2

Tibia

338.0

All tooth positions had survived and were available for inspection (italics indicate teeth that were missing postmortem, while those indicated in bold script had been lost during life): 18 17 16 15 14 13 12 11 21 22 23 24 25 26 27 28 48 47 46 45 44 43 42 41 31 32 33 34 35 36 37 38

Cloacae or drainage through the bone associated with chronic dental abscess were present at the following tooth positions: 17 16 46

14 13

23 24

The abscesses in the maxilla were particularly severe. The lesions associated with the molars had drained superiorly into the maxillary sinuses bilaterally. Both sinuses would have had chronic infections, with associated inflammation, pain and a persistent discharge of pus into the nasopharynx. The large number of abscesses was related to the high level of dental attrition noted in the remains. This was severe on all surviving teeth, but particularly so in the maxilla, where the two surviving teeth were worn to mere stumps. The surviving mandibular teeth were not quite so heavily worn, but in all cases, all occlusal enamel had been removed by wear. Degenerative changes had set in at both temporomandibular joints. These were relatively

230â&#x20AC;&#x192; high island: excavation of an early medieval monastery

mild bilaterally and were probably secondary to the heavy degree of attrition. None of the surviving teeth had evidence of caries. Deposits of dental calculus were large on most of the lower teeth and small on the two surviving teeth in the maxilla. The smaller size of the deposits on the maxillary dentition was also related to the severe dental attrition. Periodontal bone defects were common in both mandibular and maxillary dentitions and were related to the bone destruction caused by the abscesses. Porotic hyperostosis, thought to be indicative of irondeficiency anaemia, was noted on both parietals and on the occipital. This was in a mild form and the lesions were active at the time of death. The only pathological condition noted in the post-cranial skeleton was in the form of degenerative joint disease (osteoarthritis). Moderate to severe degenerative lesions were noted in the bones of the neck, where the second and third cervical vertebrae were fused at the posterior joints. The centra, or bodies, of the vertebrae were not affected. Moderately severe changes were noted in the bones of the upper and middle back, both shoulders, both hands and both hips.

Phase 1 Grave 4; Area 1 These remains (F40) were also found in an exceptionally well-built stone-lined and stone-paved grave (No. 4) outside the east wall of the church (Fig. 5.17). The primary phase of the grave consisted of sidestones with possibly a decorated headstone, a decorated footstone and most probably also a decorated recumbent slab (all missing). The grave was altered in Phase 2, when the church was built. An upper layer of sidestones was added and a substantial decorated headstone (Cross-slab 8) marked the west end. The socket for a footstone and sufficient space for a recumbent slab left the excavators with the impression that both these stones had existed. The remains in the grave consisted of the mostly complete skeleton of an adult male, dated by radiocarbon analysis to the late 10th to early 11th century (cal. AD 980â&#x20AC;&#x201C;1023 (UB-4155, 1027Âą19 BP)). The skeleton was in a fragmentary state, though some bones were moderately well-preserved. Evidence uncovered during the excavation suggested that the grave may have been disturbed during the Phase 2 restructuring, and again in the 19th century. The morphology of the bones of the pelvis indicated that this person was a male. The bones were not as robust as those of some of the other individuals recovered from the site. Fragments of both the maxilla and the mandible were recovered. The degree of wear present on the remaining teeth indicated that this was an older adult. This contrasts somewhat with the form of the post-cranial skeleton, where few degenerative changes were noted in the skeleton and those that

were encountered were mild. It would be unwise to overinterpret this difference as tooth-wear is not a reliable single indicator of age and the pattern of attrition noted in the other skeletons from the site may suggest that significant levels of tooth-wear was characteristic of life on the island. The bones of the jaws were relatively complete. The maxilla was weathered, suggesting that it had been exposed to the elements at some time in the past. This may relate to the fact that it is suggested that the headstone was added to the grave at the time of its restructuring during Phase 2. The following tooth positions were available for inspection (italics indicate teeth that were missing postmortem, while those indicated in bold script had been lost during life): 17 47

16 46

15 45

14 44

13 43

21 31

22 32

23 33

24 34

25 35

26 36

The portions of the jaws containing the remaining tooth positions were not recovered. As noted above, dental attrition was severe on all of the surviving teeth, with little or no enamel remaining on any of the occlusal surfaces. It is likely that the high level of tooth-wear had led to the development of chronic dental abscesses at the roots of three teeth: the upper right premolars and canine. Dental caries was not noted in any of the surviving teeth. There were very large deposits of dental calculus on all of the surviving mandibular teeth and on two of those from the maxilla (the upper right second molar and the upper left second premolar). The remaining maxillary teeth were so weathered, it is likely that any calculus that was present was lost as a result of this process. The large deposits of calculus and the high levels of attrition were associated with marked periodontal bone defects at all tooth positions where the teeth were present at the time of death. Most of the bones of the post-cranial skeleton were recovered, though many of these were in a fragmentary state. Some of the long bones could be reconstructed in order to measure their maximum lengths. These measurements indicate that this man had a living stature of c. 165cm. Table 8.2 Grave 4, long bone maximum lengths (mm).

Bone

Left

Femur Tibia

Right 435.5

352.0

350.5

The only pathological condition noted in the post-cranial skeleton was age-related degenerative joint disease. Mild degenerative changes were noted in the right hip,

the burial record 231

knee and ankle. This distribution might suggest that the pathological changes related to a specific activity, which produced repeated strains to the major joints of the right leg. A single traumatic episode could also have resulted in conditions where such degenerative changes could occur. No other evidence of such an event was noted, however. Fragments of some of the cervical, thoracic and lumbar vertebral joints were available for inspection (C1, C4–C6, T2–T12, L1–L5). None had any traces of degenerative disease, though the fragmentary nature of these bones must be stressed.

Phase 2 Grave 1; Area 1 These remains (F41) were retrieved from a simple, clay-cut grave (No. 1) that was located outside the extant church and was the only burial in this Area that post-dated this structure (Fig. 5.28). The remains consisted of the complete skeleton of a robust, older adult male, radiocarbon dated to the mid-11th to mid-12th century (cal. AD 1033– 1169 (UB-4156, 913±19 BP)). Unfortunately, the remains were all very fragmentary and in a very poor condition. At the time of his interment, the legs of this man were crossed below the knee and the excavators were left with the impression that the cadaver had been too long for the grave that had been prepared for him. The sex of this individual was indicated by the surviving portions of the pelvis, skull and mandible, but it was not possible to estimate the living stature of this person due to the poor state of the bones. The general size of the remains indicated that this was a robust individual. The distribution and severity of degenerative changes and the degree of wear noted in the dentition combined to suggest that this individual was a middle-aged or older adult. The following teeth were present in the jaws at the time of death (the tooth in italics was missing post-mortem, while those in bold were lost during life): 18 17 16 15 14 13 12 11 21 22 23 24 25 26 27 48 47 46 45 44 43 42 41 31 32 33 34 35 36 37 38

The bone in the area of the upper left third molar was not recovered. Cloacae, or drainage channels, for pus associated with chronic dental abscesses were present at the roots of the central and lateral upper right incisors, the upper right first premolar and the upper left first molar. Almost all of the surviving teeth had very large deposits of dental calculus. Dental caries was noted on one tooth: the upper left first molar had cavities on the mesial surface of the tooth and at the cementum-enamel junction. In all of the teeth that were recovered, the degree of occlusal wear could be classified as being either heavy or severe

and there were periodontal bone defects at the following tooth positions: 18 17

26 46

Many of the articular surfaces of the joints were available for inspection. Degenerative changes (osteoarthritis) were noted at both shoulders, at both wrists, in the right hand and in the left hip; these changes were mostly mild. More severe degenerative changes were noted in the vertebrae. These bones were particularly poorly preserved, but severe changes were present in the posterior joints of the neck and upper back. The posterior joints of the lower thoracic and lumbar vertebrae had less serious osteoarthritic changes. Very few of the vertebral centra were intact. Those of the neck had minor degenerative changes. The thoracic centra were missing, but those of the lower back were recovered. The bodies of the first and second lumbar vertebrae were compressed anteriorly. Multiple Schmorl’s nodes were found on the twelfth thoracic and first lumbar vertebrae. These nodes are the result of the herniation of the gelatinous core of the intervertebral disc. As this core solidifies in adulthood, the nodes indicate that the vertebrae were subjected to severe compressive strains during adolescence or early adulthood. All of the lumbar centra that were available for inspection had severe vertebral osteophytosis, a degenerative change that is the cumulative result of the strains exerted on the back during life. The sacrum was fused to the ilium on the left side only. There were some unusual inclusions in the bag that contained the vertebrae of this skeleton. The first was a desiccated piece of what appeared to be soft tissue, which measured 21.5mm x 15.4mm x 2.2mm thick. This fibrous material resembled dried muscle tissue. The second inclusion consisted of ten plaques of what may be calcified soft tissue. These plaques averaged about 30mm in length, but the largest measured 48.6mm x 43.1mm, while they varied in thickness from 2.2mm to 1.3mm. The plaques were porous and most were concave. They were similar in colour to the bones recovered from the site (light brown) on one side, and an off-white colour on the other side. The brown sides were striated in places. The plaques resembled descriptions of calcifications of the pleura, which are usually secondary to infection of the membrane that lines the lung. There was no evidence of associated inflammation on the pleural surfaces of the ribs.

Phase 2 Grave 11; Area 8 These remains (F8026) were found in a shallow, clay-cut grave (No. 11) outside the north-east passage (F414/F8040)

232 high island: excavation of an early medieval monastery

through the church enclosure wall (Fig. 5.34). The location of the burial had been marked by a stone-lined grave (F856) that was built some time after the remains had been interred. The skeleton appeared to be that of an adult, but the bones were in an extremely poor state of preservation, so this is stated with diffidence. The remains were dated by radiocarbon analysis to the early 11th to mid-12th century (cal. AD 1021–1159 (OxA-13665, 956±22 BP)). It was possible to identify the skull, some vertebrae, the pelvis and the lower limbs while the skeleton was in situ. The remains had been buried in boulder clay (F8000), part of which appeared to have been a redeposited, dark brown, micaceous sandy clay with many charcoal flecks (F8033). Some lighter-coloured clay (fire-reddened?) was also attached to a few fragments. Most of the bones were recovered with much of this sandy matrix still attached. Six bags of material, representing the body parts listed above, were submitted for analysis. None of this material consisted of intact bone. All of the remains were highly fragmented and friable and had they not been packed in aluminium foil at the time of their recovery in such a way as to preserve their original shape, most of the bones would not have been identifiable. Significant decalcification appeared to have taken place in the entire skeleton. The remains of the skull and the shafts of the long bones that were recovered had all been crushed antero-posteriorly, presumably by the weight of the overburdening fill. In a few cases some of the adhering soil was removed, but as this usually resulted in further crumbling of bone, it was avoided with most of the remains. The bag containing the skull was washed over a fine sieve. This resulted in the loss of much of the bone material, which disintegrated during the washing process. It did, however, facilitate the recovery of the remains of two teeth and, fortuitously, it also allowed the determination of the sex of the individual represented in the remains. Two small fragments of burnt animal bone were found in the charcoal-rich soil that was attached to the remains of the skull vault. It was possible to identify the following skeletal elements in the remains: • the bag of remains recovered from the region of the skull contained fragments of the occipital, right and left temporals (the petrous portions of both bones), a fragment of the posterior portion of the second cervical vertebra and two loose teeth. The latter consisted of a complete lower right second or third molar and an enamel shell from the crown of an upper right third molar. The presence of the latter and the degree of wear on the former indicated that this was the skeleton of an adult. The degree of tooth-wear suggested that this was a middle-aged or older adult, though this is stated with diffidence as dental attrition was particularly heavy at

•

the site. The morphology of the external occipital protuberance, one of the few fragments of that bone to survive, indicated that this individual was male. Fragments of the hip bones: these were identified by the excavator at the time of recovery and appear to be fragments of the ilia. The shaft and distal end of the left femur: in this case, the side identification was based on observations made by the excavator while the remains were still in the ground. The anterior face of the shaft of the bone was very poorly preserved, while the posterior side was slightly better preserved. A relatively rugose linea aspera was patent on the latter side. This is compatible with the identification of this individual as a male. The shaft and distal end of the right femur: the side determination was made at the time of excavation. Some relatively large fragments of the distal shaft of this bone were recovered. These were crushed anteroposteriorly and the linea aspera was not observable. The left tibia: this bone was very badly crushed and its identification and siding were based on the observations made during the excavation. The recovered portion consisted of very small and friable fragments of the shaft only. The right tibia: the side identification was based on observations made by the excavator while the remains were in situ. Some of the proximal articular surface survived.

Phase 2 Grave 9; Area 2 This burial (F208) lay in a simple, unmarked, clay-cut grave (No. 9) inside the church (Fig. 5.12). The remains were dated by radiocarbon analysis to the mid-12th to early 13th century (cal. AD 1163–1230 (UB-4000, 849±16BP)). Apart from the feet, the skeleton was relatively complete, though poorly preserved. The cervical vertebrae were not retrieved and only the remains of the skull vault were found. All of the bones that were recovered were in a fragmentary state. The remains of the cranium consisted of portions of the frontal, the right parietal, some of the left parietal and a fragment of the occipital. There was significant post-mortem damage to the outer table of the right parietal. Two artefacts (208:1, 208:2, Sections 6.1–6.3) had been placed with the burial, which is unusual in Christian burials of this period. The bones of the pelvis indicated that this was a male. The state of fusion of the epiphyses of the long bones suggested that this man was a younger adult at the time of his death. It was not possible to estimate the living stature of this man due to the fragmentary state of the bones. With the exceptions of the bones of the feet and neck,

the burial record â&#x20AC;&#x192; 233

most of the post-cranial joints were available for inspection and the only degenerative changes noted were in the right shoulder, where the lesions were mild. Most of the thoracic and lumbar vertebrae were available for inspection and these did not have any degenerative changes. The only other pathological changes noted in the remains were to the skull. Pitting in both orbits indicated that this individual had cribra orbitalia as a child. The related condition, porotic hyperostosis, was found on the frontal, on both parietals and on the occipital. This condition was active at the time of death. The lesions were moderately severe on the frontal, whereas mild lesions were noted on the parietals and occipital. Both of these conditions are thought to be related to iron-deficiency anaemia. The dentition of this individual was not recovered in situ, but a number of small fragments of human bone and teeth were recovered from three associated contexts inside the church: Features 201, 203 and 215. There was evidence of substantial rabbit disturbance in the entire area and it is likely that this activity accounts for the incomplete nature of the skeleton (F208) and for the dispersal of some of the remains. The fragments of human bone and teeth recovered from the three secondary contexts mostly consisted of portions of the jaws and facial skeleton of an adult and seem likely to be the missing portions of the skeleton (F208). A single human tooth was recovered from F201. This was a lower right canine with a moderate degree of wear. A large band of hypoplasia across the middle of the tooth indicated that this individual suffered some form of systemic stress when aged about two to three years old. This could have been in the form of an episode such as an infectious disease, a bout of malnutrition (many children in the pre-industrial era would have experienced a drop in nutritional standards at weaning) or an acute bout of parasite infestation. The remains from F203 consisted of a portion of a left mandible as well as fragments of the left and right zygomatic bones, a piece of an occipital bone and fragments of two foot phalanges. The portion of mandible was quite small and was clearly not from a robust individual. The following teeth were present in the mandible at the time of death: 35 36 37 38

The premolar and first molar were missing post-mortem and those portions of bone containing the remaining tooth positions were not recovered. A loose, unsided canine was also recovered. The degree of wear on the two remaining molars was heavy, but not excessive, which is

consistent with this bone belonging to a relatively young adult. However, this is stated with considerable diffidence as the opposing teeth could have been lost during life. Neither of the surviving teeth had caries, but both had large deposits of calculus. The remains retrieved from F215 consisted of two loose teeth and a fragment of a phalanx. It was not possible to determine if the latter is from a hand or a foot. The teeth were a lower left first molar and a lower left second premolar. These may be the teeth missing from the portion of mandible retrieved from F203, described above. The degree of wear on the premolar matched that of the second and third molars in the fragment of mandible. The rate of attrition on the first molar was more severe, but this is a common occurrence because this tooth erupts before the rest of the permanent dentition and is therefore exposed to attritional forces for a longer period of time.

Discussion

The excavations at High Island produced four burials (F37, F70, F40, F41) in two separate phases from the elaborate, stone-built graves located outside the church, one burial (F8026) from the passage through the enclosure wall, and one burial (F208) from within the church. The human remains, probably disturbed by rabbit activity, recovered from a number of associated features within the church are likely to belong to F208. Radiocarbon dating suggests the skeletons represent a sequence of interments beginning in the late 9th or 10th century and continuing until the mid-12th or 13th century. In all cases, the poor condition of the remains hindered analysis. It was possible to determine the sex of the skeletons in each of the six burials: all were male or probably male. This is consistent with the monastic nature of the site. If women or children were present on the island when the complex was in use, they did not merit interment within the enclosure built around the church. While the site had a number of phases of interment, age-at-death may have been a factor in meriting burial in the stone-built graves to the east of the church. Three of the individuals in these graves were older men, probably at least 55 years old at the time of death: F37 (Grave 5, Phase 1); F70 (Grave 3, Phase 1); and F41 (Grave 1, Phase 2). The remains in the fourth of the stone-built graves, F40 (Grave 4, Phase 2), were ambiguous in relation to estimating age-at-death. The dental remains, which were incomplete, suggested that this was an older adult, but this contrasted with the lack of age-related degenerative changes in the post-cranial skeleton. There was evidence that this burial was disturbed in antiquity and in the 19th century, so it is possible that the mismatch between the cranial and postcranial remains reflects this activity. The adult male (F8026) buried in Grave 11 outside the entrance passage to the enclosure (Phase 2) was probably not a

234 high island: excavation of an early medieval monastery Table 8.3 Comparison of ante-mortem tooth loss rates from Irish sites (n= number of teeth observable; AM = number of teeth lost ante-mortem; % = percentage of teeth lost ante-mortem).

Site

Date

High Island

9th/10th–12th/13th centuries

Fishamble St and John’s Lane, Dublin

9th–11th centuries

Wood Quay, Dublin

110

22.7

O’Donnabhain forthcoming

388

2.8

12th century

O’Donnabhain forthcoming

561

4.8

Tintern Abbey

14th–16th century

O’Donnabhain 2010

894

151

16.9

Temple Lane, Dublin

13th–16th century

O’Donnabhain et al. 1994

457

6.1

younger adult, while the one relatively late burial (F208) inside the church, in Grave 9, was a younger adult (latter end of Phase 2). It was possible to estimate the living stature of two individuals, both of whose remains were recovered from the stone-lined graves at the east end of the church. These men were of short to medium height (c. 162cm and c. 165cm). At nearby Omey Island, where the skeletal remains are dated to the latter half of the first millennium AD on the basis of the presence of lintelled graves and a leacht, Scott (2006) was able to estimate the stature of sixteen adult males and reported a mean stature of 169cm (range 163.0–178.8cm). Buckley (2005) reported an average stature of 171cm among seven males from Illaunloughan, Co. Kerry, dated between the 7th and 9th centuries AD, while Wells (1981) recorded an average stature of 163.1cm among four males from Iona (range 161.2–165.0cm). The latter burials may date from the 7th to the 14th centuries. McKenzie (2010) reported a mean average stature of 167.1cm (range 152.6–179.4cm) in a sample of 189 males from Ballyhanna, Co. Donegal, that mostly date to between the 13th and 16th centuries AD. A survey of large samples of Irish males undertaken in the 1930s (Hooton and Dupertuis 1955) reported stature means of 172.98cm for West Galway (range 155–196cm) and 174.48cm for the Aran Islands (range 158–193cm). The estimates obtained from the skeletons at High Island are at the lower end of the ranges from all of the comparative assemblages. Most of the dental remains recovered at the site were associated with the four burials (F37, F70, F40 and F41) in the stone-lined graves. In the portions of jaws recovered, 110 tooth positions were available for inspection. Of these erupted positions, 25 teeth (22.7%) had been lost during life. While the small sample size must be borne in mind, this rate of ante-mortem tooth loss is very high. It seems most likely that this was related to a particular diet,

Source

which produced high levels of tooth-wear, combined with poor standards of oral hygiene. Table 8.3 lists the rate of ante-mortem loss at a number of Irish sites from which data are available. As can be seen, the rate from High Island is significantly higher than any other. In vivo tooth loss is usually caused by dental caries, attrition, periodontal disease, trauma, or a combination of these factors. In the High Island skeletons, caries does not seem to have been a major contributing factor. In the sample, 67 teeth were available for inspection. Only one (1.5%) of these had caries. While the small sample size must be stressed again, this rate is very low. Wells (1981) noted an even lower caries rate (0.4%) in an assemblage of 7th- to 14th-century date from Iona, while Buckley (2005) did not find caries among five individuals of early medieval date at Illaunloughan, Co. Kerry. The rates for these and other, comparable sites are summarised in Table 8.4. The low rate of caries among the skeletons from High Island is related to very high rates of dental attrition. A high rate of tooth-wear offers some protection against caries as decayed enamel is worn away before cavities form. At High Island, the high degree of tooth-wear must have been the main contributory factor to the high rate of tooth loss. The high level of dental attrition in the skeletons suggests the presence in the diet of abrasive particles. This is often attributed to the accidental inclusion of fine grit particles as a result of using stone to grind and mill cereals. While the burials in the stone-lined graves represent a considerable temporal span, the pattern of severe dental attrition was maintained over that time, suggesting continuity in dietary practices. It is unfortunate that the dental remains from later skeletons at the site were poorly preserved. The burial (F8026) in the passage through the church enclosure wall was so poorly preserved that it is not possible to compare his dentition with that of the men buried in the stone-

the burial record 235

Table 8.4 Comparison of caries rates (n= number of teeth observable; caries = number of teeth with caries; % = percentage of teeth with caries).

Site

Date

High Island

9th/10th–12th/13th centuries

Illaunloughan

Early medieval

Iona

7th–14th century

Omey Island

Late first millennium

St Marys of the Isle, Cork

13th–16th century

Temple Lane, Dublin Tintern Abbey

caries

1.5

n/a

0.0

463

0.4

Scott 2006

615

4.4

Power 1995

101

2054

4.9

12th–16th century

O’Donnabhain et al. 1994

378

9.5

13th–16th century

O’Donnabhain 2010

635

13.7

lined grave. The latest burial encountered at the site, F208 (Grave 9), is thought to date to when the monastery may no longer have been in full-time occupation. The degree of tooth-wear observed in this individual is less severe than that seen in the earlier monastic burials. This might be a function of age. It might also indicate that this man had a diet that differed from those represented in the monastic burials. The similarity in wear patterns among the latter may be the result of following a particular dietary regimen. Plaque is the primary cause of periodontal disease. Plaque can mineralise to form calculus and this leads to localised irritation of gum tissues and can contribute to tooth loss. Periodontal bone defects were common among the burials from the stone-lined graves. Large deposits of calculus were found on the teeth associated with the four individuals. This implies that standards of oral hygiene techniques were very low among these men. It also reinforces the impression of continuity over time in dietary practices. Calculus accumulates on the teeth more rapidly in an alkaline oral environment, something that is provided by diets high in protein and/or carbohydrates (Roberts and Manchester 2005). The skulls of three of the individuals from the stonelined graves and that of the individual buried inside the church were recovered (that from the burial placed outside the entrance to the church enclosure was fragmentary and too poorly preserved to permit analysis). Two of the burials in the stone-lined graves had porotic hyperostosis, while the later burial inside the church had evidence of cribra orbitalia and porotic hyperostosis. Both conditions manifest as pitting of the bones of the skull. These conditions are generally accepted as being the result of

Source

Buckley 2005 Wells 1981

iron-deficiency anaemia. This is thought to represent an adaptive response by the body as it reacts to some form of biological stress. This latter could have taken the form of malnutrition, infectious diseases, micro-organism infestation or a combination of such factors. There is evidence to suggest that this was a life-long struggle rather than the effects of a period of debilitation prior to death. In the earliest of the remains from the site (F37 from Grave 5, Phase 1), the lower canines had numerous bands of linear enamel hypoplasia. This suggests that during childhood, this individual suffered from systemic stress on a regular basis. The cranial lesions indicate that this is likely to have continued into adulthood. Similarly, the presence of scarring indicative of cribra orbitalia in the orbits of the burial inside the church (F208 from Grave 9, Phase 2) is suggestive of childhood stress, while the presence of porotic hyperostosis in the same skull indicates that similar biological stresses were encountered in adulthood. The pattern of degenerative changes varies between different individuals. This is probably a function of differences in age. The remains from Graves 5, 3 and 1 (F37, F70 and F41; Phases 1 and 2, respectively) were older individuals and had more age-related degenerative changes. While many joint surfaces were not available for inspection, some patterning is evident in the distribution of these lesions and the general impression is of people whose lives were characterised by hard, physically active work. In the older individuals, the bones of the neck and shoulders were arthritic. This pattern of degenerative changes suggests that these individuals may have habitually carried heavy loads on the head or the upper back. Similar patterns of degenerative

236 high island: excavation of an early medieval monastery

changes have been noted in skeletons from a number of periods from Ireland. Photographs and descriptions from the 19th century indicate that the Irish proletariat of that period regularly carried large loads in baskets, or creels, on the upper back, with a strap across the shoulders or around the head. The archaeological evidence suggests that this type of activity had a long ancestry. Schmorl’s nodes are lesions incurred in childhood or adolescence when the vertebrae are subjected to weightbearing strains that result in the herniation of the gelatinous core of the intervertebral disc. This core solidifies in adulthood. The presence of these lesions in the vertebrae of the remains in Grave 1 (F41; Phase 2) suggests that this individual was introduced to adult labour practices at an early age; reliance on the labour of children and adolescents is common in pre-industrial societies. The compression fractures of the vertebrae of this individual may also have been caused by these activities. However, these compression lesions may also be related to the plaques of what may be calcified pleural tissue recovered with that individual. Calcification of the pleura can occur in a number of conditions, including fungal infections such as aspergillosis. Several species of Aspergillus act as opportunistic invaders in debilitated individuals. The fungus enters through the respiratory tract and the principal internal lesions are in the lung. However, haematogenous dissemination to the skeleton can occur, albeit rarely. This usually involves the ribs or vertebrae. In vertebral aspergillosis, the centra can collapse. The condition of the few centra recovered from this skeleton would be consistent with such a diagnosis in this case.

8.2 Strontium and oxygen isotope analysis J. Cahill-Wilson, C.A. Taylor, H. Usborne, P. Ditchfield, and A.W.G. Pike

Samples of two human burials (F41, Grave 1; F208, Grave 9) and one faunal sample (F401) from High Island, Co. Galway, were submitted for strontium and oxygen analysis to the Department of Archaeology and Anthropology, University of Bristol. The strontium analysis was carried out at the laboratories at University of Bristol, while the oxygen analysis was carried out at the Research Laboratory for Archaeology and History of Art (RLAHA), University of Oxford. The purpose of the analysis was to investigate a likely origin for the individual (F208) buried in an unmarked grave (No. 9) inside the church in the mid-12th to early 13th century (cal. AD 1163–1230 (UB-4000, 849±16 BP)). The burial was unusual in both the positioning and the find of two grave goods (208:1, 208:2, Sections 6.1–6.3), one placed on each shoulder. Grave goods are not com-

mon with burials in the Christian tradition and would be more usually linked to a person with Nordic associations. For comparative purposes samples of enamel were also taken from the tooth of an individual (F41) buried in what the excavator believed to be indicative of more traditional Christian form. In order to assess a likely localised strontium isotope ratio for the island, a further sample was taken from the tooth of a rabbit (F401).

Methodology: strontium isotope analysis

The enamel surface of an intact molar was first cleaned using a dental burr and hand drill. A wedge of enamel (c. 0.5mm wide, 1mm deep), representing the complete growth axis of the tooth enamel, was removed using a flexible diamond-impregnated dental disk. Any dentine adhering to the enamel section was then removed using a dental burr, and the remaining enamel sample was cleaned in an ultrasonic bath. The whole tooth enamel section was dissolved in 3ml 7N HNO3. The supernatant was dried and redissolved in 3N HNO3. An aliquot of this solution, representing c. 3mg of solid enamel, was removed, and made up to 0.5ml with 3N HNO3 to be loaded onto ion exchange columns. The strontium was separated using standard ion exchange chromatography, using 70 μl of Eichrom Sr Spec resin (50–100 µm). Samples were loaded into 0.5ml 3N HNO3 and washed with 4N HNO3. Strontium was eluted in 1.5ml MilliQ water. The elutant was dried down and loaded, using a few µl 10% HNO3, onto rhenium filaments preconditioned with 1µl TaCl5 solution and 1µl 10% H3PO4. Strontium isotope analysis was performed on a Thermo Fisher Scientific Triton thermal-ionisation multi-collector mass spectrometer at the University of Bristol. Methodologies employed for Sr isotope analyses follow those published in Haak et al. 2008. (See also Evans et al. 2010.) The typical internal precision of a single 87Sr/86Sr analysis (expressed as two relative standard errors (S.E.) of the mean of the 288 cycles comprising one analysis) was 10 ppm. External reproducibility of 87Sr/86Sr over a five-month period (expressed as two relative standard deviations (S.D.) of the mean) for analyses of both strontium isotope standard NIST SRM987 (n = 20) and an in-house seal tooth standard (n = 10) were <30 ppm.

Methodology: oxygen isotope analysis

Each tooth was abraded with a tungsten carbide dental burr to remove any calculus and particulates adhering to the surface. A longitudinal sample, representing the whole enamel growth period, was removed as a V-shaped wedge using a flexible diamond-edged rotary dental saw. The enamel was cleaned of all adhering dentine using a burr, to produce a sample of core enamel tissue. The samples were then transferred to clean, labelled glass vials, weighed,

the burial record 237

Table 8.5 Results of analysis completed on the teeth of two human skeletons and one faunal skeleton from High Island, Co. Galway, at the University of Bristol and Research Laboratory for Archaeology and History of Art, Oxford.

Lab ref.

Burial type

87Sr/86Sr

δ18Oc

δ18Odw ‰ SMOW

HI F41

Ext. adult male

0.709261±.000004

27.74

-6.81

HI F208

Ext. adult male

0.709536±.000004

26.97

-7.70

29.08

-5.25

HI Rabbit F401

0.709148±.000003

rinsed with MilliQ water, ultrasonicated for three minutes and the liquid pipetted off. The washing was repeated three times and the samples were dried on a hotplate. Each sample was transferred to a clean agate mortar and ground to a fine powder under methanol. After evaporation of the methanol, the powder was collected, weighed and transferred to prepared mini-spin tubes and sent to the RLAHA, Oxford University, for analysis. Oxygen isotope analysis of the carbonate fraction of tooth enamel was performed at the Research Laboratory for Archaeology and the History of Art, University of Oxford, following published methodologies (Pollard et al. 2012). The mean δ18O value of the in-house Carrara marble standard (NOCZ) was -1.867‰ (n=8, 2 S.D. = 0.0523), with the expected value being -1.906‰. The oxygen isotope signatures of enamel samples are reported in parts per mil (‰), firstly as Craig-corrected (Craig 1957) δ18O values in the carbonate fraction (δ18Oc) relative to the VPDB international standard, then relative to SMOW (an international reference standard for isotopes in water known as Standard Mean Ocean Water)1 following Coplen (1988). Data are then converted to δ18O of dental phosphate (δ18Op) following Iacumin et al. (1996), and finally converted to δ18O of drinking water (δ18Odw) following Longinelli (1984).

Comment

Although the underlying geology of the island is known to be a micaceous schist, there are older lithologies present, and a complete geological assessment is important in determining the variation in the strontium isotope range in both the immediate and wider environs. In the absence

of a full geological report, and following recommendations in Bentley et al. (2005), a faunal sample, Rabbit F401, was used to assess a localised 87Sr/86Sr level. The results suggest that there is a variation between this at 87Sr/86Sr = 0.7091 and the two human skeletons, Burial F41 87Sr/86Sr = 0.7093 and Burial F208 87Sr/86Sr = 0.7095. Although the oxygen isotope ratios are within a range found across most of Ireland and Britain according to recently published research on oxygen isotopes in modern drinking water in Ireland, this range is not recorded in the west of Ireland, where a more enriched level of δ18Odw = -4 to -5 has been recorded from a broad range of sites (Darling et al. 2003). This is confirmed by the oxygen isotope result of the faunal sample (F401), which is in keeping with expected oxygen isotope ratios for this area of Ireland, F401 = δ18Oc = 29.08 (δ18Odw = -5.25). Burial F208 (Grave 9) has a lower value of δ18Oc = 26.97 (δ18Odw = -7.7). Although the strontium 87 86 Sr/ Sr isotope ratio is not in itself conclusive, the oxygen supports an interpretation that this male is unlikely to have originated in the west of Ireland. It is unlikely, however, that he is of Norse or Scandinavian origin, as the 87Sr/86Sr ratio is not indicative of an origin on these older lithologies, where comparative data is available these have been in the range of 87Sr/86Sr = > 0.712 (Chenery et al. 2014), and furthermore we would expect a significantly different oxygen isotope ratio c. δ18Odw = > -9. An origin for the individual is more likely to be in an area of similarly dated geology within a similar oxygen isotope range in either Ireland or Britain.

1 For a complete outline of the methods devised for the University of Bristol for strontium isotope analysis and University of Bristol in collaboration with Research Laboratories at Oxford for oxygen isotope analysis, see J. Cahill-Wilson, C.A. Taylor, H. Usborne, P. Ditchfield and A.W.G. Pike (2012), ‘Strontium and oxygen isotope analysis on Iron age and early historic burials around the Great Mound at Knowth, Co. Meath’, in G. Eogan (ed.), Excavations at Knowth 5: the archaeology of Knowth in the first and second millennia AD, 775–88.

238

9 the environmental reports 9.1 The faunal remains Margaret McCarthy

Introduction

The excavations at High Island produced a moderate-sized assemblage of animal, bird and fish bones, recovered in varying quantities from all areas and phases of the monastic site. Over half of the faunal material dates to the 8th- to late 12th-/early 13th-century (Phases 1 and 2) occupation of the island, although the presence of specimens from large, perhaps modern livestock species together with the occurrence of a few modern butchery traces indicates that a certain degree of mixing has occurred. This disturbance was probably due to the burrowing rabbits and the activities of the 19th-century miners and other visitors to the island. The bulk of the micro fauna present in the samples are almost certainly later intrusions and this combined with the presence of numerous rabbit bones throughout the deposits means that the samples cannot be viewed as a totally uncontaminated assemblage. The analysis has nonetheless indicated that the bones represent a good sample of the meat diet of the various occupants of the site. The overwhelming majority of the excavated features consisted of general layers of accumulated organic soils, hearth ash, construction deposits and localised spreads of shell representing episodes of dumping food waste by the inhabitants of the island over several centuries. These bone-producing organic layers and deposits accumulated within and around the focal area of the site, i.e. the church and pre-church features, as well as around the domestic buildings and in the area around the enclosing monastic wall. Dates provided by radiocarbon analysis indicate that the first probable occupation of the site was some time after the early 8th century, but before the mid–late 10th or early 11th century (in Phase 1). The relatively substantial accumulation of animal, bird and fish bones from this early phase indicates extensive activity during this period. However, not all of this activity was necessarily associated with a single, continuously occupying monastic community as a significant proportion (16.8%) of the fauna derived from redeposited hearth ash deposits (F221) that had been dumped over the primary structure (i.e. the ‘Paved Area’ (F245), see Section 5.2) located at the heart of the monastic site. It has been suggested by the excavator (Section 5.2) that this substantial and distinc-

tive deposit may represent detritus from a community that occupied the settlement during an apparent hiatus in the monastic occupation of the site. This hiatus, if such it was, occurred prior to the construction of the secondary structure (i.e. the slotted features F260, F263) and the later church. Similar deposits (F93a–j) thought to date to the same period were located outside the church enclosure and account for a further 51.4% of the fauna from this period. Indeed, when the faunal remains from these two deposits are extracted from this early phase, the relatively small amount of bone from the earliest level at the site may suggest that the monastery was not a busy and active settlement at this stage (or that they did not eat much meat). It must be borne in mind, however, that the amount of domestic faunal material recovered is also influenced by the location of the excavation areas. The faunal remains from pre-church deposits are included in Phase 1 (Table 9.4). The most active period of the monastery dates from when the church was built, probably in the mid–late 11th century or, at the very latest, the early 12th century (Phase 2). During this period and up until the late 12th or 13th century, the monastery appears to have flourished and expanded, with the construction of at least two beehive huts (Cells A and B) and an extended monastic enclosure wall. Large amounts of bone were recovered from this phase of occupation across all excavation areas. The faunal remains from deposits contemporary with this period are included in Phase 2 (Table 9.5). The monastery is suggested to have fallen into decline or possibly have been abandoned around the end of the 12th or start of the 13th century, i.e. Phase 3. Following a period of abandonment, when a number of the structures imperative to the proper functioning of the monastery were no longer in use, the monastery appears to have been reoccupied. Whether this reoccupation was permanent or temporary (ecclesiastical or secular) is unknown. However, relatively substantial amounts of bone were recovered from this period, which is suggested to have lasted to around the early–mid-15th century (end Phase 3). Throughout this time, it is probable also that the island was visited by a transient pilgrim population. The faunal remains from this period of activity are included in Phase 3 (Table 9.6). Intermittent activity may have continued after the

the environmental reports 239

15th century and into the post-medieval period (Phase 4). Little is known about the activities at the site during the earlier part of this period (15th–18th centuries), however antiquarian records from the early 19th century describe an almost physically intact monastery (see Section 3). The excavated evidence suggested that some of the buildings may have been intermittently occupied during this time, possibly by tenant farmers, fishermen and latterly by the copper miners, who were the only other group of people (apart from the monks) recorded to have lived on the island. The miners were based on the island in the early 19th century for what has been estimated to be a period of about six months. The faunal remains from this period are included in Phase 4 (Table 9.7). The analysis of the faunal assemblage has been directed towards finding differences in the meat diet of the various occupants of the site over the various phases of activity identified during the excavation. The faunal samples are dominated by the remains of domestic livestock species and small, indeterminate fragments from the bones of these animals. In addition, there are many fish and bird bones and where sieving was carried out, there were numerous small fragments of fish bone, indicating that the resources of the sea formed a very significant aspect of the diet. The analysis of the fish and seabird remains has not only provided information on the additional food preferences of the various occupants of the site but also on the range of habitats exploited. Fishing and the capture of local colonies of seabirds provided a significant additional source of protein and the location of the monastery on High Island was ideally suited for the exploitation of these resources.

Methods

The faunal material was identified using the reference collections housed in University College Cork and the Natural History Museum in Dublin. Data was recorded onto an Excel database, which includes categories for butchery, ageing and sexing as well as species and element identification. Bones that could not be identified to species were categorised, where possible, according to the relative size of the animal represented. The material recorded as ‘large mammal’ (LM) in the tables, for instance, is likely to belong to cattle, but was too small to eliminate the possibility of horse and red deer (although the former species was not present and just two red deer bones were identified). Similarly, specimens that in all probability were sheep, but which may have also originated from goat, pig or large dog, were recorded as ‘medium mammal’ (MM). Due to the anatomical similarities between sheep and goat, bones of this type were assigned to the category ‘sheep/goat’ unless a definite identification was made using guidelines from Boessneck

(1969) and Prummel and Frisch (1986). There were no definite identifications of goat and the remainders of the elements allowing for discrimination between the two species were all identified to sheep. Ageing data were determined using procedures outlined by Silver (1971) for long bones and by Grant (1975) for mandibles. The relative proportion of the different species was assessed using the fragments total and the minimum number of individuals represented. The principal aim of the fish bone analysis was to identify the different kinds of fish present in the deposits. This was achieved by using the comparative collection of modern skeletons housed in the Department of Archaeology, University College Cork. Fish size estimates were by direct comparison with modern specimens of known weight and length, and information on habitat and shoaling patterns was obtained from Hill (1992) and Wheeler (1978). A considerable amount of the large number of bones in a fish skeleton cannot be identified to species and ribs, fin rays and branchiostegal rays were all included in the indeterminate category. The numerous fragments of fish scales recovered from the sorted residues of soil samples were also included in the indeterminate category.

Recovery

All of the mammal bones and most of the bird bones were retrieved through hand excavation. The remains of smaller fish and some small bird bones were noted in the deposits during excavation and a soil sampling programme was adopted to ensure that a representative sample of these two faunal groups was recovered. A rigorous sampling strategy was implemented and many deposits, especially from the earlier phases of occupation, were sampled for bulk processing. These were sieved and sorted in the laboratory and resulted in the recovery of many small bones of fish and birds that may not otherwise have appeared in the record. Undiagnostic fins and branchial rays, ribs and vertebral fragments were particularly abundant in the soil samples. Not all the sorted residues produced fish remains and the numbers recovered in a standard sample varied considerably. One of the largest samples of fish bones came from the sub-church redeposited hearth ash deposits (F221) during Phase 1 (8th to mid-11th century). The sieving programme not only enlarged the assemblage of fish bones but also added to the number of species identified. The species present in the sorted residues that were not recovered through hand excavation tended to be the remains of young fish and vertebrae from small sprat-size fish, which may have arrived at the site in the stomachs of larger fish.

Condition and taphonomy

In general, the faunal material was reasonably well preserved and the condition of the unaltered bone was

240 high island: excavation of an early medieval monastery

generally uniform across the site and through all phases represented. Bone preservation in Phase 1 (8th to mid-11th century) and Phase 2 (mid-11th to late 12th or early 13th century) material was generally recorded as good, with less evidence for pre- and post-depositional taphonomy, such as weathering and erosion. The bones from surface features associated with Phase 3 (late 12th or early 13th century to mid-15th century) and Phase 4 (mid-15th to late 20th century) suggested that a certain degree of mixing had taken place, with eroded brittle fragments being found together with well-preserved bone. A high proportion of the fragments suffered from moderate surface erosion due to post-depositional processes. Less than 5% of the identified bones showed signs of butchery, which is often an indication of post-depositional disturbance and erosion. Fragmentation rates throughout were high, resulting in large numbers of bones that could only be classified as large and medium mammal remains and many small fragments that were indeterminate to either species or element (Table 9.1). The comparatively small number of measurable specimens and the fact that sizeable numbers of loose teeth were recovered from all areas and phases of the site suggest that the assemblage was in a fragmentary state, which is confirmed by the generally low percentages of identified material (29%). The incidence of loose teeth was particularly high in the later phases, again suggesting that preservation conditions were poor, teeth being denser elements that generally survive better on sites where bones have been left scattered about on the surface, exposed to weathering and trampling. Human settlement indicators, such as charring and burning, were noted on a very high proportion (34%) of the assemblage and these fragments are presumed to relate to domestic activity, such as cooking and/or refuse disposal. Much of the Phase 1 and Phase 2 material displayed evidence for burning and charring and totally calcined bones often accounted for over 60% of the remains from any one feature. These burnt bones would seem to reflect the cooking methods of the various inhabitants of the site, suggesting a large amount of roasting. The burnt bones were subjected to varying degrees of heat, with some fragments being barely scorched while others were burnt to a blue/grey and white appearance. The calcined nature of the bones may relate to food processing, such as roasting meat joints, bird and fish carcasses over an open fire, or may indicate that food waste was cast into the fire as a means of disposal and remained there for a sufficient time to take on the white, cracked appearance of heat-shattered bone. Other reasons for the extremely calcined nature of the bones may be explained by the knowledge that lacustrine organic-rich mud was being used as a fuel by some of the occupants during the earliest Phase 1 period (Lancaster, this volume, p.284). Unburnt animal and fish bones

left lying scattered on the surface would therefore have been a natural part of the mud deposits selected for use as fuel. A certain amount of the bones from deposits associated with the later medieval and post-medieval occupation of the site were also scorched and blackened and this type of damage probably occurred while joints of meat were being roasted over an open fire. There was no evidence for gnawing on the bones, which is probably not surprising as dogs and cats were absent from the identified assemblage. A certain proportion of the assemblage appears to have derived through natural means, with rabbits and microfauna inhabiting the site either during or after occupation. Rabbit bones are present in most phases of occupation and it is likely that these animals caused considerable disturbance to the earlier deposits through burrowing. Rabbits may have been brought onto the island as a food source by the 19thcentury copper miners, who themselves are suspected of causing a certain amount of damage to the monastic site. There were no traces of utilisation on any of the rabbit bones recovered from Phase 2 or 3, which suggests that these animals were not present on the island during this time or during the later medieval period; the indications are that they are relatively recent intrusions. A certain amount of the seabirds and fish from the later phases of occupation may also represent natural occurrences, although the presence of butchery marks on some bird bones and the evidence for extensive burning on both fish and bird remains from the earlier phases indicates that they represent food waste generated by the monks and other early medieval visitors to the island.

Analysis

The bones have been divided into the four phases of activity established during the excavation and the details of the findings are described below according to these phases. The species and faunal groups present do not differ radically between periods. The data for mammals, birds and fish are summarised in Tables 9.1–9.3 in terms of the number of individual specimens (NISP) for each species by individual phase, while context-specific information is supplied for each phase in Tables 9.4–9.7. In all, 8477 bones were recovered through a combination of hand retrieval and sieving. These comprised 6,791 animal bones, 958 fish bones and 728 bird bones (Tables 9.1–9.3). Animal bones were present from each phase of occupation, although the remains were not distributed evenly across the site, being concentrated in certain areas and certain phases of occupation. For Phases 1 and 2 (Tables 9.4, 9.5), the two largest individual samples of animal bones came from deposits associated with the earliest period, i.e. Phase 1, although it has to be taken into account that over two-thirds of the fauna were derived

the environmental reports 241

Table 9.1 Chronological distribution of mammals (NISP). (S/G* Sheep/Goat, LM* Large mammal, MM* Medium mammal, Indet* Indeterminate)

Phase 1

Phase 2

TOTAL

Phase 3

Phase 4

TOTAL

GRAND TOTAL

Cow S/G* Pig Deer Rabbit Mouse Shrew Seal Whale LM* MM* Unid*

55 211 41 10 3 1,209 980 1,314

23 85 1 17 1 44 196 522

78 296 42 27 1 3 1,253 1,176 1,836

89 92 179 9 88 174 234

157 382 1 2 105 18 105 258 186

246 474 1 2 284 9 18 193 432 420

324 770 43 2 311 1 9 3 18 1,446 1,608 2,256

TOTAL

3,823

889

4,712

865

1,214

2,079

6,791

Table 9.2 Chronological distribution of fish (NISP).

Phase 1

Phase 2

TOTAL

Phase 3

Phase 4

TOTAL

GRAND TOTAL

Cod Ling Hake Haddock Pollack Whiting Gadid (cod family) Ballan wrasse Sea bream Wrasse/Bream Mackerel Scad Plaice Conger eel Dogfish Sprat-size fish Unid

3 5 2

19 2

3 5 19 4

1 2 -

1 5 5 -

2 5 2 5 -

2 5 5 5 5 19 4

45 15 15 9 2 2 132

46 5 2 2 1 3 81

91 15 20 11 2 2 1 5 213

61 12 3 19 70

203 1 1 5 178

264 12 1 3 1 5 19 248

355 15 32 1 14 2 3 6 24 461

TOTAL

230

161

391

168

399

567

958

from just two deposits (F221 and F93a–j), both of which may have been associated with a possible hiatus in the monastic occupation of the site. For later medieval activity at the site, i.e. Phases 3 and 4, a more uniform amount of bone was found in a larger number of contexts (Tables 9.6, 9.7).

Of the total assemblage, just 21% were identified to species level, while the remainder were classified as fragments of large and medium mammal remains or were indeterminate. A large proportion of the indeterminate sample consisted of small fragments of calcined animal bone as well as indeterminate fish bones. All of the identi-

242â&#x20AC;&#x192; high island: excavation of an early medieval monastery Table 9.3â&#x20AC;&#x201A; Chronological distribution of birds (NISP).

Phase 1

Phase 2

TOTAL

Phase 3

Phase 4

TOTAL

GRAND TOTAL

Dom fowl Dom goose Shag Cormorant Storm petrel Manx shearwater Gannet Puffin Guillemot Herring gull LBB gull GBB gull Woodcock Snipe Corncrake Plovers Blackbird Passerines Unid

1 3 3 19 4 19

1 24 4 106 7 2 8 1 3 2 17 83

2 24 7 106 10 2 8 1 3 19 2 21 102

11 6 1 5 8 2 3 2 13 23

8 6 92 66 5 30 6 3 3 11 1 3 2 1 1 109

8 6 103 72 6 35 14 5 3 11 1 3 2 1 3 3 13 132

10 6 127 79 112 45 16 13 4 14 1 3 19 4 1 3 3 34 234

TOTAL

258

307

347

421

728

fied taxa have been found on other early medieval coastal sites in Ireland, although some of the small wild fauna are almost certainly recent intrusions.

Phase 1: 8th century to mid-11th century

Animal bones dated to the earliest phase of occupation of the monastery were recovered from deposits within four separate excavation areas: from inside the church enclosure, including the fill deposits of a number of early graves as well as a limited area outside the wall to the north and south (Area 1); from pre-church deposits inside the church (Area 2); from pre-cell deposits within Cell B (Area 3); and from the trench excavated through the monastic enclosure wall to the west of Cell A (Area 4) (Fig. 5.1). The deposits associated with this phase produced the largest quantity of faunal material, with a total of 4,103 bones being recovered (Table 9.4). The various categories of unidentified material formed a comparatively large percentage of these samples, with just 14.37% of the bones being positively identified to species. The high degree of ancient fragmentation may be an indicator of maximum utilisation of the carcasses, but given the incidence of erosion on the bones could also mean that some were damaged through

trampling and post-depositional weathering. Identification of the bones revealed that the larger domestic animals formed the greater part of the assemblage, though the bulk of these were not identifiable to species level. A high percentage (62.66%) of the 4,103 bones recovered from Phase 1 deposits was found in Area 1. However, a large proportion of these (82.03%) originated from the retained burnt deposits (F93aâ&#x20AC;&#x201C;j) outside the church enclosure. The majority of these bones were variously burnt, charred and scorched from cooking and refuse disposal activities and of the 2,109 bones recovered, only 8.67% were identifiable to species level. These deposits also contained a disproportionate amount of pig bones. A range of fish and bird species was also identified from this area, mostly from deposits within and around the pre-church graves. A high proportion of the fragments from here were indeterminate. The second largest assemblage of Phase 1 deposits came from redeposited burnt debris (F221) within Area 2. This deposit sealed the earliest structural remains at the heart of the monastery, suggesting a possible hiatus in the monastic occupation of the site or that the monastery may have been occupied during this time by another community. Soil analysis has indicated that this deposit could have built up

the environmental reports 243

Table 9.4 Phase 1: Species composition (NISP). (S/G* Sheep/Goat, LM* Large mammal, MM* Medium mammal, Indet* Indeterminate)

Area

Feature no. Cow S/G* Pig

Fish

Bird

Seal Mouse Rabbit LM* MM* INDET* TOTAL

Area 1

38 42 56 57 68 69 75 93a–j 95 96 105

1 33 3 -

1 100 14 -

24 -

2 7 9 1 9 8

1 3 2 1 -

1 -

1 5 1,115 14 -

1 3 406 1 55 -

110 15 31 6 77 12 419 72 5

3 113 25 40 7 80 22 2,109 1 158 13

Area 2

207 216 221 224 232 246 250 262 265 267

8 5 -

47 12 1 2 -

12 -

1 2 175 1 2 11 1

5 2 35 -

1 9 -

1 32 -

1 2 148 88 9 -

13 267 25 2 11 -

3 22 692 9 166 2 11 3 22 1

Area 3

347/351

Area 4

4030 4032 4034 4037 4039 4043 4045 4046 4047 4050 4051 4052 4060 4064 4070 4073

3 1 1 55

17 1 2 2 2 1 1 6 1 1 211

5 41

1 230

32 5 1 1,209

166 23 3 17 20 26 3 3 2 3 980

27 17 8 8 179 10 1,314

250 1 19 34 11 7 1 18 22 32 3 184 2 1 3 10 4,103

TOTAL

244â&#x20AC;&#x192; high island: excavation of an early medieval monastery

over a period of between 40 and 200 years (Lancaster, this volume, p.284). A total of 692 bones were recovered from this deposit, accounting for 74.32% of the 931 bones identified from the Phase 1 deposits within this area (Table 9.4). Due to the significant location of the deposit, the recovered bone finds were therefore potentially very interesting, in the expectation that they would highlight differences in the diet of the individuals living at the site during this period. The most obvious difference in bone composition between the F221 deposit and other deposits within this Area is the very large quantities of fish bone from the former deposits, although there may be taphonomic reasons for this in that many more soil samples were taken from this deposit, which will invariably result in the recovery of large amounts of fish bone. Interestingly, however, this deposit accounted for the second highest number of pig bones (twelve) from the site, after the retained hearth deposits (F93aâ&#x20AC;&#x201C;j) outside the church enclosure. The three main livestock animals were represented within deposits from Area 2 and, like those in Area 1, sheep/ goat accounted for the majority. Many of the bones were totally calcined and a large proportion (34.15%) consisted mostly of tiny fragments of indeterminate animal bone. A collection of 42 bird bones was recovered, of which a noteworthy sample of nineteen bones representing at least three woodcock (Scolopax rusticola) were recovered from a layer of charred vegetation (F232) covering natural clay, possibly an old ground surface. There were no traces of butchery on any of the woodcock bones, but the presence of disproportionate amounts of upper leg and wing bones and slight charring on one chest bone indicate that these birds may have been captured for food. Ten rabbit bones found in two layers (F221 and F224) indicate that the deposits here were subjected to some disturbance from burrowing animals. The bones represent both juvenile and adult individuals. Excavation Area 3 accounted for a tiny proportion of the faunal remains from Phase 1, with just three fragments of indeterminate burnt animal bone recovered from within the wood and charcoal fill of a smithing pit (F352). Excavation Area 4 accounted for 14.57% of the bone recovered from Phase 1 deposits and nearly all of these came from fill deposits associated with the construction of the primary monastic enclosure wall (F4040) and from deposits outside, i.e. north of the wall. The quantity of material found would suggest that this area, removed somewhat from the central foci of the site, was used for the disposal of domestic waste. The charred appearance of much of this material indicates that it represents food waste. Of the domesticated mammalian fauna, the predominant species was sheep/goat, which accounted for 211 individual fragments, representing at least ten individuals. Almost all parts of the sheep skeleton were present,

although the bones of the head, particularly loose teeth, were more common than those from other areas of the skeleton. Most of the main meat-bearing upper elements of the body were also present as well as some of the lower extremities, such as phalanges and metapodial bones. Within Areas 1 and 2, at least six individuals were present based on the recovery of adult distal humerii and bones from juvenile individuals. Ageing evidence was limited to observing epiphyseal fusion rates on the long bone as no sufficiently complete jaws were recovered to allow for an assessment of the eruption sequence of teeth. A proximal portion of femur came from an animal over 3.5 years of age at death. The long bone fusion data indicated that 78% of the sheep were less than two years of age at death, indicating that the animals were bred primarily for their meat. Five of the long bones were chopped crudely through the midshaft and there were some instances of superficial knife marks on rib fragments, associated with skinning and the removal of meat. A further three individuals were represented in the bones from Area 4. The faunal material from these deposits was poorly preserved, indicating that it was lying in a secondary position when excavated. Ageing data was sparse, but over 75% of the limb bones had unfused epiphyses, suggesting that most individuals were killed at their prime meat-bearing stage (c. one to two years) and not at an age to indicate that sheep husbandry focused primarily on the production of wool. One of the mandibles was sufficiently complete to be aged and the eruption stage indicated that the individual was immature at death, slightly less than one year old. Poor preservation conditions resulted in the recognition of very few butchering traces on the sheep bones. A few limb bones showed knife marks associated with the removal of meat and a radius had been chopped through the distal articulation during the dismemberment of the carcass. Cattle were the second most commonly represented species in terms of identified fragments (17.91%) and an estimate of the number of individuals present (six) also indicated that animal husbandry centred on the keeping of sheep. Ageing data was extremely scarce and the few bones with surviving articulations suggested that most animals were around one to two years of age at death. The single cow vertebra from wall fill (F4030) was small and porous and came from an individual that was killed before it had reached its second year. Many of the midshaft fragments were porous and quite thin-walled, indicating that they originated from young animals, probably less than a year old, and as with the sheep/goat sample the emphasis appeared to be on meat production. The unfused proximal portion of a humerus from (F4037) indicates that this individual was less than three and a half to four years of age at slaughter. A high proportion of the loose teeth were

the environmental reports 245

also from juvenile individuals. However, five very fragmentary cattle teeth were recovered from the possible ‘old ground surface’ (F232) and although these belonged to an adult individual, it was not possible to assess the degree of wear on the teeth. In terms of body parts, the cattle sample was best represented by meat-bearing elements, such as tibiae, femora and humerii, and the breakage pattern, despite the presence of very few butchery marks, was consistent with these animals having been butchered and eaten. Superficial knife marks on two metapodial bones relate to skinning and dismemberment. Pig was the third most common domesticated mammalian fauna present during Phase 1 and, interestingly, pig was not represented with any significance in any of the other phases. In Phase 1, pig bones accounted for 41 bones from three contexts: 24 bones originated from the retained hearth debris (F93a–j) outside the church enclosure; twelve bones came from the redeposited burnt material (F221) over the primary structure, i.e. the ‘paved area’ (F245) beneath the church; and five bones came from a fill layer (F4030) within the primary monastic enclosure wall. Of the 24 bones recovered from the retained hearth debris (F93a–j), the sample consisted of burnt and unburnt fragments of a selection of bones from all regions of the carcass, indicating local slaughter and consumption of pigs. At least two individuals were present, one over two years of age at death and the other a young suckling piglet. The sample of twelve bones from the redeposited hearth beneath the church consisted mostly of fragments, and the most numerous elements were phalanges and loose teeth fragments. The five bones from the fill deposit (F4030) within the monastic enclosure wall comprised two teeth, a second phalanx, a skull fragment and a metacarpus. These were all totally calcined from being in contact with intense heat. None of the bones showed any signs of having been butchered and the only ageing evidence advanced was from a scapula, which represented an individual less than one year old. The wild mammal component consisted of seal and rabbit. Seal was present in the retained hearth debris (F93b and F93c) as three metapodial bones from an adult individual. The bones make a good match with the common seal, Phoca vitulina, which is one of the most common seal species off the west coast of Ireland. This was the only context in the entire excavation that produced evidence for seal bones and their presence indicates that these sea mammals were occasionally exploited. No traces of butchery were noted on the limb bones, but it is assumed that this seal was captured either on dry land or during a fishing expedition and was brought to the monastery to be eaten. The small samples involved do not allow for meaningful comment on sea mammal hunting strategies and the few bones recovered probably represent

an opportunistic catch. It is noteworthy, however, that seal did not appear to form part of the diet of the later inhabitants of the monastery. Fish bone was recovered in relatively large amounts in a variety of contexts, but this is more a reflection of the sample sieving strategy employed rather than evidence for survival. Notwithstanding this, a significant proportion of fish bones (175 of the 230 identified) originated from the redeposited burnt debris (F221) beneath the church. Over 20% of the bones were calcined and these are clearly associated with food preparation by people using the monastic buildings at this time. The identified sample represents one of the widest ranges of fish species from the site, including, in order of frequency, ballan wrasse (Labrus bergylta), seabream (Sparidus sp), scad (Trachurus trachurus), haddock (Melanogrammus aeglefinus), hake (Merluccius merluccius), conger eel (Conger conger), Gadid and spratt-sized fish (Table 9.2). Most fish bones were recovered through hand recovery, including the larger wrasses, hake, haddock and conger eel. The sorted residues of soil samples produced bones from young fish as well as most of the small indeterminate fragments. Fragmentation rates were extremely high and small indeterminate fragments, including fish scales, fin rays and branchial rays, from the sieved samples formed a major component of the recovered assemblage. Skeletal element distribution for the material showed that the fish reached the site intact, with all fish being supplied with their heads still attached. The small sprat-sized fish may have accumulated naturally on the site in the stomachs of larger fish. A collection of 49 bird bones from c. four or five species was recovered from this phase and, excluding the high proportion of unidentified bones (38.77%), the single largest sample came from a deposit (F232) at the earliest excavated level within the church, which may represent an old ground surface. Nineteen bones were recovered from this layer, representing the remains of at least three woodcock (Scolopax rusticol). There were no traces of butchery on any of the bones, but the presence of disproportionate amounts of upper leg and wing bones and slight charring on one chest bone indicate that these birds may have been captured for food. Another species of bird present was the exclusively pelagic Manx shearwater (Puffinus puffinus), three wing bones being recovered from a deposit of charcoal and dark soil (F56) between Graves 5 and 8 in Area 1. A further three bones (the humerus and sternum) of a cormorant (Phalacrocorax carbo) were also recovered and four bones from passerines, of which a number were wing bones. The amount of bird bone from this level is negligible and the deliberate capture of seabirds does not appear to have been undertaken during the earliest phase of occupation of the monastery. The occurrence of a single fowl leg bone in the fill (F38) of one of the graves (No. 4)

246â&#x20AC;&#x192; high island: excavation of an early medieval monastery

is noteworthy and the presence of medullary bone within its cavity indicates that the monks kept a few hens for the production of eggs.

Phase 2: mid-11th to late 12th/early 13th century

The second phase of activity marked the most active period of monastic settlement and animal bones were recovered, in varying quantities, from most areas across the site. Bone was recovered from six excavation areas, with the densest concentration found within Area 3, i.e. Cell B (Fig. 5.1). The bones recovered from this area may suggest that cooking and/or consumption of meat was carried out in this building. Many of the smaller bones were recovered from redeposited layers either within the fill deposits of the monastic enclosure wall in Area 4 or within grave-fill deposits within Area 1 (Table 9.5). A total sample of 1,309 animal bones was recovered, although little economic or dietary information could be obtained as a large proportion (58.32%) of the bones were extremely fragmented and calcined and could not be identified to species level (Table 9.5). The largest sample from mammalian fauna was from sheep/goat, indicating that yet again sheep/goat continued as the dominant domesticated food-producing species on the island during this time. Within the 85 sheep/goat bones identified, the minimum number of individuals present was eight, including two lambs. The sample included teeth, limb bones, vertebrae and skull fragments. A relatively high proportion of sheep bones with unfused epiphyses were found, indicating that the emphasis in the management of sheep appears to have been on meat production. One mandible was sufficiently well preserved to estimate the age of the animal by using tooth eruption as a guide and this indicated that the individual was around two years of age at death. Seven lamb bones were recovered from hearth debris (F319) within Cell B, indicating that the occupants of this building enjoyed a prime cut of lamb, perhaps on a special occasion. Lamb was also present in the fill (F66) of Grave 3 and a neo-natal fatality was recorded in the fill (F45a) of pseudo Grave 2. Evidence for butchery on the bones included the medial and lateral division of two vertebrae as well as skinning marks on two other vertebrae. Only 23 bones could be identified as originating from cattle (Table 9.5). An astragalus and three upper limb bones were recovered from within Cell B and two fragments of a humerus from surface layers around the graves in Area 1. One of the limb bones from Cell B was identified as calf. The small quantity of sub-adult cattle bones from other areas consisted entirely of elements from the extremities of the carcass. Very fragmentary loose teeth were recovered from deposits F478 and F4017 and the

collection of thirteen bones from deposits (F803n, F803s and F844) around Cell B consisted entirely of lower leg bones, including distal metapodia and phalanges. These bones represent primary butchery waste, but as they were recovered from redeposited soil, it is not possible to state with certainty that the slaughter of cattle was undertaken in this area of the settlement. At least one pig was present, based on the recovery of a single tooth from a deposit of hearth ash (F478) outside Cell A and sealed beneath the extended monastic enclosure wall (F4025 SS). A total sample of 161 fish bones was collected from five excavation areas (Table 9.5). The identified fish species, in order of frequency, included whiting, ballan wrasse (Labrus bergylta), whiting (Merlangus merlangus), ballan wrasse/sea bream (Labrus bergylta/ Sparidus sp), scad (Trachurus trachurus), plaice (Pleuronectes platessa), bones from members of the cod family (gadidae), and a small member of the elasmobranch family, probably dogfish (Scyliorhinus canicula) (Table 9.2). The whiting bones were quite small, from individuals less than 20cm in length, and were found together in a single deposit (F45a) within pseudo Grave 2 in Area 1. The single largest collection of 61 fish bones came from a deposit (F702) within the mural chamber in the western side of the monastic enclosure wall. Identification rates were low as the material was highly fragmented and totally calcined. The sample of bird bones amounted to 259 fragments, of which 175 were identified to species level. Wild species, in particular seabirds, dominated the assemblage and identified species, in order of frequency, included storm petrel (Hydrobates pelagicus), shag (Phalacrocorax carbo), puffin (Fratercula artica), Manx shearwater (Puffinus puffinus), cormorant (Phalacrocorax aristotelis), herring gull (Larus argentatus), snipe (Gallinago gallinago), gannet (Sula bassana), guillemot (Uria algae) and a species of small passerine. Domestic fowl appeared again as an almost complete coracoid from an adult bird. The seabird sample was dominated by storm petrel remains (106 specimens), representing adult and juvenile individuals. Over 34% of the bones were from young individuals that could have been taken when the birds were nesting on the island in late summer. Today, storm petrels nest in burrows on the island and in walls or piles of stones and it is presumed that they would have been captured in similar locations during the early medieval period. However, whilst it is possible that the large quantity of storm petrel bones represent food waste, the fact that they were recovered from surface deposits in the area of the graves could mean that they died from natural causes. There is no supporting evidence on the bones to indicate that these birds were deliberately caught and eaten. In contrast to this, most of the 24 shag bones recovered from a fill layer

the environmental reports â&#x20AC;&#x192; 247

Table 9.5â&#x20AC;&#x201A; Phase 2: Species composition (NISP). (S/G* Sheep/Goat, LM* Large mammal, MM* Medium mammal, Indet* Indeterminate)

Area

Feature no.

Cow

S/G*

Pig

Fish

Bird

Rabbit

LM*

Area 1

11a 21a 25a 45a 46 66 98

1 1 -

3 4 2 2 -

2 19 1 2 2 -

78 2 26 11 17 1

5 1 -

7 -

4 -

91 8 53 3 13 22 1

Area 2

206 256

6 -

2 -

10 2

Area 3

319 321 335

4 -

39 3

21 15

1 3 -

9 -

3 -

31 -

72 -

180 3 18

Area 4

448/486 457 476 478 4010 4011 4012 4013 4016 4017 484/4079

1 1 1 1 -

1 18 -

1 -

4 24

4 1 92 6

1 1 6 -

23 35 1 10 7 1 5 24 -

2 12 8 7 104 274

23 43 1 4 13 12 34 2 13 227 304

Area 7

702

Area 8

803 833 844 874 884

5 7 1 23

3 2 2 85

10 161

12 4 258

4 7 15 46

5 1 42 196

8 13 22 522

39 8 34 80 2 1,309

Total

(F4017) in the extended monastic enclosure wall (F4025) were burnt from being roasted over a fire, and a cormorant bone from a deposit of limpet shells (F321) within Cell B bore knife marks associated with meat removal. The two gannet bones, a radius and a tibiotarsus, both from clay packing (F803n) around the north side of Cell

MM* INDET*

TOTAL

B, were from juvenile individuals that may have been taken from nests on the high cliffs around the island in the late summer.

248â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Phase 3: late 12th/early 13th to mid-15th century A relatively large sample of animal bone was recovered from deposits associated with the use of the site during this phase. Animal bones were found scattered in most areas of the site, with notable concentrations in Areas 2, 3, 4 and 8 (Table 9.6) (Section 5, Fig. 5.1). Relatively sizable amounts of cattle and sheep/goat bones came from within Cell B (Area 3), and while it has been suggested by the excavator that the site may have been abandoned by the monks at this stage, it would appear that Cell B was used for shelter and the dumping of food waste during this time. Various deposits within the church and in the area surrounding Cell B also produced relatively large quantities of bone, and the dumping of refuse seems to have focused on these areas. In contrast, very little mammalian bone was found within Cell A (Area 5) and in deposits excavated in the northern section of the site (Area 4). The features dated to this phase of activity collectively yielded a sample of 1,105 bones (Table 9.6). In contrast to the earlier phases, both livestock species of sheep/goat and cattle were almost equally represented in terms of minimum number of individuals and total fragment counts. Sheep/goat was marginally the dominant species, contributing 92 bones from a minimum of four individuals. Most parts of the body were present, indicating that the meat consumed at the site during this period did not come in the form of prepared joints but as live animals on the hoof, which were slaughtered and butchered in the vicinity of the monastery. While it is suggested that the monks may have left the island by this stage, the site continued to be occupied by groups of people who erected new structures and were also probably squatting in the abandoned monastic ruins. Several bones were chopped and split and there were a few instances of axial division of the carcasses, typical of medieval butchery methods. Five of the sheep mandibles bore evidence of dentition, of which just one had a fully erupted tooth row; from the wear pattern, it belonged to an individual over four years of age at death. One of the mandibles from Area 1 came from a lamb less than six months old and two other lamb bones were recovered from a deposit (F215) within the church. The largest quantities of cattle bones from the site derive from this phase of activity, which would seem to indicate that increased numbers of cattle were brought onto the island for grazing following the demise of the monastery in the late 12th or early 13th century. A total sample of 89 bones was identified and it included virtually all elements from the carcass, such as skull fragments, upper limb bones, peripheral elements, teeth and vertebrae. There were more teeth and vertebrae than any other element and five of the vertebrae were chopped axially

along the midline of the body, a procedure that would usually have required the suspension of the carcass prior to butchery. As with sheep, the ageing data indicated that there was a high incidence of animals between one to two years of age, which is consistent with the slaughter of cattle primarily for their meat. There was also evidence for the slaughter of calves for their meat, with 18 calf bones alone being recovered from deposits within Cell B. Some of these bones bore fine knife marks associated with skinning and the removal of meat. No pig bones were identified and the wild mammal component was dominated by rabbit bones (181 specimens) representing a wide range of age groups, from newborn to adult individuals. Rabbits were introduced into Ireland by the Anglo-Normans in the 13th century and while it is possible that they were brought onto the island during this high medieval phase of activity, it is more likely that they were introduced to the island in the early 19th century as a food source for the copper miners. It is always difficult to determine whether or not the remains of rabbits are contemporary with the occupation of a settlement as these animals often burrow into archaeological layers. At the time of excavation High Island was overrun with rabbits (they have since disappeared) and the absence of butchery marks on the bones and any other traces of utilisation associated with consumption suggests that the lagomorph remains represent recent intrusions into the medieval deposits through burrowing. Fish and bird bones were also quite common in the Phase 3 deposits, although the fresh-looking appearance of some of the bird bones indicated that they represent natural fatalities. The sample of 168 fish bones was made up entirely of saltwater species and was dominated by wrasse vertebrae and head bones. The three other potential food species, cod (Gadus morhua), hake (Merluccius merluccius) and scad (Trachurus trachurus), collectively yielded just six bones and these species were clearly not significant food items. Nineteen vertebrae from a small sprat-size fish were also recovered from the sieved samples, but it is unlikely that these were exploited for food. The widest range of identified bird species was present in the Phase 3 deposits despite the recovery of a relatively small assemblage of 74 bones. Most of the identified elements represented seabirds, including, in descending order, shag, gannet, cormorant, Manx shearwater, puffin and storm petrel. Three bones from a species of plover (Charadrius sp.) were identified as well as blackbird (Turdus merula) and a species of small passerine.

Phase 4: mid-15th to late 20th century

The latest phase of activity at the site contributed the second largest quantity of material to the bone assemblage. Chronologically, the excavated deposits span a period of

the environmental reports â&#x20AC;&#x192; 249

Table 9.6â&#x20AC;&#x201A; Phase 3: Species composition (NISP). (S/G* Sheep/Goat, LM* Large mammal, MM* Medium mammal, Indet* Indeterminate)

Area Area 1

Feature no. 4

Cow

S/G*

Rabbit

Shrew

Fish

Bird

LM*

MM* INDET* TOTAL

Area 2

202 214 215 217

4 1 19 9

19 -

9 -

2 28 29 1

2 11 10 -

3 -

12 7 3

88 -

20 140 84 13

Area 3

318 324

18 17

22 13

76 81

9 -

21 20

10 52

114

156 297

Area 4

445 463 464 472 479 482 4002 4005 4013

1 1 1

4 1 -

3 2 82 1 3 -

6 20 -

1 2 1 -

25 21 17 1 7 1

15 -

33 8 124 33 1 5 9 1 2

Area 5

508 509

2 -

3 -

2 2

6 -

13 2

Area 6

613/623

Area 8

837 847 849 852 890

33 1 11 89

12 1 2 92

179

1 16 168

2 3 5 74

20 9 3 86

4 10 174

12 5 234

59 26 39 2 24 1,105

Total

almost 500 years, therefore the faunal remains may be associated with late medieval pilgrims to the site or may have derived from the activities of the copper miners who, it is suggested, lived on the island for at least six months during the early 19th century. It is also likely that at least some originate from shepherds who may have spent time on the island tending to livestock, or to fishermen taking shelter on the island or coming ashore to make small fires and cook, as a long-lived tradition for this remains in the area. The bones were scattered across most areas of the site with the exception of Area 5, Cell A. Bone was

found in relatively large numbers between the church and the enclosure wall and the area immediately outside the wall on both sides (Area 1), from within the church (Area 2) and also within Cell B (Area 3). There was little evidence for occupation of the site at this time and many of the remains, especially the fish and bird bones, may have accumulated by natural means. In general, the faunal assemblage was poorly preserved, with the limb bones, in particular, displaying the weathered and eroded appearance of bones that were not lying in a primary context when excavated.

250 high island: excavation of an early medieval monastery Table 9.7 Phase 4: Species composition (NISP). (S/G* Sheep/Goat, LM* Large mammal, MM* Medium mammal, Indet* Indeterminate)

Area Area 1

Feature Cow S/G* Pig Deer Rabbit Whale Fish Bird LM* MM* INDET* TOTAL no. 2 4 1 2 1 10 9 27 47 9 1 18 28 82 2 7 17 47 73

Area 2

200/201 201 211 212 213

8 1

3 4 10 4 6

23 3 1

3 1 -

7 1 1 3 2

1 -

2 5 4 -

6 4

35 11 17 26 14

Area 3

310 311 313 314 315

1 28 1 -

4 42 32 38 -

1 -

5 10 23 6 1

1 -

1 2 -

1 4 1

11 28 4 -

27 -

11 67 143 49 2

Area 4

401 416 425 429 430 433 434 436 441 442 447 449 451 452 461 473

7 6 11 1 1 4 3 2 1 -

2 19 4 14 18 2 8 13 3 -

1 -

2 10 2 1 1 -

142 44 4 11 -

12 12 13 12 3 3 7 102 -

7 5 13 11 -

4 9 6 13 2 21 24

6 12 7 1 -

16 13 53 29 202 32 57 27 1 7 3 152 4 2 21 24

Area 6

600 601 606

1 -

1 1 1

Area 7

700

Area 8

801 805 806 807 809 811 814 816 818 822 823 829 831 840

4 6 4 24 1 2 1 23 1 157

2 9 1 8 1 30 4 1 6 1 1 73 2 1 382

1 105

2 5 2 128 4 26 4 2 2 399

2 10 4 23 50 20 8 1 347

5 2 2 12 5 2 24 105

2 8 5 44 4 4 1 57 258

8 12 52 12 30 186

8 41 23 171 17 238 45 8 13 2 1 217 5 2 1,960

Total

the environmental reports â&#x20AC;&#x192; 251

Altogether, 1,960 bones were scattered across seven of the excavated areas and these came from general surface spreads and deposits, rubble layers and pockets of limpet shells. Identification of the bones revealed that the two main livestock animals formed the greater part of the mammalian assemblage (Table 9.7). Sheep/goat reclaimed their role of most numerous livestock animal, though the proportion of cattle was also quite high, particularly in terms of the minimum number of individuals present. There was just one pig bone: a complete second phalanx from an individual less than two years of age from a rubble deposit (F436) north of the monastic wall. Ageing data for sheep/goat indicated the presence of animals that were slaughtered at their prime meat-bearing stage, with unfused epiphyses suggesting the presence of animals less than two years of age. A similar age pattern was established for cattle, with 67% of the surviving epiphyses and a few teeth-bearing mandibles representing individuals less than two years old at slaughter, while a relatively high proportion (27%) of cattle were slaughtered in their first year. In terms of skeletal element representation, both the meatbearing upper limb bones and peripheral elements were present, indicating that the deposits contained a mixed assemblage of primary butchery waste and kitchen waste. Evidence for butchering on the bones was not extensive, consisting mostly of chop marks, associated with the separation of the major joints, and knife marks, caused during the removal of flesh from the bones, which were noted on some scapulae and vertebrae. The range of other mammalian taxa was limited, confined to an indeterminate species of whale, red deer (Cervus elaphus) and rabbit (Oryctolagus cuniculus). While most of the rabbit bones may be intrusive, superficial knife marks on a tibia from a deposit (F313) inside Cell B (Area 3) indicated that at least one rabbit was captured and skinned by the later inhabitants of the island. Red deer was identified from two antler fragments that were sawn at both ends and are interpreted as representing waste from antler-working at the site at some stage during this period. Eighteen fragments of a whale skull were found in a single deposit (F47) between the church and the enclosure wall. Two of these were carefully sawn on all sides and appeared to represent off-cuts from bone-working. The fragments were extremely small from being worked and it was not possible to identify the species of whale that was exploited. The sample of bird bones from the latest phase of activity at the site is by far the largest (F347) in the entire assemblage, although a certain amount of these may have accumulated naturally. The domestic fowl and geese bones from Area 1 (F2), Area 2 (F212) and Area 3 (F310, F311) are clearly associated with human activity in and around the monastery during the post-medieval/early modern period, and the presence of medullary bone in one of the fowl leg

bones indicates that laying hens were still present on the island for at least some part of this phase. Seabirds, not surprisingly, constituted the largest proportion of identified species and included substantial amounts of shag (Phalacrocorax aristotelis), cormorant (Phalacrocorax carbo) and Manx shearwater (Puffinus puffinus) bones (Table 9.3). These represented a mixture of young and adult individuals and the presence of fine knife marks on two shag leg bones from a deposit (F811) around Cell B indicates that seabirds continued to be captured and eaten during this period. The less commonly occurring seabirds, such as gannet (Sula bassana), guillemot (Uria aalgae), storm petrel (Hydrobates pelagicus), puffin (Fratercula artica), herring gull (Larus argentatus), lesser black-back gull (Larus fuscus) and great black-back gull (Larus marinus), may well have been nesting in the ruined buildings of the monastery and died there from natural causes. Gannet chick bones were recovered from a deposit (F430) around Cell A. Incidental finds of land-based wild birds included snipe (Gallinago gallinago) and corncrake (Crex crex). The recovery of a corncrake wing bone from the interior of the church is worthy of mention, though a single occurrence in a surface location suggests this bone represents a natural death. Seven species of fish were recognised, although the overwhelming majority (92%) of the identified sample belonged to wrasse, with the remains representing at least six medium- to large-sized individuals. All other fish present were represented by fewer than ten bones, including cod (Gadus morhua), ling (Molva molva), Pollack (Pollachius pollachius), mackerel (Scomber scombrus), conger eel (Conger conger) and a small species of the elasmobranch family, probably dogfish.

Discussion

The numbers and varieties of mammals, birds and fish represented by the bones from the excavations at High Island suggest that the monastic community and other later occupants of the island obtained protein from many different sources. While most of the bones examined were from domestic livestock, seabirds, fish and shellfish also formed a comparatively large component of the monastic diet. The bones from the later phases of activity show remarkable continuity in food consumption, with the earlier monastic community indicating that the later inhabitants of the island relied almost equally on domestic livestock, fish and seabirds for their subsistence. The bulk of the material from all phases of activity is characteristic of domestic refuse, consisting of food debris and butchery waste, which became incorporated into the general accumulation of soils throughout the various phases of occupation. In terms of the monastic community, i.e. Phases 1 and 2, with the probable exception of the deposits recovered

252 high island: excavation of an early medieval monastery

from the redeposited burnt deposit (F221) over the primary structure beneath the church and those recovered from the retained hearth debris deposits (F93a–j) outside the church enclosure, it was possible to determine general patterns in bone disposal and in turn suggest those areas of the settlement where meat preparation, cooking and refuse disposal may have taken place. It is suggested that butchery may have been undertaken away from the focal area of the settlement, but any recognition of specific butchery areas has to be tempered with the knowledge that a high proportion of the bones, in particular those from deposits used in the construction of the monastic enclosure wall, may have been redeposited from elsewhere on the site. The fact that bones were found in such quantities around Cell B indicates a relationship with domestic buildings and suggests that this was a central zone of activity for cooking and eventual disposal of food waste. It also seems significant that relatively large quantities of bone were recovered from within Cell B (Area 3), indicating that food was prepared and consumed in this building. This is corroborated by the archaeological evidence uncovered from the excavation, which suggests that this structure may have functioned as a building where food was prepared and probably eaten. The overall scarcity of bones within the church and around the pre-church structures reflects the non-domestic function of these buildings, and the only substantial sample of bones from this area came from redeposited burnt debris (F221) thought to have been introduced deliberately during a possible hiatus in the monastic occupation of the site. The data also show some variation between individual features and excavated areas to highlight systems of rubbish disposal and, as previously mentioned, to identify areas perhaps of specialised activity, such as slaughter or food-processing, away from the focal part of the site. Disproportionate amounts of primary butchery waste were identified in Area 4, for instance, and there was evidence for the consumption of lambs and a calf within Cell B. It has also to be considered that not all of the bones collected during the excavations relate to subsistence as animals such as rabbits, rodents and some of the birds could have died naturally at the settlement either during or after occupation. Also, some of the bones were collected from surface deposits, especially in the area of the graves (Area 1), where small birds may have nested and died of natural causes beneath dense layers of rubble that engulfed the site prior to excavation. Some of the less common fish species could also have entered the deposits in other ways in that large seabirds may have collected stranded fish from the shore and deposited them at the site.

Domestic and wild mammals

All the common livestock animals, sheep/goat, cattle and

pigs, were found in Phase 1 and Phase 2 deposits (Table 9.1) and were presumably brought to the island by boat when the monastery was first established. In examining the faunal material from the earliest period of occupation (Phase 1), it should be taken into account that a significant proportion of the bones from this phase derived from just two deposits (F221, F93a–j) and that there are some variations in the relative frequencies of identified species from these deposits. In terms of land-based resources, the monks were fed overwhelmingly with mutton and mortality data indicate that the meat came predominantly from sheep reared specifically for this product. It was difficult to assess the ratio of sheep to goat, but as there were no definite goat bones in the samples it is assumed that the ovicaprid material represents sheep, which follows a typical trend for most sites of this date (8th to late 12th/early 13th century). A small proportion of the bones came from adult sheep that would have been kept for maintaining herd levels as well as for the processing of wool, which was widespread in Ireland by this time. There was also evidence for the consumption of lamb during Phases 1 and 2, with a cluster of lamb bones being recovered from a hearth deposit within Cell B. The occupants of this building would seem to have eaten choice lamb, perhaps on the occasion of a special feast, such as at Easter. In addition, the recovery of a calf bone from the same level, while by no means conclusive on the basis of a single find, suggests that the occupants of Cell B enjoyed better quality foodstuffs than the other brethren. Cattle ranked second in importance in both Phases 1 and 2, including the deposits associated with the possible hiatus of monastic occupation at the site. The marginality of the landscape on a remote off-shore Atlantic island is reflected in the kill-off patterns, which indicated that cattle were slaughtered at an age when they had reached their optimum for meat yield. The slaughter of animals less than a year old was also common and the few bones from individuals over four years of age presumably represent cows that were kept to maintain the herd and to supply milk. The high juvenile mortality rate amongst cattle can be paralleled with other early medieval coastal sites and has been interpreted as representing a shortage of fodder in the winter months as hay was not harvested in Ireland until the 13th century, with the arrival of the Anglo-Normans (McCormick 1991; Murray and McCormick 2005). Pigs were the least well represented of the main livestock animals for the early medieval period, with their bones occurring in any significant number only in the earliest phase of occupation. Most of these come from just two deposits, however, the retained hearth debris deposits (F93a–j) outside the church enclosure and the redeposited hearth debris (F221) over the primary structure beneath the church. Both these deposits are associated with a pos-

the environmental reports â&#x20AC;&#x192; 253

sible hiatus of monastic occupation at the site, and the retrieval of pig bones in these deposits might suggest that the group of people who inhabited the site at this time ate more pork than the monks. A small number of pig bones were also recovered from fill deposits within the primary monastic enclosure wall. Ageing analysis indicated that most of the pig bones from the two deposits (F221 and F93aâ&#x20AC;&#x201C;j) are from individuals that had reached their maximum size in terms of meat production and there was no evidence for very young or old individuals. In contrast, the very limited ageing data obtained from the monastic levels (F4030) indicated that all age ranges were represented, suggesting that the monks reared a few adult pigs, probably useful as scavengers of domestic refuse and to provide occasional variety to a land-based meat diet of mutton and beef. The presence of almost entire carcasses of cattle and sheep from both phases of early medieval occupation indicates that slaughtering was carried out within the monastic site and although butchery marks were not particularly extensive, most bones were variously smashed and broken in what can be described as relatively primitive attempts at carcass preparation. The high rate of fragmentation of the bones considerably reduced the number of measurements that could be taken and although metrical analysis was carried out on some suitable bones, the numbers involved are too small to merit detailed discussion. An average shoulder height of 51cm was obtained for a small collection of sheep bones, which fits within the range for other early medieval assemblages. The sample of measurable cattle and pig bones was too small for valid size analysis, but the few depths and breadths taken compare with those obtained for other contemporary sites. Wild land-based mammals do not feature in the early medieval monastic faunal assemblage. The recovery of three common seal bones from just one context (F93aâ&#x20AC;&#x201C;j) that was associated with the period of a possible monastic hiatus is further evidence that the group of people who inhabited the site at this time may have had a different subsistence basis from the monks, in that they seem to have kept larger numbers of pigs and occasionally hunted sea mammals for food (they also used mud as fuel, a feature largely unseen in other contexts: see Lancaster, this volume, p.284). Relatively large amounts of animal bones were recovered from deposits associated with the use of the site in the later medieval and post-medieval period, i.e. Phases 3 and 4. The major part of the sample consisted of the bones of cattle and sheep and, in contrast to the early medieval period, both livestock species were almost equally represented in terms of minimum individuals present and total fragment counts. Despite the limited sample sizes, there is evidence that the numbers of live cattle on the

island increased during the period post-dating the late 12th or early 13th century and for this period (Phase 3) in particular, their bones come to occupy an almost equal position to sheep. The apparently increased presence of cattle on the island following its probable abandonment by the monastic community may indicate that the tillage fields reverted back to pastoral land, with farmers from the mainland bringing drystock to High Island for grazing. In stark contrast, the number of pigs being kept on the island seems to have reduced remarkably, with just one bone being recovered from the later phase of occupation. The wild mammal component was dominated by rabbits, which are either relatively modern in origin or represent animals that were brought onto the island by the 19thcentury miners as a food source. Red deer appear in the archaeozoological dataset for the first time, identified from two antler fragments in deposits associated with the final phase of activity at the site. Both pieces had been sawn and are interpreted as representing waste from antler-working in the vicinity of Cell B during the later stages of the late medieval or the post-medieval period. Eight small skull fragments of an indeterminate species of whale were also recovered in a single deposit between the church and the enclosure wall, two of which had been carefully sawn. The predominance of sheep on High Island, relative, in particular, to the small quantity of pig bones, is attributed not just to the hardiness of these animals but also to the lack of tree cover on the island. The rooting activities of pigs would have destroyed the very thin soils and the poor grazing would not have sustained large numbers of cattle. Local environmental factors may also explain the paucity of pigs as the barren, treeless nature of the island would not have been suitable for animals that relied mostly on pannage in the past. This model of environmental reconstruction and animal exploitation is open to scrutiny, however, as the charcoal evidence from the site indicates that the island supported tree growth at some stage, with twelve different species of tree being identified (see Gale, this volume, p.269). While we know that trees were present on the island in the past, it is likely that they were over-exploited following the arrival of human groups in the late prehistoric period, and the lack of pigs at the site may therefore be an indication that the woodland was severely depleted by the early medieval period. The pollen evidence from the site indicates that cereal-type pollen, including barley, wheat and rye, was prominent and an increase in arable activity is linked to the cultivation of fields by the monastic community, despite the fact that the thin soil cover and the exposed, treeless terrain would not have provided ideal conditions for cereal cultivation (White Marshall and Rourke 2000, 242). The local environmental conditions, on the other hand, were very suitable for sheep farming and the by-product of manure

254â&#x20AC;&#x192; high island: excavation of an early medieval monastery

would have provided a rich source of fertiliser for the tillage fields, and possibly also fuel for the hearths. Hunted animals did not form part of the diet in any period, which is not surprising given the limited range of wild fauna on the island. Rabbits represent the only nondomestic mammalian resource and these animals would not have been available during the primary phases of occupation as they were first introduced into Ireland by the Anglo-Normans in the 13th century.

Marine resources

The monks and the later occupants of the island made full use of the marine environment in which they lived and fish remains constituted a common element in the vertebrate assemblage throughout all the periods of settlement. No butchery was noted on any of the bones, but a large number were charred and calcined, either as a result of cooking or refuse disposal. In the early deposits within Phase 1 it is suggested that much of the burnt bone resulted from lacustrine mud being used as a source of fuel, perhaps both by the monks and by the group of people who inhabited the site during a suggested hiatus of monastic occupation (see Lancaster, this volume, p.284). Eleven species of fish were identified from the early medieval period (Phases 1 and 2), with the most abundant species, in order of frequency, being ballan wrasse, seabream, whiting, scad, haddock, hake, conger eel, plaice and dogfish, as well as indeterminate bones from members of the gadid family (Table 9.2). A quantity of tiny vertebrae from sprat-size fish were recovered from the sorted residues, but these most likely arrived at the site in the stomachs of the larger fish. It should be taken into account that 45% of the recovered fish bones from the early medieval period came from the redeposited hearth debris (F221) over the primary structure beneath the church, therefore it is unwise to include these in any assessment of monastic diet at the site. The identified sample from this deposit represented one of the widest ranges of fish species from the site and included the largest quantity of scad bones, providing further evidence that the overall picture of subsistence for this period is different. The monastic samples are dominated by ballan wrasse, a fish species that is particularly abundant during the summer months and that is easily caught without the use of boats as it tends to live inshore in a rocky marine environment. While several sizes of wrasse are present, the majority of the bones represent fish with estimated lengths between 30cm and 45cm, with a few smaller specimens under 20cm in total length. The seabream bones are probably those of red seabream, which is a summer visitor to the western coastline of Ireland and was formerly caught on rod and line in large numbers by island communities. There is little variation in the size of seabream present in the samples, with the vast majority

deriving from fish with lengths between 35cm and 40cm. All other species occur sporadically, suggesting that they represent incidental catches during organised fishing trips for wrasse and seabream. In terms of numeric frequency, the most common member of the cod family (Gadidae) present was whiting, although these were recovered as a cluster of bones from a (surface grave-fill) deposit (F25a) in Phase 2 and represent a single fish, c. 20cm in total length. As previously mentioned, the remains of scad, being recognised from vertebrae and their diagnostic lateral scutes, were found mostly in the redeposited hearth debris (F221). Scad is an offshore species, but it is rarely found deeper than 100m. In the recent past these fish were often seen split and dry-salted round the west coast of Ireland (Hill 1992). Small quantities of scad were also recovered from the early medieval monastic site on Illaunloughan Island (Hamilton-Dyer 2005) and it was by far the most dominant species in a 13th-century refuse pit at Drumcliffe (McCarthy, unpublished). During the later phases at the site (Phases 3 and 4), and in particular during Phase 4, wrasse continued to be the dominant species, followed by considerably smaller amounts of dogfish, ling, pollack, scad, cod, hake, mackerel, plaice and conger eel. Four species of fish â&#x20AC;&#x201C; pollack, cod, ling and mackerel â&#x20AC;&#x201C; make their appearance for the first time, but the numbers involved are too small to read much significance into this. The few recovered bones for the three gadids (pollack, cod, ling) belonged to mediumsized fish, c. 60cm in length. Other species, such as conger eel and dogfish, were represented by fewer than ten bones and there were just single occurrences of mackerel and plaice, the latter being the only species of flatfish identified at High Island. To summarise the evidence for marine resources, wrasse and seabream together constitute 81% of the total identified assemblage, with wrasse being by far the most commonly eaten fish in all the periods of settlement. These fish are locally common and were of significant value not just to the occupants of the early medieval monastic site but also to the later medieval and post-medieval visitors to the island. Analysis of body part representation indicates that the fish were brought back to the site whole, as bones from all regions of the body were present. Among the less common species, members of the cod family (Gadidae) are the most numerous, including ling, cod and pollack. Pollack favour rough ground habitats and the large fish represented by the five bones from the latest phase of activity (Phase 4) would have been best caught in the autumn, when they moved inshore. There seems to be a slight increase in gadid remains through time and the apparent arrival of three new gadid species in Phase 3 and Phase 4 deposits may be indicative of fluctuating seawater temperatures in the later phases of settlement. The num-

the environmental reports â&#x20AC;&#x192; 255

bers involved are statistically insignificant, however, and any consideration on climate change and fluctuating sea temperatures must therefore be treated with caution. Using information on habitat, biology and presentday methods of capture (Wheeler 1978), some suggestions can be made concerning the type of fishing undertaken by the former occupants of High Island. All of the fish species identified occur commonly off the west coast and most are still valued as food fish. The nature of the coast around High Island is favourable for many kinds of fish, but especially those that prefer rocky habitats. The main point to emerge from the analysis of the fish bone assemblage is that, while a wide variety of species were identified, ballan wrasse clearly formed the staple seafood diet throughout all phases of activity. These fish are very common along the Atlantic seaboard and would have been caught with relative ease from the rocky headlands projecting into the sea at High Island. The young are often found in shore pools, provided there is sufficient algae and rock, and on High Island they could have been fished from the rocky shoreline or from the cliffs with relatively little effort using a baited hook and line. From the generally small size of the individuals it seems that fishing for this species was mainly practised during the summer. The gadids and other fish in the assemblage are medium-sized individuals, which probably lived near the coast during the summer.The capture of seabream could only have taken place during the summer, when they frequent southern Irish waters (Hill 1992; Wheeler 1978). Whiting and scad are generally captured with nets, although the few individuals from High Island are considered to be the result of opportunistic captures during fishing expeditions for wrasse and seabream. In terms of resource availability, there would have been no particular need to venture far out to sea and most fish would appear to have been captured in inshore waters, particularly during the summer months when they occur in larger numbers. The larger pollack from Phase 4 may have been captured from deep water by a fishing boat. Surplus fish were probably preserved for the winter by drying, smoking or salting and racks of dried fish were once a common sight on the islands off the west coast of Ireland. There is little evidence to indicate the use of nets, traps or boats. Regrettably, fish-hooks or definite fishing tools were not found during the excavations as these would have provided information on the methods of capture. The only artefact recovered from the excavations possibly associated with fishing was a weight or line-sinker (95E124:208:2), found with the mid-12th to early 13th-century burial in Grave 9, inside the church (see Section 5.2.1).

Bird remains

A total assemblage of 728 bird bones was examined and the species identified illustrate well the part played by avi-

fauna in the diet of the various inhabitants of the island. The largest sample (47.66%) of bones came from Phase 4 (mid-15th to late 20th century), and it is clear that seabirds, not surprisingly, predominated, with many of the species being those that would be expected to inhabit the sea and cliffs surrounding High Island (Table 9.3). In contrast to the large amounts of fish bones that were recovered from the redeposited burnt deposit (F221) over the primary structure beneath the church, very few bird bones (just two) were found in these deposits, suggesting that the group of people who lived at the site at this time did not engage in the hunting of seabirds. The proportion of bird bones from both of the later phases of activity (Phases 3 and 4) was high, indicating that the capture of seabirds continued into the post-medieval period as a means of providing seasonally available fresh food and of supplementing land-based resources. In terms of understating the formation history of the recovered bird bone assemblage, it has to be considered that many of the bones came from surface deposits under rubble, especially in Area 1, above the graves, where nesting birds may have died of natural causes. As a result, it is unclear if the recovered evidence represents deliberate exploitation of these birds as a food source. There are few apparent butchery marks and other traces of utilisation to provide supporting evidence for the exploitation of seabirds for food, but if one assumes that the commonly occurring birds were being consumed, then it is likely that the carcasses were cooked whole and dismembered by hand at mealtime. The most abundant bird species in the early medieval period were storm petrel, shag and woodcock, with the latter occurring as a single deposit of bones representing three individuals in the old ground surface within the earliest structure beneath the church. All other wild species, including cormorant, Manx shearwater, puffin, guillemot, herring gull, gannet and snipe, are represented by fewer than ten bones each and represent either opportunistic catches or natural fatalities. For the commonly occurring species, most carcass components were recovered, including skull, wing and leg bones, indicating that intact birds were brought back to the monastic site after a fowling expedition. The deposits were dominated by storm petrel bones, representing a mixture of adult and juvenile individuals, but these may represent natural fatalities as there are no traces of utilisation of the bones to suggest that these birds were cooked and eaten. The Blasket islanders ate storm petrels on a regular basis, but the birds were roasted whole over a fire in order to reduce their fatty content (Lysaght 2000). Such food-processing methods would have resulted in the bones becoming totally calcined, a modification not noted on the storm petrel remains from High Island. In contrast, the recovery of a relatively large sample of totally calcined shag bones from a fill layer in

256â&#x20AC;&#x192; high island: excavation of an early medieval monastery

the monastic enclosure wall indicated that these birds had been roasted whole over a fire. The less common seabirds included gannet and guillemot and three species of gull, and these may have been deposited in the samples by chance after dying nearby. It is unlikely that the gulls were eaten due to their unpalatable flesh, though young gannets were regarded as a delicacy by many North Atlantic island communities and fowling for the meat and feathers of gannet chicks was practised commercially on the Little Skellig into the 20th century (Lysaght 2000). Guillemot and puffin are both still commonly eaten in Iceland. The small quantity of young gannets from High Island indicates that they did not form a regular part of the diet. The meat values of these young birds would have been minimal owing to their small body sizes and unless their bones are recovered in large numbers, it is difficult to interpret them as a regular part of the food supply. It is likely that the less common land-based species are incidental remains of birds that died nearby and became incorporated into the deposits by natural means. It is interesting that birds such as lesser black-back gull, greater black-back gull, oystercatcher and fulmars are not present in larger numbers as these are now very common on High Island and could easily have been caught and eaten if they had been present at the time the monastery was functioning. It is known that fulmars only arrived in Ireland and Britain at the beginning of the 20th century, which explains why they are referred to as â&#x20AC;&#x2DC;new gullâ&#x20AC;&#x2122; in Ulster (Anderson 2008). Various species of gull may well have been present in large numbers during the early medieval period, but they were probably not chosen as a source of food as their flesh was not considered very palatable. The sample of bird bones also included the remains of domestic fowl and goose, although these were found in very small quantities and mostly from the later phases of activity. The recovery of two domestic fowl bones from deposits associated with the monastic settlement shows that the monks kept hens as regular providers of eggs. In terms of avian remains, the sieved samples predominantly yielded quantities of bone from small passerine species that were living nearby and their presence indicates that some trees and shrubs were growing in the local area. The later medieval and post-medieval phases (Phases 3 and 4) contributed the largest assemblage of bird bones from the site, although a certain amount of these may have accumulated naturally. The widest range of identified species was in Phase 4. In common with the early medieval period, seabirds were the dominant species, with smaller amounts of other bird species, including blackbird, a species of plover and some small passerines. Although it is not unknown for members of the crow family to be eaten, it is unlikely that the few blackbird bones recovered from the later deposits at High Island represent food

refuse. Domestic fowl and geese were recorded in Phase 4 deposits and the presence of medullary bone in a few fowl bones indicates that laying hens continued to be kept at the settlement during the later phases of occupation. A number of gannet chick bones were collected from around Cell B, but their small quantity indicates that there was no deliberate exploitation of this resource. The avian sample from the later phases of occupation is also noteworthy for producing the only corncrake bone from the site. Breeding colonies for all the seabirds identified in the assemblage would have been located in the nearby cliffs for puffins and gulls, in burrows on high ground for Manx shearwater and in crevices in walls for storm petrels and this is the most likely source of exploitation for the various occupants of the island. Shags are exclusively marine birds, nesting on rocky coasts, while cormorants are largely restricted to shallow inshore waters. Both species would have been readily available for capture throughout the year, though the presence of bones of immature individuals indicates that birds were mostly taken during late summer, following the breeding season. The Manx shearwater is exclusively pelagic and only comes ashore to breed in subterranean burrows on the high ground on the island. Guillemots and puffins also have to come ashore to breed and would have been available for exploitation along the cliffs during the summer months. Storm petrels are the smallest European birds of the open sea, being of a particularly shy nature they come ashore to breed on isolated islands and remote areas of the Atlantic coastline. They breed in loose rubble and stones, where they creep into hollows, and this behaviour would have made the abandoned monastic site a favoured spot for storm petrels. The preponderance of storm petrel in the surface deposits associated with the graves east of the church (Area 1) is striking and may indicate that these individuals were captured, perhaps on a single fowling expedition. The bones belong to a mixture of adult and juvenile individuals, indicating that they may have been captured shortly after the nesting season and before the young took flight, although the absence of butchery and other traces of utilisation indicate that the sample could also represent natural fatalities. The storm petrel was regarded as a delicacy on the Blasket Islands, though the fact that it was swallowed whole rather than chewed suggests that it was not very palatable (Lysaght 2000). The seasonal exploitation of seabirds would, in some cases, have required considerable courage on the part of the fowler, who often had to scale treacherous heights and either seize the birds by hand or capture them with wide nets. The 19th- and early 20th-century occupants of St. Kilda, off the west coast of Scotland, subsisted almost entirely for food in the form of meat and eggs on seasonally available seabirds (Fenton 1997). Gannets and shearwaters

the environmental reports 257

were taken in April, fulmars in May and August and puffins were snared in July. The harvesting of eggs of gulls, fulmars, puffins and guillemots was undertaken in May. On the Great Blasket Island, off the Kerry coast, the eggs of oystercatchers, guillemots, razorbills, puffins and seagulls were eaten as well as the flesh of most of these birds (Lysaght 2000). Overall, it would seem that seabirds, and presumably their eggs, were welcome seasonal supplements to the diet at High Island rather than being staple food items.

Regional integration

The Irish monastic diet is relatively well understood and zooarchaeological evidence is available from inland sites and coastal mainland sites as well as monastic and secular sites on marine islands. The faunal assemblage from High Island has provided another important body of data to further our understanding of animal husbandry strategies adopted by early medieval monastic communities. Monastic marine island assemblages have been examined from Church Island (Roche 1958), Illaunloughan (Murray & McCormick 2005; Hamilton-Dyer 2005) and Skellig Michael (Murray 2011; Hamilton-Dyer 2011), all in Co. Kerry; from Inishmurray, Co. Sligo (O’Sullivan 2008); and in Co. Galway considerable samples of bones, including a small sample of fish bones (Hamilton-Dyer 1994), have been recovered from St Féichín’s other monastery, at the mainland coastal tidal island site of Omey Island. Reasonable faunal samples have also been recovered during excavations at the two secular sites on the Aran Islands: Dún Aonghasa (McCormick and Murphy 2012; McCarthy 2012) and Dún Eoghanachta (McCormick and Murphy 2012; McCarthy 2012). A few coastal mainland sites in the region, including the two sandhill sites at Doonloughlan (Hamilton-Dyer 2000) and Rathgurreen ringfort (Hamilton-Dyer 2002), both in Co. Galway, have also produced faunal samples and the results are incorporated into this discussion. Not all of the assemblages from the above sites are ideal for comparison as some are clearly too small in quantity to allow meaningful comment on diet and economy at the sites. At Inishmurray, for instance, a very small sample of bones was recovered during the excavations at Relickoran, but as most of these came from topsoil deposits, the remains may not necessarily reflect the diet of the early medieval occupants of the island (O’Sullivan and Ó Carragáin 2008). The samples from Church Island (Roche 1958) seem to have been quite substantial, but the published report does not include specific counts and the results cannot therefore be incorporated into Table 9.8. The precipitous nature of the Skellig Michael topography renders comparison with other island sites difficult. Collections from other sites, such as Reask (Fanning 1981, 167), are also too small to be of any value.

Table 9.8 Relative percentages (from MNI) of livestock species at other contemporary coastal sites (based on total number of identified mammalian fauna). (Site 2* Just 16 bones from this site.)

Cattle % S/G % Pig % Marine Islands High Island, Co. Galway Phases 1 & 2

Illaunloughan, Co. Kerry Phase 1 Phase 2 Phase 3 Phase 4

29 33 39 28

43 43 22 24

14 9 18 18

Dún Eoghanachta, Co. Galway Level 4 AD 778–944 Level 3 AD 903–1022 Level 1 AD 956–1013

14 33 26

42 17 34

14 17 13

Mainland coastal sites Doonloughan, Co. Galway Site 1 Site 2*

29 56

29 25

14 19

Oughtymore, Co. Derry

Rathgurreen, Co. Galway Phase 1 AD 374–601 Phase 2 AD 664–856

42 41

21 20

21 23

The most useful site for comparison is Illaunloughan Island in Co. Kerry (Murray and McCormick 2005) as the excavations here produced a large representative collection of all the major faunal groups. In many respects the samples from all of the aforementioned sites have striking similarities with those from High Island, both in the types of species present and in the mortality patterns of the livestock. The relative proportions of the various species identified compare well to the other contemporary island assemblages, both monastic and secular, confirming the significance of domestic livestock in the diet and economy of these sites (Table 9.8). Marine island assemblages are especially characterised by a dominance of sheep bones where local environmental conditions may have been more suited to the management of sheep. The highest proportion of sheep bones

258 high island: excavation of an early medieval monastery

is from High Island, with this species alone accounting for 57% of the land-based mammalian assemblage. A relatively high proportion (43%) of sheep is also recorded for the two earliest phases of activity at Illaunloughan (Murray and McCormick 2005), with an increase in the proportion of cattle during the later phases, a husbandry pattern also noted in the later medieval phase of activity at High Island (Phase 3). Sheep are also the dominant livestock animals during the first phase of activity (8th to 10th century) at the secular site of Dún Eoghanachta (McCormick and Murphy 2012) on the Aran Islands, with cattle again increasing in numbers in the later periods. In striking contrast, the mainland coastal ringfort at Rathgurreen, Co. Galway (Murray 2002), shows a very high proportion of cattle to sheep, while the sandhill sites excavated at Doonloughan (Murray and McCormick 2012) indicated that cattle and sheep bones were present in almost equal proportions. At Oughtymore, an open mainland site in Co. Derry, sheep are marginally more frequent than cattle (Mallory et al. 1984). The considerable effort involved in transporting adult cattle across the open sea, in particular to a precipitous island such as High Island, may in part explain the dominance of sheep in this assemblage. Although the ageing sample from High Island is very limited when compared to that at Illaunloughan, the observed mortality patterns are consistent with other island assemblages and indicate that the majority of sheep and cattle were slaughtered at an age when they had reached their optimum age for meat production, with a very high proportion of cattle being killed under one year old at Illaunloughan (Murray and McCormick 2005). Lambs and calves were slaughtered and consumed at High Island, including some individuals as young as three to four months of age. In general, the data from marine island assemblages seem to support an early slaughter of livestock, which is interpreted as representing the extreme marginality of these remote sites where animals could not be overwintered due to shortage of fodder (Murray and McCormick 2005). The tertiary role of pigs at High Island also seems broadly typical of general trends in pig husbandry observed at the other sites, indicating that environmental conditions seem to have had a considerable influence on the nature of animal husbandry. Murray and McCormick (2005) have postulated that the higher incidence of sheep at western coastal sites is a reflection of the open, unwooded landscape of these regions, although the charcoal evidence from High Island indicates that the island sustained at least ten different species of native tree albeit these could have been depleted by the early medieval period. In contrast to Illaunloughan (Murray and McCormick 2005) and Church Island (Roche 1958), horses do not appear in the faunal record for High Island nor is there any

evidence for wild fauna, with the exception of three bones of common seal, but as these were recovered from redeposited hearth deposits associated with a possible hiatus in the monastic occupation of the site, these bones may not form part of the early medieval diet of the monks. Seals were exploited to a greater extent at Church Island, where large amounts of young and adult grey seal bones were identified in the domestic refuse deposits (Roche 1958). The majority of seal bones recovered from early medieval contexts on Skellig Michael were immature specimens and several displayed evidence of butchery (Murray 2011, 428). Seal bones were also found at Illaunloughan (Murray and McCormick 2005) and Inishkea North (Henry 1945). The sea was an important source of food for these island monastic communities and fish was a significant part of the diet in view of the strict Church rules on fasting (Murray and McCormick 2005). Church regulations approved the consumption of seabirds, such as Manx shearwaters, on the many fast days in the week as well as during Lent because the flesh of these birds was considered to have a salty, fish-like taste (Anderson 2008). The range of fish from High Island was broadly similar to that recorded from other contemporary assemblages, where there seems to have been a reliance on marine species, especially wrasse and seabream (Table 9.9). At Illaunloughan and Skellig Michael (Hamilton-Dyer 2005; 2011), the fish samples are dominated by the remains of seabream, which differs strikingly from High Island and the secular island site at Dún Eoghanachta (McCarthy 2012), where ballan wrasse are the dominant species (64% and 87%, respectively). Local marine environmental conditions have probably influenced the composition of the fish fauna from High Island and the Aran Islands, as ballan wrasse live in relatively shallow waters close to cliffs and rocky shorelines. Relatively large quantities of wrasse were also identified at Doonloughan (HamiltonDyer 2000a), Co. Galway. The dominant fish species at the sites, i.e. wrasse and seabream, would all have been influenced by sea temperature and would only have been available during the summer months when sea temperatures had risen. The presence of seabream at Dún Aonghasa (McCarthy 2012), and Dún Eoghanachta (McCarthy 2012) on the Aran Islands, together with bones from young wrasse and gadids, indicates that fishing was mostly undertaken in the summer months by the early medieval secular inhabitants of these islands. In terms of resource availability, there seems to have been no particular need for the inhabitants of the sites to have ventured too far out to sea and most of the identified fish would have been easily captured in inshore waters by netting or by long line from the cliffs. Scad is not usually regarded as a prime food today, yet it appears in small numbers at many early medieval coastal

the environmental reports 259

Illaunloughan, Co. Kerry

Dún Eoghanachta Co. Galway

Dún Aonghasa, Co. Galway

Drumcliffe, Co. Sligo

Rathgurreen, Co. Galway

Doonloughan, Co. Galway

Cod Ling Hake Haddock Pollack Whiting Gadidae (cod family) Angler Ballan wrasse Seabream Wrasse/Bream Mackerel Scad Gurnard Plaice Bass Brill Mullet Conger eel Common eel Tope Dogfish Ray Shark sp. Salmon/trout Herring Sea scorpion Sprat-size fish Total

High Island, Co. Galway

Table 9.9 Composition of fish species at other contemporary coastal sites (NISP).

3 5 19 4 91 15 20 11 2 2 1 5 178

17 1 120 1 88 11 183 657 1238 81 11 40 9 35 2 25 2 3 2,524

8 137 3 4652 544 1 10 1 2 1 5,359

6 10 2 18

682 3 2 14 701

1 1 9 12

17 107 24 184 36 8 376

sites. At Doonloughan (Hamilton-Dyer 2000a), scad were numerically the most frequent fish with a common element at the coastal ringfort at Rathgurreen, though the sample of fish bones from here is too small to be significant (Hamilton-Dyer 2000a). In common with many other early medieval sites, both monastic and secular, the proportion of cod and other large gadids is low, though Roche (1958) records that cod was one of the dominant species at Church Island. Surplus fish were probably preserved for the winter by drying or smoking and racks of

dried fish were once a common sight on the islands off the west coast of Ireland. The results of the avian analysis are comparable with the assemblage of bird remains excavated at Dún Eoghanachta (McCarthy 2012) and at Illaunloughan (Murray and McCormick 2005) and emphasise the importance of seabirds as an additional source of protein for early medieval communities living along the Atlantic seaboard. The data from both monastic and secular coastal sites suggest a consistent exploitation of the edible seabirds of the region,

260 high island: excavation of an early medieval monastery

Illaunloughan, Co. Kerry

Dún Eoghanachta, Co. Galway

Rathgurreen, Co. Galway

Table 9.10 Composition of bird species at other contemporary coastal sites.

High Island, Co. Galway

although many of the species would probably not be relished today. Species representation at the sites is more or less consistent, although the one notable feature from Illaunloughan is the very high proportion of Manx shearwater, this pelagic species alone accounting for over 70% of the identified avifauna (Table 9.10). The bird bones from Skellig Michael were also dominated by Manx shearwaters (Hamilton-Dyer 2011, 437). This would appear to have a very definite ecclesiastic link as Manx shearwater was one of the birds approved by the Church for meat consumption (Anderson 2008). In marked contrast, there is no evidence for Manx shearwater at the nearby site of Church Island, with shag, cormorant and gannet dominating the identified seabird assemblage from this site (Roche 1958). White-fronted goose and wild duck were also captured and presumably eaten at Church Island. Various species of gull appear at a number of sites, such as High Island, Inishmurray and Illaunloughan, but the small amounts involved indicate that these birds were not considered good for eating. On the off-shore islands, seabirds would have been seasonally available while ashore for breeding and nesting in the rocky cliffs, and chicks and eggs may have been taken from their nests relatively easily in many instances or, if necessary, by scaling the cliffs. In contrast to High Island, the avian samples from Dún Eoghanachta revealed that guillemot and cormorant were the seabirds of most importance, accounting for 91% of the identified assemblage. Shag, razorbill and puffin were also present in sufficient numbers to indicate that they were exploited for their food. Breeding colonies for these species would have been located in the nearby cliffs and the presence of bones of immature individuals indicates that the birds were taken during the breeding season. Referring to Galway, Lewis’s Topographical Dictionary of Ireland (Lewis 1837) describes men lowering themselves down cliffs to net seabirds and he also mentions that marine species were exploited for their feathers. Although O’Flaherty (1935) states that the guillemot was the only bird sought after by the Aran islanders for its flesh and feathers, the data from early medieval features excavated at Dún Eoghanachta indicate that cormorant was equally favoured by earlier residents (McCarthy 2012). These seabirds would have come ashore to nest on islands and coastal promontories, arriving about April, and young birds would have been taken in August when large enough to be left alone by their parents or as potential adults to shortly return to sea. The presence of domestic fowl bone, albeit in small amounts, at most of the identified phases is of note and the figures are adding to the increasing evidence for these birds at remote Atlantic early medieval sites. Domestic fowl bones have also been found at the monastic sites of Inishmurray, Co. Sligo (O’Sullivan and Ó Carragáin

Domestic fowl

Goose sp.

Eagle?

Shag

Cormorant

462

Razorbill

Storm petrel

106

Manx shearwater

405

Gannet

Puffin

Guillemot

595

Kittiwake

Herring gull

LBB gull

GBB gull

Woodcock

Snipe

Corncrake

Plover sp.

Godwit

Woodpigeon

Crow

Blackbird

Passerines

TOTAL

205

577

1,151

2008), and at Caherlehillan, Co. Kerry (McCarthy, in prep.) as well as at the coastal ringfort of Rathgurreen in Co. Galway (Hamilton-Dyer 2000a).

Conclusions This report examined the extent to which the excavated animal, avian and fish bones from High Island could provide information on the subsistence strategies of the early monastic community as well as that of the later occupants and more transient visitors to the site. The island was ideally suited for settlement as it provided good grazing, a

the environmental reports 261

supply of fresh water, abundant marine resources and colonies of seasonally available seabirds. The presence of a horizontal watermill, with associated ponds, mill races and a water reservoir, provides some insight into the character of the settlement. The indications are that the island supported a large monastic community that grew from its incipient beginnings, possibly in the mid–late 7th century or some time in the 8th century, to the time it was probably abandoned by the monastic community in the mid-12th to early 13th century. The data from the animal bones suggest a mixed farming economy supplemented seasonally by marine resources and seabirds. The animal management strategy predominantly involved sheep husbandry, which fits well within a regime of mixed farming, possibly involving cereal production. The greater occurrence of sheep over cattle has been noted as a common husbandry practice on marine islands, where the marginality of the terrain was less well suited to the dietary requirements of cattle (Murray and McCormick 2005). The ecclesiastic community on High Island seems to have been independent of the mainland for its food supply and the management of livestock animals was undertaken by the monks as part of a self-sufficient subsistence strategy that included breeding sheep and cattle primarily for their meat and possibly also cereal cultivation. The abundant marine life surrounding the island was also exploited and finds of fish, seabirds and shellfish attest to the exploitation of these local resources. The strict dietary requirements that would have been imposed on the monastic community by the Church led to a higher rate of consumption of marine resources, a situation paralleled at other island monastic sites around the Atlantic seaboard.

9.2 Non-wood plant macro-remains Meriel McClatchie

Introduction

Soil samples from 37 contexts were presented for archaeobotanical analysis, 25 of which were found to produce charred non-wood plant macro-remains. Most of the plant remains were recovered from locations in the vicinity of the extended monastic enclosure wall and Cell A, with smaller quantities being found around the church and in other areas. Cereal grains and chaff were regularly recorded in small quantities, with occasional deposits producing relatively large numbers of grains. Barley was predominant, including the six-row hulled variety, with oat also present. Small numbers of weed seeds were recovered from a number of deposits, with bracken and acorn fragments also recorded. Analysis of the plant remains from High Island has demonstrated that the island’s inhabitants had access to cereals and other plants over several centuries. While

the location of cultivation is unconfirmed, it does appear that crops were at least processed on the island during the earlier phases of activity, rather than being introduced as grain that had been processed elsewhere. The excavated remains have been divided into four phases: Phase 1 (8th to mid-11th century) and Phase 2 (mid-11th to late 12th/early 13th century) relate largely to the early medieval period (5th–12th centuries), with Phase 2 being the most active period of occupation at the monastery; Phase 3 (late 12th/early 13th to mid-15th century) and Phase 4 (mid-15th to late 20th century) relate, for the most part, to the later medieval (12th to 16th centuries) and post-medieval (post c. 1550–1600) periods. This report will detail the types and locations of plant remains recorded in each phase, in addition to interpreting the remains in their wider context. Tables 9.14–9.16 provide a quantified list detailing the plant remains present.

Methodology

The soil samples were processed using conventional flotation methods, with the smallest sieve mesh-aperture measuring 300µm. Identification of the archaeobotanical material in all samples was carried out using a stereomicroscope, with magnifications ranging from x6.3 to x50. Each sample was scanned in order to confirm the presence of archaeobotanical material, which was then extracted and sorted into general groupings on the basis of visual comparison of morphological features. The archaeobotanical material was identified by comparison to reference material in the collection of modern diaspores held at the Department of Archaeology, University College Cork, as well as the drawings and photographs from various seed keys (Anderberg 1994; Beijerinck 1947; Berggren 1969; 1981; Katz et al. 1965). Cereal grains were recorded as ‘whole’ if the embryo was present. Some of the material was distorted or fragmented, and was identified to genus or family level only. The identified taxa are listed in Tables 9.14–9.16. With the exception of cereals, botanical names are listed following the order and nomenclature of Flora Europaea (Tutin et al. 1964–80), and common names follow those provided in New flora of the British Isles (Stace 1991). The seeds, achenes and utricles of plants are referred to as ‘seeds’ throughout the text for convenience (see Tables 9.14–9.16 for botanical names).

Non-wood plant macro-remains recorded Phase 1: 8th to mid-11th centuries

Seventeen contexts associated within Phase 1 were presented for archaeobotanical analysis. Non-wood plant macro-remains were absent from eight contexts (F38/39; F93g; F93j; F97a; F216; F221; F246; F347). Plant remains were present in the remaining nine deposits (Table 9.11).

262 high island: excavation of an early medieval monastery

F42 produced a small number of grains of Hordeum vulgare subsp. vulgare (six-row hulled barley) and Hordeum vulgare L. (hulled barley), as well as Cerealia (indeterminate cereal) culm node fragments, seeds of Chenopodium/ Atriplex sp. (goosefoot/orache) and a seed of Plantago lanceolata L. (ribwort plantain). Species of the goosefoot and orache genera can be found growing in a variety of environments, including disturbed ground around settlements and in arable fields. Ribwort plantain can be found growing in grassy places and meadows, and on cultivated and waste land. Two charcoal-rich deposits (F56; F68) associated with the pre-church levels in the area of pseudo Graves 6 and 7 produced a small quantity of plant remains, including grains of six-row hulled barley, hulled barley and Avena sp. (oat), as well as a seed of goosefoot/ orache. Redeposited fill (F69) within Grave 3 contained a seed of Stellaria media L. Vill. (chickweed), which represents a plant that can be found growing in cultivated, waste and bare places. A deposit of burnt debris (F93i) produced a small number of barley and oat grains, as well as possible Quercus sp. (acorn) fragments and a frond fragment of Pteridium aquilinum L. Kuhn (bracken). Bracken can often be found growing in woods and on heaths, usually being associated with dry, acid soils rather than calcareous soils. Acorns could have been collected, perhaps locally, in order to provide food for both humans and animals. Three deposits (F105; F232; F267) in the vicinity of the pre-church ‘paved area’ (F245) produced occasional cereal chaff fragments – including a culm (stem) base fragment and culm fragments of indeterminate cereal – as well as a seed of Silene vulgaris Garcke (bladder campion) and a fragment of possible acorn. Bladder campion can be found growing in grassy places and on open and rough ground, generally on dry, calcareous soils. A fill deposit (F4052) associated with the primary phase of the monastic enclosure wall (F4040) produced a single indeterminate seed.

Phase 2: mid-11th century to late 12th/early 13th century

Ten contexts associated with Phase 2 were presented for archaeobotanical analysis, all of which contained nonwood plant macro-remains (Table 9.12). Redeposited grave-fill deposits (F25a; F46) produced a barley grain, as well as seeds of Atriplex patula L. (common orache) and Silene sp. (campion). Common orache can be found growing on disturbed and waste ground, while species of the campion genus can be found growing in a range of environments. Burnt debris (F319; F335) within Cell B produced a small number of barley grains, a cereal culm node fragment and a bracken frond fragment. Four of the examined deposits were associated with the extended monastic enclosure wall. Burnt debris (F457) located within the wall contained a small number of seeds of common orache and Polygonum persicaria L. (redshank). Common orache can be found growing on disturbed and waste ground, while redshank can be found growing on cultivated, waste and bare ground. Three deposits (F478; F484; F4079) in the vicinity of the south face of the wall produced the largest quantity and widest range of remains from Phase 2. These three deposits contained grains of sixrow hulled barley, hulled barley and oat, as well as cereal culm and culm node fragments. Seeds of bladder campion, goosefoot/orache, Rumex acetosella agg. (sheep’s sorrel), Rumex acetosa L. (common sorrel) and Gramineae (grass) were also recorded. Sheep’s sorrel can be found growing on meadows, bare ground and acid soils, while common sorrel can be found growing in a range of habitats, including grassy places. A deposit of burnt debris (F515) located beneath the paved floor within Cell A produced a small number of barley and oat grains, as well as a seed of Chenopodium album L. (fat-hen). Fat-hen can be found growing on cultivated, waste and bare ground. A single barley grain was recovered from the silt layer (F703) within the wall chamber.

Table 9.11 Phase 1 contexts containing non-wood plant macro-remains.

Context

Interpretation

Charcoal-rich deposit within pre-church level of Grave 1

Charcoal-rich deposit within pre-church area of pseudo Grave 6

Charcoal-rich deposit within pre-church area of pseudo Grave 7

Grave-fill within pre-church phase of Grave 3

93i

Burnt debris retained between wall (F91) and the north wall of church enclosure

105

Charcoal-rich deposit pre-dating church

232

Charred vegetation, possibly ‘old ground surface’

267

Burnt debris in vicinity of ‘paved area’ (F245)

4052

Silt deposit within primary phase of monastic enclosure wall (F4040)

the environmental reports 263

Table 9.12 Phase 2 contexts containing non-wood plant macro-remains.

Context

Interpretation

25a

Uppermost fill layer within pseudo Graves 6 and 7

Fill of pseudo Grave 2

319

Burnt debris over hearth floor within Cell B

335

Burnt debris within Cell B, fill of pit (F341)

457

Burnt debris within extended monastic enclosure wall (F4025)

478

Charred deposit outside extended monastic enclosure wall above F484/4079

484

Deposit sealed beneath south face of extended monastic enclosure wall outside Cell A

515

Burnt debris beneath paved floor within Cell A

703

Silt layer within wall chamber

4079

Deposit sealed beneath south face of extended monastic enclosure wall outside Cell A

Phase 3: late 12th/early 13th to mid-15th century

Phase 3 represents a period of possible reoccupation after the decline of the monastery. Six contexts associated with Phase 3 were presented for archaeobotanical analysis (Table 9.13), all of which contained non-wood plant macro-remains. Most of the remains recorded in Phase 3 were recovered from burnt debris (F509) located within Cell A. Approximately 500 whole components were identified, representing one-eighth of the whole sample, which was deemed representative. Barley and six-row hulled barley grains were predominant, with a smaller quantity of oat grains also being recorded. These cereal grains were radiocarbon dated to a period between the late 13th and early 15th century (cal. AD 1287–1424 (UB-4988, 580±40 BP)). Cereal chaff and weed seeds were absent from this deposit. Two fills (F445; F464) of a disused water-collection and drainage channel to the east of Cell A produced a much smaller quantity of six-row hulled barley, barley and oat grains. Burnt debris (F472) to the north of Cell A contained a barley grain and an indeterminate cereal grain fragment. Two deposits in the vicinity of the church also produced small quantities of weed seeds. A deposit

(F4) contained common orache seeds, while a fill (F214b) of pseudo Grave 10 produced bladder campion and goosefoot/orache seeds. Common orache can be found growing on disturbed and waste ground, and bladder campion can be found growing in grassy places and on open and rough ground, generally on dry, calcareous soils.

Phase 4: mid-15th century to late 20th century Phase 4 represents intermittent later medieval and postmedieval activities at High Island. One deposit (F474) associated within Phase 4 was presented for archaeobotanical analysis. This deposit did not contain any non-wood plant macro-remains.

Discussion Cereal remains The recovery of cereal and associated remains from a range of archaeological deposits at High Island indicates that cultivated crops were being used by the island’s inhabitants over several centuries. The presence of an early horizontal water-mill at High Island further underlines the importance of cereals at this location. Construction of the mill,

Table 9.13 Phase 3 contexts containing non-wood plant macro-remains.

Context

Interpretation

Deposit located above paved surface (F29), between church and enclosure

214b

Fill of pseudo Grave 10 (F214) located in north-west corner of church interior

445

Fill of disused water-collection channel (F8025) in vicinity of Cell A

464

Fill of disused water-collection channel (F8025) in vicinity of Cell A

472

Burnt debris located north of extended monastic wall (F4025), partly sealed by east wall of subrectangular building (F427)

509

Burnt debris on paved floor (F510) within Cell A

264 high island: excavation of an early medieval monastery

which was accompanied by ponds, a water reservoir and mill-races, was an enormous undertaking, and had the potential to provide food-production facilities for a sizeable population (Rynne 2000). It has been suggested that the mill may have been constructed at some date between the mid-9th and mid-10th century (White Marshall and Rourke 2000, 55). The date of abandonment of the mill is unknown, but it is probable that the mill was still in use during Phase 2 (mid-11th to late 12th/early 13th century), which was the main period of monastic activity recorded during excavations at High Island. Further evidence for the use of cereals at this location may be reflected by the recovery of a rotary-quern fragment (95E124:10:1) in the masonry of the east wall of the church. Two dished grinding-stones (95E124:4034:1 Phase 1; 95E124:450:1 Phase 4) were also recorded during excavations. While querns can be used to grind and reduce a range of materials, both rotary and dished grinding stones are most often associated with the processing of cereal grains (Connolly 1994a; Kelly 1998, 245). A review by EMAP (O’Sullivan et al. 2010, 47–8) of ‘special’ deposits in early medieval structures has noted the deliberate incorporation of quern stones at a number of other early medieval structures, such as at Leacanabuaile and Dressogagh, where broken rotary quern stones were placed into roundhouse walls and wall slots, respectively, while at Rinnaraw, Drumaroad and ‘the Spectacles’, broken quernstones appear to have been deliberately deposited near the doorway thresholds of early medieval houses. It is suggested that the placing of querns within structural elements may represent the marking of special events, such as the establishment or abandonment of structures. Such actions underline the importance of social aspects with regard to crop-related activities in early medieval communities. The importance of cereals during the early medieval period is reflected in the extensive range of evidence for arable activities at this time. Cereal remains have regularly been recorded in archaeological deposits of this period (Monk 1991; Monk et al. 1998), and tools associated with soil preparation, harvesting and food processing are often recovered (for example, Hencken 1936 and 1942; Collins 1955; Farrell et al. 1989; Rynne 1998; Monk and Kelleher 2005). The cultivation, harvest and trade of cereal products are frequently referred to in contemporary documentary sources (for example, AL i 125.22–27; AL iv 309.5–6, 307.7–8; AL v 167.10–12, 167.25–27, 393.12–13, 485.21; Kelly 1998, 320, 332, 468, 482–5; Rynne 1998, 87; Hancock et al. 2000). The medieval period in Ireland also provides regular evidence for crop production and consumption (e.g. Monk 1987a; 1987b; 1987c; McClatchie 2003; Murphy and Potterton 2005).

Types of cereal remains recorded The predominance of barley, particularly six-row hulled barley, and oat at High Island, with wheat and rye being absent, reflects a more general trend that can be detected at other early medieval sites throughout Ireland (Monk 1991, 317; Monk et al. 1998, 72). The cultivation of barley is relatively low-risk, as barley will yield at least part of its crop even after a poor season. Barley will also grow equally well on light and heavy soils. Oat grains, rather than chaff, were recorded at High Island. Oat chaff (e.g. glume bases) is required for the identification of this cereal type to species level. The oat recorded at High Island cannot, therefore, be identified as being of the wild or cultivated varieties; it is not known if oat was occurring as a weed of the barley crop or if oat was being cultivated as a crop in its own right. It should be noted, however, that there is much evidence for the cultivation of oat at this time (Kelly 1998, 226–7). Oat is well suited to the humid and wet Irish climate and will tolerate poorer soils that may have discouraged the cultivation of other cereal types. Contemporary documentary evidence also suggests that barley and oat would have been more commonly encountered during the early medieval period. A list of cereal types provided in the 8th-century law tract Bretha Déin Chécht (Kelly 1998, 219) places cereal types in the following order: bread wheat, rye, possible spelt wheat, possible two-row barley, possible emmer wheat, six-row barley and oat. This order represents the relative prestige of each type of cereal, which is correlated with a particular ranking in society. Bread wheat is equated with the rank of a superior king, bishop or chief poet whereas, at the other end of the scale, oat is equated with the commoner (Kelly 1998, 219). Barley and oat are more regularly recovered from early medieval archaeological deposits in Ireland, reflecting their lower status (Monk 1991). Bread wheat and rye are rarer in the archaeological record, as their status was perceived as being higher. Cereals were therefore regarded not just as a source of sustenance but also as cultural symbols that could distinguish social classes (Fredengren et al. 2004). The cereal types regularly recorded at High Island are representative of the more common cereals that would have been available during this period, with no evidence for ‘high-status’ cereals. Cereal remains have previously been recovered from a number of other early medieval ecclesiastical island sites, such as Church Island, Co. Kerry (Scannell 1958), and Illaunloughan, Co. Kerry (Murray and McCormick 2005). At Church Island, plant remains were recovered from floor and associated deposits in the stone oratory and round stone house. A range of cereal types was identified, including wheat, possible rye, possible barley and oat grains. At Illaunloughan, oat was predominant in the large cereal assemblage, followed by barley, while wheat was

the environmental reports 265

Table 9.14 Non-wood plant macro-remains from Phase 1. BOTANICAL NAME

HYPOLEPIDACEAE Pteridium aquilinum (L.) Kuhn FAGACEAE cf. Quercus sp. CHENOPODIACEAE Chenopodium/Atriplex sp. Chenopodium/Atriplex sp.

frond fragments nut fragments utricles utricle fragments seeds seeds seeds

CARYOPHYLLACEAE Stellaria media (L.) Vill. Silene vulgaris Garcke PLANTAGINACEAE Plantago lanceolata L. GRAMINEAE Hordeum vulgare subsp. vulgare Hordeum vulgare L. Hordeum vulgare L. Avena spp. Cerealia Cerealia Cerealia INDETERMINATE Indeterminate

PLANT PART

grains grains grains grains culm base fragments culm node fragments culm fragments seed fragment

Phase Feature number Sample number COMMON NAME bracken

1 42 ’96G

1 56 ’96Y

1 68 ’97E

1 69

1 93i

1 105 ’01B

1 232

1 267 ’98AQ

1 4052 ’02A

possible acorn goosefoot/orache goosefoot/orache

13 8

1 -

10 -

1 -

chickweed bladder campion ribwort plantain six-row hulled barley hulled barley barley oat indeterminate cereal indeterminate cereal indeterminate cereal indeterminate Total

1 -

2 -

1 1

1 2

1 1

recorded in just a small number of samples. Recent excavations at Skellig Michael have also uncovered evidence for small quantities of cereal remains, including oat and barley (Bourke et al. 2011). During the earlier part of the medieval period, barley and oat continued to be important, particularly in areas beyond Anglo-Norman control (McClatchie 2003). A variety of other Irish medieval rural sites, such as Ballyveelish (Monk 1987a) and Drumlummin (1987c), both in Co. Tipperary, have produced evidence for barley and oat remains.

Cereal products Cereals would have been cultivated for their food value; grains could have been used in a range of food products, including breads, gruels and porridges, as well as in brewing and as animal fodder (Kelly 1998, 82–3, 220, 226, 245, 330, 332–5; Sexton 1998). Cereal straw would have been utilised in structures, roofing and bedding (Kelly 1998, 111, 240). Cereals could also have been used as a form of currency, for example in food-rent (op. cit., 219, 333).

Documentary sources of the early medieval period regularly list bread as one of the main foods consumed in monastic diets, and where cereal type is mentioned, the bread is often made of barley (Kelly 1998, 219; Sexton 1998, 82; Murray et al. 2004, 180). Barley would have produced a coarse, dense loaf (Sexton 1998, 82), in keeping with the penitential existence associated with monasteries. Tigernach of Clones, Co. Monaghan, for example, is said to have lived exclusively on barley bread, water-cress and hot water (Kelly 1998, 344). It should be noted, however, that documentary sources were often concerned with issues of fasting and dietary penance; it is possible that extreme eating regimes, like Tigernach’s mentioned above, were not continually observed at all monasteries (Murray et al. 2004, 179–81; White Marshall and Walsh 2005, 130).

Cereal chaff and arable weeds Cereal chaff (culms) was recorded in both phases associated with the monastic occupation (Phase 1 and Phase 2), but was absent from Phase 3 deposits. Weeds that can be found

266 high island: excavation of an early medieval monastery Table 9.15 Non-wood plant macro-remains from Phase 2.

Phase Feature number

2 25a

2 46

Sample number

PLANT PART frond fragments achenes achenes achenes utricles utricles

COMMON NAME

bracken

redshank sheep’s sorrel common sorrel fat-hen common orache goosefoot/ orache goosefoot/ orache bladder campion campion six-row hulled barley hulled barley barley

BOTANICAL NAME HYPOLEPIDACEAE Pteridium aquilinum (L.) Kuhn POLYGONACEAE Polygonum persicaria L. Rumex acetosella agg. Rumex acetosa L. CHENOPODIACEAE Chenopodium album L. Atriplex patula L. Chenopodium/Atriplex sp. Chenopodium/Atriplex sp. CARYOPHYLLACEAE

utricles utricle fragments

Silene vulgaris Garcke

seeds

Silene sp. GRAMINEAE Hordeum vulgare subsp. vulgare Hordeum vulgare L. Hordeum vulgare L.

seeds

Hordeum vulgare L. Avena spp. Cerealia Cerealia Cerealia Gramineae

grains grains grains grain fragments grains grain fragments culm node fragments culm fragments grains

2 515

2 703

’02ii

’02K

1 2

1 -

16 -

1 -

1 2

4 6

4 2

1 -

barley

oat indeterminate cereal

1 12

indeterminate cereal indeterminate cereal grass Total

growing in arable fields, including fat-hen, redshank and chickweed, were also represented in many of these deposits. These weeds may have been inadvertently harvested along with the cereal crop. The presence/absence of cereal chaff and arable weed seeds in archaeological deposits can indicate the stage of processing that a crop had reached, with the final stage being represented by an absence or very small percentage of ‘contaminants’, such as cereal chaff and weed seeds. The processing of crops can only be performed using a restricted number of techniques in non-mechanised societies, as the morphological attributes of various crop types determine the manner and sequence of processes applied (Hillman 1981; 1984). The stage in processing

2 2 2 2 2 2 319 335 457 478 484 484 ’98F, ’00AJ ’02H ’98G

2 2 4079 4079

that a crop had reached when deposited can therefore be ascertained through the presence or absence of various elements. The sequence of processing for free-threshing cereals, which include barley and oat, begins with harvesting, which removes the crop from the plot in which it was growing. The next stage is threshing, which releases grains from straw and chaff. Raking, in order to eliminate coarse straw and weeds, follows this stage. Winnowing will then remove light chaff and straw fragments, in addition to light weed seeds and heads. The next stage is coarse sieving, which eliminates large weed seeds, weed heads, unthreshed ears and straw nodes. Fine sieving is the next stage of processing and will remove small weed seeds. Glume wheats, which were absent from deposits

the environmental reports 267

Table 9.16 Non-wood plant macro-remains from Phase 3. BOTANICAL NAME CHENOPODIACEAE Atriplex patula L. Chenopodium/Atriplex sp. CARYOPHYLLACEAE Silene vulgaris Garcke GRAMINEAE Hordeum vulgare subsp. vulgare Hordeum vulgare L. Avena spp. Cerealia

PLANT PART utricles utricles seeds grains

Phase Feature number Sample number COMMON NAME common orache goosefoot/orache bladder campion six-row hulled barley

3 4 3

3 214b ’96H 1 1

3 445

grains grains grain fragments

barley oat indeterminate cereal Total

1 5 6

at High Island, require additional stages of processing to release grains due to their enclosure by glumes. As a result of the damp Irish climate, cereal grains may have been dried before winnowing or storage, perhaps in a kiln constructed for this purpose (Monk and Kelleher 2005), although evidence for a formal kiln was not recorded at High Island. Culm (straw) fragments were found to outnumber cereal grains in a number of the smaller Phase 1 and Phase 2 assemblages, suggesting that these deposits may represent chaff residues from the earlier stages of cropprocessing activities, as outlined above. This evidence suggests that at various stages during Phases 1 and 2, cropprocessing is likely to have been carried out on the island. The absence of cereal chaff and the very small quantity of weed seeds recorded from Phase 3 deposits suggests that these later assemblages were almost fully processed. During Phase 3, crop-processing may have taken place at another location on the island, where the residues (chaff and arable weed seeds) were not preserved as they did not come into contact with fire. It should also be noted that charred plant material is subject to differential preservation of plant components, whereby cereal chaff is less robust than cereal grain and will be destroyed at a lower temperature when charred (Boardman and Jones 1990). Alternatively, the Phase 3 deposits could represent crops that were introduced to the island in an already processed form, indicating a change in crop-related activities following the decline of the monastery and possibly representing a period when the mill had gone out of use. Early medieval documentary sources often note that crops were cut high on the stalk, just removing the ear (Kelly 1998, 238). There is also a reference, however, to a corn-rick in an 8th-century law text, which implies that the crops were sometimes harvested in sheaves, and there are further references to this practice in later texts

3 464 7

3 472

27 34

1 1 2

3 509 121 390 14 72 597

(Kelly 1998, 239). Early medieval written sources emphasise reeds for thatch and rushes for floor-covering, rather than cereal straw, but there seems to be an increased use of cereal straw in later periods, being used in roofing, as tinder, fodder and bedding (Kelly 1998, 240). The presence of a culm base fragment in Phase 1 suggests that at least some of the crop was being uprooted, rather than being harvested high on the stalk with a sickle. Collection of straw in Phase 1 is also suggested by the presence of a number of low-growing arable weeds, e.g. chickweed, which would not enter an assemblage if only ears were being harvested. It appears, therefore, that straw as well as grains may have been harvested for use at High Island.

Locations of cereal remains Most of the cereal remains recorded in deposits dating to the monastic period at High Island were clustered in Phase 2 deposits in the vicinity of Cell A and the nearby extended monastic enclosure wall. This evidence suggests that these areas may have been a focus of food preparation activities at this time. Cereal remains were present, although less well represented, in areas directly associated with the church. Interestingly, cereal remains were also scarce at Cell B, which was interpreted as the monastic refectory. This may indicate that Cell B was more often used for food consumption rather than preparation, as there appear to have been fewer opportunities for cereals coming into contact with fire at this location and thereby being preserved in the archaeological record. At Illaunloughan, food remains were similarly clustered in certain areas, suggesting that some activities were limited to specific locations (White Marshall and Walsh 2005, 128). At High Island, it appears that boundaries, whether physical or conceptual, may also have limited the locations of various activities, including food production. It appears that efforts were made to restrict the debris of food-related activities to ‘domestic’ areas such as

268 high island: excavation of an early medieval monastery

Cell A, perhaps in order to create a boundary between the ‘domestic’ and more obviously ‘religious’ areas at the monastery.

Comparison of plant macro-remains and pollen records Analysis of pollen remains was previously carried out at High Island, indicating that rye was the predominant crop during the early medieval and later periods, with much less evidence for barley and wheat being recorded (Molloy et al. 2000, 240). The pollen evidence does not correlate well with the plant macro-remains evidence, as rye is absent from the latter. Rye is wind-pollinated and is known to disperse pollen more freely than other cereals; barley, oat and wheat are self-pollinating plants that generally produce relatively low amounts of pollen, which is poorly dispersed (Molloy et al. 2000, 241; Bakels 2000). It appears, therefore, that rye may be over-represented in the pollen record, due to the nature of its pollen dispersal. This does not fully explain the absence of rye from the plant-macro remains record at High Island, however. It is possible that rye was present at High Island but did not enter deposits associated with the excavated areas, due to the differential use of crops in various locations. It has been noted above that rye was considered a ‘high-status’ crop in early medieval Ireland, and perhaps the monastic community resident within the church enclosure did not feel that the consumption of rye was appropriate in this location. It has also been noted above that plant food remains were largely restricted to certain areas, and perhaps restrictions also applied to the types of crop that could be consumed within these areas. Rye could instead have been consumed on other parts of the island, beyond the monastery, or could even have been introduced to the island for milling and then removed again to the mainland. This in turn would suggest that a wider community, perhaps from the mainland, who consumed a wider range of cereal types, also had access to the mill.

Location(s) of cereal cultivation In their analysis of the pollen remains recorded at High Island, Molloy et al. noted that the thin soil cover, rocky terrain and lack of shelter from wind and salt spray would have provided less than optimal conditions for cereal cultivation on High Island (2000, 242). They do, however, interpret the presence of rye, barley and wheat in the pollen record as representing cultivation, perhaps small-scale, on the island. This small-scale cultivation may have been supplemented by the importation of further crops from the mainland (Molloy et al. 2000, 242), which would have necessitated only occasional trips to and from the mainland, as crops are easily storable. The cultivation of crops at High Island would cer-

tainly have been challenging, but Rynne (2000) has noted that the people who built the mill were technologically advanced. It is possible that the island’s inhabitants also introduced innovative farming methods that would have enabled cultivation in such a ‘marginal’ location. Indeed, early medieval monasteries have frequently been considered instrumental in large-scale land improvement or reclamation (Ryan 2000, 35). Furthermore, plant cultivation was particularly associated with early medieval monasteries; the gardener is listed among the seven officers of the church (Kelly 1998, 250). It is possible, however, that the pollen remains of cereals represent the processing, rather than cultivation, of crops at High Island. Crop-processing can release significant quantities of pollen (Behre 1981; Hall 1988), and the plant macro-remains evidence suggests that crop processing was being carried out on the island. It is therefore possible that crops were being imported onto the island, where they were then processed, with the pollen being released at this time. Furthermore, weeds that are usually associated with acid soils were found in association with some cereal deposits at High Island, while weeds that are more often found growing on calcareous soils were found in other cereal deposits. This evidence indicates that the crops derived from a variety of locations, suggesting that at least some of the crops brought from the mainland to High Island.

Non-cultivated remains Many plants that would today be considered economic pests could have been seen as useful by past societies, including the inhabitants of High Island. Some of the ‘weeds’ recorded at High Island may have been utilised in foodstuffs, as well as in activities such as dyeing. There is also special emphasis in the Irish law-texts dating to the 7th and 8th centuries on the growing of plants with medicinal properties (Kelly 1998, 250). The presence of medicinal plants today in the environs of Irish and Welsh monastic sites has led some commentators to suggest that these plants may represent relicts of former cultivation (Synnott 1979; Connolly 1994b). In terms of foodstuffs, early medieval and medieval documentary sources contain references to the consumption of sorrel (Kelly 1998, 311; Sexton 1998, 83). Gerard’s medieval Herball described how common sorrel provided a ‘profitable sauce in many meats’ and was ‘pleasant to the taste’ (Gerard and Johnson 1633, 398). The recovery of redshank seeds in Viking Dublin skeletal deposits led Mitchell (1987, 23–4) to suggest that this species may have been intentionally consumed. Sheep’s sorrel is said to have been used by the Irish to flavour fish (Moloney 1919, 39). Chickweed is often mentioned as both an edible and medicinal plant in documentary sources, where, for

the environmental reports 269

example, it is noted as providing relief against inflammations (Keogh 1735, 29; Salmon 1710, 180). It appears that fat-hen seeds were collected for consumption since prehistoric times in Europe, and a recent study has even explored the possibility that fat-hen was cultivated during the Iron Age (Stokes and Rowley-Conway 2002). Sorrel is among the plants that may have been used in dyeing (Moloney 1919, 39), and it has been noted that processing of bracken can also provide a yellow-green dye (Kelly 1998, 263). Interestingly, the highest concentration of sorrel remains at High Island (from Phase 2 hearth F4079) was located in an area that also contained the highest concentration of hones, stone discs and tools of unknown function. Excavations on the island ecclesiastic site of Inishkea North, Co. Mayo, previously uncovered evidence for dyeing in the form of dogwhelk shell remains (Henry 1952). The presence of sorrel remains at High Island is a less convincing indicator of dyeing, but the potential for such an activity at this location should be considered. As well as being used as a dye-plant, early Irish law-texts also noted the economic importance of bracken as tinder (Kelly 1998, 338, 380–81), and the presence of bracken in medieval deposits at Drogheda was suggested to have represented its use in animal bedding (Dickson and Mitchell 1984, 219). The presence of acorns in High Island deposits may represent the use of this resource as fodder for pigs, as described in early medieval sources (Kelly 1998, 83). Acorns may also have been used as a flour substitute, particularly in times of bad harvests and famines (Sexton 1998, 81). Although there is little evidence for oak at High Island in the pollen record at this time (Molloy et al. 2000, 239), oak was identified in the majority of archaeological deposits examined for their charcoal content, perhaps suggesting the acorns were gathered from a local source (G. Scally, pers. comm.). Cultivated plants, rather than gathered plants, are, however, predominant at High Island. It should be remembered that preservation by charring is biased in favour of plants that are more likely to come into contact with fire. Cereals may have been exposed to fire during the drying of crops and cooking activities, and such plants are therefore more likely to be represented in charred assemblages when compared with plants that are more often eaten raw or boiled, such as fruits and vegetables. Many vegetables will also be consumed before they set seed, which means that they are unlikely to be represented in the archaeological record. It is therefore probable that the inhabitants of High Island would have made use of an even wider range of plants than that represented in the examined deposits. Early medieval documentary sources mention the consumption of a range of vegetables at this time, including onions and cabbage (Kelly 1998, 248, 255), as well as fruits

such as apples, plums and berries (Kelly 1998, 259, 261, 306–7). Other wild resources could also have included a greater variety of nuts and seaweed (Kelly 1998, 304, 306).

Conclusions

The analysis of archaeobotanical material from excavations at High Island has provided an insight into food choices over several centuries at this site. The types of foods consumed are likely to be associated with social choices, as well as economic concerns and environmental constraints. In terms of cereals, barley (including the six-row hulled variety) was predominant, with oat also recorded. This focus on barley does not appear to have changed throughout Phases 1 to 3. According to early medieval sources, the consumption of barley, particularly in bread form, is often associated with monasteries in Ireland. The pollen record provides evidence for rye at High Island, although this cereal type is not represented in the macro-remains record. Cultivation of cereal crops at High Island has not been confirmed, although the presence of cereal chaff (culms) and arable weeds indicates that at least the processing of crops is likely to have been carried out at this location. Other potential plant resources were also recorded, including ‘weeds’ that have a number of potential uses, as well as bracken and possible acorns. Fruits were absent from the archaeobotanical record. It should be remembered, however, that preservation by charring is biased in favour of plants that are likely to come into contact with fire, such as cereals. It is therefore probable that the inhabitants at High Island would have made use of a far wider range of plants than that represented in the examined deposits.

9.3 The charcoal remains Rowena Gale

Introduction

An extensive programme of environmental sampling was undertaken during the excavations on High Island. Charcoal was recovered from a large number of bulk soil samples and, in addition, fragments were hand-collected, producing a total of 105 samples for analysis. The samples were taken from both in situ and redeposited burnt debris, most of which represented the residues of domestic hearths, and from the fill of the smithing hearth pit in the east of the monastic enclosure beneath Cell B. In addition to the use of wood fuel, burnt peat was also relatively common in the samples. Species identification was undertaken to indicate the character of the fuels used and their sources of supply, and also to obtain environmental data relating to the vegetation resources on the island. Secure dates for features uncovered in the excavations were

270â&#x20AC;&#x192; high island: excavation of an early medieval monastery Table 9.17â&#x20AC;&#x201A; Charcoal from Phase 1. (Key: h = heartwood; r = roundwood (diameter <20mm); (r) = includes roundwood (diameter <20mm); s = sapwood (diameter undetermined, Quercus only). The number of fragments identified is indicated.) Feature Alnus Betula Corylus Ericaceae Fraxinus Ilex Pomoideae Prunus Quercus Salix/ Juniperus Picea/ Taxus no. alder birch hazel heather ash holly hawthorn spinosa oak Populus juniper Larix yew spruce/ group blackthorn willow/ larch poplar Phase 1: 8th century to mid-11th century 42

93 a

93 b

cf. 1

93 c

8h, 17s

93 d

3h, 2s

93 e

93 g

2h, 2s

93 h

93 i

21r

12r

15h

93 j

93 m

25r

101

105

216

221

44 (r)

10h, r, 3s

232

cf.1

250

cf. 1

262

265

346

3h,2r

347/351

357h,2r

4034

4036

4037

89 (r)

32h, 1s

4038

36h, 21s

4039

2h, 1s

4051

4052

4055

4072

705

?1r

the environmental reports 271

obtained through radiocarbon dating by selecting charcoal from short-lived species.

Methodology

Bulk soil samples were processed by flotation and they were sieved using tiered sieves (2mm, 1mm and 0.5mm meshes). Intact segments of narrow roundwood were relatively frequent. Charcoal fragments measuring greater than 2mm in radial cross-section were considered for species identification. The samples from the fill of the smithing hearth pit were 50% sub-sampled prior to examination (F347; F351). The condition of the charcoal varied from firm and well preserved to poor and friable. The samples were prepared using standard methods (Gale and Cutler 2000, 12–13). The anatomical structures were examined using incident light on a Nikon Labophot-2 compound microscope at up to x400 magnification, and they were matched to prepared reference slides of modern wood. When possible, the maturity of the wood was assessed (heartwood or sapwood) and stem diameters and the number of growth rings recorded. It should be noted that charred stems may be reduced in volume by up to 40%.

Results

The taxa identified are presented in Tables 9.17–9.19. Tables 9.20–9.22 show the stem diameters and the ages of selected roundwood. Tables 9.23–9.25 provide a comprehensive list of the excavated features from which the samples were examined and their associated sample numbers and comments. Classification follows that of Flora Europaea (Tutin et al. 1964–80). Group names are given when anatomical differences between related genera are too slight to allow secure identification to genus level. These include members of the Pomoideae (Crataegus, Malus, Pyrus and Sorbus), Salicaceae (Salix and Populus) and Ericaceae (Calluna and Erica). When a genus is represented by a single species in the native flora, it is named as the most likely origin of the wood, given the provenance and the period, but it should be noted that it is rarely possible to name individual species from wood features. The anatomical structure of the charcoal was consistent with the following taxa or groups of taxa: 1. Aquifoliaceae, Ilex aquifolium L., holly; 2. Betulaceae, Alnus glutinosa L., Gaertner, European alder; Betula sp., birch; Corylaceae, Corylus avellana L., hazel; Ericaceae, Erica sp. and Calluna vulgaris L., Hull, heathers and ling. Many members of the heather family are anatomically similar. 3. Fagaceae, Quercus sp., oak; 4. Oleaceae, Fraxinus excelsior L., ash; 5. Rosaceae, subfamilies: Pomoideae, which include Crataegus sp., hawthorn; Malus sp., apple; Pyrus sp., pear; Sorbus sp., rowan, service tree and whitebeam.

6. 7. 8.

These taxa are anatomically similar, one or more taxa may be represented in the charcoal; Prunoideae, Prunus spinosa L., blackthorn; Salicaceae, Salix sp., willow, or Populus sp., poplar. In most respects these taxa are anatomically similar; Cupressaceae, Junipererus communis L., juniper; Pinaceae, Picea sp., spruce, or Larix sp., larch. These taxa are anatomically similar. Neither is native to Ireland; Taxaceae, Taxus baccata L., yew.

Discussion

In addition to deposits of burnt peat, the evidence indicated the common use of wood fuel during all phases of activity on High Island. The broad range of species identified indicates access to mixed woodland, some of which appears to have been coppiced. The following discussion examines these deposits in relation to context and phase and considers possible sources of fuel supply and the environmental implications.

The Iron Age

The silt deposit in the north of the site (F93L) produced the earliest dates from the excavations: 360–102 cal. BC (UB6452 (AMS), 2161±33 BP) and 387–170 cal. BC (UB-4986, 2201±38 BP). Although little is otherwise known regarding this Iron Age period of occupation, the identification of hazel, juniper and hawthorn (Sorbus group) gives some indication of the local woodland. Thirty-seven roundwood stems of fast-growth hazel (diameters 8–10mm), probably indicating coppice, were present. There were 42 roundwood stem fragments of juniper (diameters 2–10 mm) and 31 fragments of hawthorn (possibly narrow roundwood, but too comminuted to confirm). Herbaceous stems or roots (dicotyledon) were included. No evidence of further activity was uncovered in the excavations until the early medieval period.

Phase 1: 8th to mid-11th century

Thirty-nine samples date to this phase, which relates to the earliest period of monastic occupation. The charcoal associated with this phase was recovered from the church enclosure, the church, the interior of Cell B, the northern side of the monastic enclosure and the mural chamber in the south-west of the monastic enclosure (Fig. 5.1; Areas 1–4, 7). Roundwood was comparatively frequent in Phase 1 deposits, particularly hazel stems (Table 9.20). It is interesting also to note the presence of yew in these deposits and two of the three contexts in which it was found were in close proximity to the pre-church features and their immediate surroundings (Table 9.17; F42, F56). Although the resinous foliage is very flammable, the wood burns

272 high island: excavation of an early medieval monastery

slowly and does not make good firewood. However, the attractive reddish-brown wood was greatly valued for the production of artefacts. In Ireland, the yew tree was considered sacred and 8th-century Irish manuscripts refer to ‘noble artifacts’ made from its wood, for example, domestic vessels, croziers and shrines (Grigson 1958, 25–6; Edward Milner 1992, 43). Owing to the longevity of the tree and its evergreen foliage, yew has been associated with religion and ritual for millennia and its fronds may therefore have been significant at the monastery for specific festivals or for burials. Although no longer common, yew occurred widely throughout Ireland from early prehistoric times (Mitchell 1986, 165). It is feasible that, in common with many early medieval ecclesiastical sites, yew trees were planted beside the church or the structure that may have existed in this location before the church was built. Charcoal was recovered from deposits in the vicinity of the Phase 1 Graves 3, 4 and 5, in the east of the church enclosure. Associated charcoal was sparse in the layer of clay and schist to the south of the graves, but hazel, heather and yew were included (Table 9.17; F42). To the north of the graves, the brown soil containing lenses of charcoal and the dark brown clay included hazel, oak, blackthorn and yew (Table 9.17; F56, F68). Hearth ash, obtained from one of the samples taken from below the paved surface (F29) in the south-west of the church enclosure, included heather roundwood, oak, peat and other waste material consistent with domestic debris (Tables 9.17, 9.23; F75). Small deposits of oak and hawthorn (Sorbus group) occurred in the second sample of silty clay, with traces of burnt residues taken below the paving in the north-east of the church enclosure (Table 9.17; F105). A relatively large quantity of narrow ericaceous stems was recovered from the schist clay located between the primary north wall of the church enclosure and its secondary wall (Table 9.17; F97). The absence of other species could imply that this is debris from burning off local heather growth, perhaps to encourage new growth as fodder for livestock, or as the burnt remains of thatch or bedding. Several layers of redeposited ash and burnt debris were located over this surface. Since there was insufficient charcoal for dating from individual layers, a date of cal. AD 728–971 (UB-4522, 1171±41 BP) was obtained from combined samples. The charcoal content of these layers consisted mainly of hazel and oak, though birch, heather, holly, hawthorn (Sorbus group), willow or poplar and juniper were also identified (Table 9.17; F93, F93a–j, F93m). Peat and herbaceous stems were often present (Table 9.23). Hazel roundwood was frequent and the fast-growth stems were typically those of coppice growth (Table 9.20). Birch and oak were recorded from another layer of burnt debris in this location (Table 9.17; F96).

Two deposits which lay beneath the church contained quite substantial amounts of charcoal. An old ground surface, overlying the natural boulder clay, included charred vegetation. Associated charcoal consisted of narrow, twiggy material, including heather, oak, juniper and possibly birch (Table 9.17; F232). Charred stems from an unidentified herbaceous dicotyledon were also present (Table 9.23; F232). Charcoal from a post-hole, associated with the paved area (F245), was sparse, but birch, hazel and oak were included (Table 9.17; F250). Evidence for the use of peat fuel was also present (Table 9.23; F250). The wood fuel in the redeposited burnt debris, probably domestic hearth residues, that extended across the paved area and spread under the walls of the later church included birch, hazel, ash, oak, juniper and yew (Table 9.17; F221). This layer was sampled across the area within the church (for example, ’96Z was taken from the northeast corner and ’97AB from in front of the centre of the altar), but no significant spatial differences were indicated in the species identified. Radiocarbon dating using hazel charcoal provided a range of cal. AD 860–1020 (OxA8918, 1110±40 BP) and a sheep bone produced a date of cal. AD 770–980 (OxA-8946, 1182±36 BP). The hazel in this layer included four-year-old stems (approximately 10mm in diameter), some of which included the rapid growth consistent with that of coppice (Table 9.20; F221). Small fragments of oak and willow or poplar occurred in the layer of clay and schist that overlay the redeposited burnt debris (Table 9.17; F216). This feature may have been contemporary with the layer of grey clay from which a single piece of hand-collected willow or poplar was recovered (Table 9.17; F262). The fill of a stone-lined slot (F260), probably contemporaneous with the clay and schist layer, produced a single piece of oak (Table 9.17; F265). The smithing hearth pit pre-dated the construction of Cell B, the beehive hut in the east of the monastic enclosure. The charcoal-rich pit fill, associated with iron-working and the processing of metals, was dated by radiocarbon analysis from a sample of hazel charcoal at 2 sigma cal. AD 870–1030 (OxA-8917, 1100±40 BP). Of the 364 pieces of charcoal examined from the hearth pit, 359 proved to be oak, which were almost exclusively heartwood from slow-growth timber (Table 9.17; F347/F351). In addition, small quantities of hazel, holly and willow or poplar were identified. The heat source of the hearth was clearly fired with large chunks of oak. It is possible that the remaining taxa provided kindling or were infill in the charcoal-making process. Small fragments of oak, hazel and willow or poplar were present in the fill of the two small stake-holes at the smithing area (Table 9.17; F346). Prior to the use of coked-coal, charcoal was an essential ingredient of iron smelting and smithing (Forbes 1956,

the environmental reports â&#x20AC;&#x192; 273

Table 9.18â&#x20AC;&#x201A; Charcoal from Phase 2. (Key: h = heartwood; r = roundwood (diameter <20mm); (r) = includes roundwood (diameter <20mm); s = sapwood (diameter undetermined, Quercus only). The number of fragments identified is indicated.) Feature no.

Alnus Betula Corylus Ericaceae Fraxinus Ilex Pomoideae alder

birch

hazel

heather

ash

Prunus

holly hawthorn spinosa group blackthorn

Quercus

Salix/

Juniperus Picea/ Taxus

oak

Populus willow/ poplar

juniper

Larix yew spruce/ larch

Phase 2: mid-11th century to late 12th/early 13th century 10

209a

209b

209c

319

52r

321

10r

335

16r

3h, 3s

448/486

5h, 1s

457

103r

21r

29h, 2s

468

478

484/4079

496

12h

4011

4012

4016

4017

4018

515

13h

621

626

702

703

707

8h, 13r

803

19h

844

5h, 2s

869

874

55h, 2s

888

889

34r

8018

8020

cf 1

47; Hammersley 1973). In areas where wood was scarce, for example on some of the Scottish islands, carbonised peat provided an effective substitute (Fenton 1978). Iron smelting and smithing were traditionally undertaken in areas of abundant fuel supply, thereby allowing wood to be converted to charcoal on site, whereas ore and clay were frequently imported. In the early medieval period,

the conversion of wood to charcoal was probably undertaken in pits, which enabled the production of small quantities of good quality charcoal (Rackham 1976, 81). A 16th-century treatise on metallurgy endorses the use of pit charcoal, especially for smithing, as the resulting fuel was harder and more efficient than that from a clamp (Smith and Gnudi 1990, 179). The production of char-

274 high island: excavation of an early medieval monastery

coal requires approximately six to seven units of wood to produce one unit of charcoal, and extensive charcoal-burning activities can rapidly deplete large areas of woodland (Horne 1972; Armstrong 1978, 64). Archaeological evidence from numerous iron-working sites across Britain, dating from the Iron Age until the post-medieval period, indicate that charcoal made from fairly substantial oak heartwood was the preferred fuel for smelting and smithing (Cowgill 2003; Gale 2003). In 1662, John Evelyn describes methods of charcoal-making for domestic use, ironworking and gunpowder. For iron-working he notes that good-sized oak was preferable to narrow roundwood, the latter being usually supplied to domestic hearths (Armstrong 1978, 41). Although the evidence from High Island attested to on-site metal-working, this was unlikely to have been practiced on a commercial scale, and it is possible that both the charcoal, which is much lighter to transport than wood, was imported. The largest deposit of charcoal associated with the primary construction phase of the monastic enclosure wall related to a layer of silty clay abutting the north face (Table 9.17; F4037). The taxa consisted predominantly of hazel roundwood and oak, but also included holly, hawthorn (Sorbus group) and willow or poplar and yew. The deposit included four-year-old hazel stems (approximately 10mm in diameter), including some of rapid growth consistent with that of coppice (Table 9.20; F4037). Hazel charcoal from this context produced a radiocarbon date range of cal. AD 778–1000 (UB-4987, 1138±48 BP). Charcoal was sparse in the remaining samples taken from the layers at the north of the monastic enclosure. Hazel, oak, heather, ash and hawthorn (Sorbus group) were included (see Table 9.17 for distribution). Within the chamber in the western side of the monastic enclosure wall, a small lens of charcoal was obtained from the schist floor in which willow or poplar and heather were identified (Table 9.17; F705). Although almost certainly the remains of fuel debris, the source of the charcoal or its associated hearth is unknown.

Phase 2: mid-11th to late 12th/early 13th century

Phase 2 related to the period when the monastery was at its most active and new buildings, such as the church, Cell B and Cell A were constructed. Thirty-two samples of charcoal remains were recovered from all eight excavation areas albeit, in some cases, in quite small quantities (Fig. 5.1). Hand-collected charcoal from within the excavated central section of the east wall of the church contained willow or poplar roundwood (Tables 9.18, 9.24; F10). Inside the church, the altar produced small amounts of charcoal identified as birch, hazel and oak (Table 9.18;

F209a–F209c). The uppermost level (F209a) within the altar had been subject to recent disturbance. Two paved hearths were revealed within Cell B. Significant quantities of stratified burnt debris were revealed overlying the largest and centrally located paved hearth floor. The fuel remains consisted almost entirely of narrow ericaceous stems, but also included oak, ash and peat (Tables 9.18, 9.24; F319). A further deposit of burnt debris lay to the eastern side of the paved hearth floor, which included hazel, ericaceous stems, oak and juniper (Table 9.18; F321). The small quantity of charcoal from the fill (F335) of the stone-lined pit adjacent to the hearth floor was comparable to the domestic debris recorded from the main mound of burnt debris (F319) inside the cell and also from other deposits found elsewhere on the site. The domestic debris in the pit was likely to have accumulated here as a result of intermittent clearing of the hearth floors. The taxa identified included birch, hazel, heather, oak, willow or poplar (Table 9.18; F335). The fuel residues also contained peat (Table 9.24; F335). The bulk of the small samples associated with the fill deposits from within the northern side of the extended monastic enclosure wall (F4025) included oak and/ or hazel (Table 9.18; F448/F486, F4011, F4012, F4016– F4018). The exception in regard to the taxa identified was that from the burnt debris beneath these deposits, which represented a hearth. The range identified here was broader than in the overlying fills and was more comparable to the taxa identified in other possible hearth residues at the inner face of the wall (see below). The taxa identified were birch, hazel, heather, ash, hawthorn (Sorbus group) and oak (Table 9.18; F457). This deposit of burnt debris included a high ratio of fast-growth hazel stems and also peat (Tables 9.21, 9.24; F457). The origin of this deposit is uncertain, but was possibly domestic related. The charcoal from the burnt debris produced a radiocarbon date range at 2 sigma cal. AD 1027–1238 (UB-4521, 888±41 BP). Two deposits, which were uncovered at the inner face of the extended monastic enclosure wall and in close proximity to Cell A, also produced a broad range of taxa. The six species identified from the layer of ash and clay were birch, hazel, heather, ash, oak and juniper (Table 9.18; F478). Six species were also identified from the other deposit, the burnt debris. These were alder, hazel, heather, ash, oak and willow or poplar (Table 9.18; F484/F4079). An unidentified conifer was also present (Table 9.24; F484/F4079). Significantly, the recovery of heat-shattered stones and hones in the ash and clay could suggest some type of localised industry that required a heat source. The charcoal from the burnt debris (F4079) produced a radiocarbon date range at 2 sigma cal. AD of 1042–1217 (UB-6454 (AMS), 888±30 BP). Charcoal from the burnt debris beneath the paved floor within Cell A

the environmental reports â&#x20AC;&#x192; 275

Table 9.19â&#x20AC;&#x201A; Charcoal from Phase 3 and Phase 4. (Key: h = heartwood; r = roundwood (diameter <20mm); (r) = includes roundwood (diameter <20mm); s = sapwood (diameter undetermined, Quercus only). The number of fragments identified is indicated.) Feature Alnus Betula Corylus Ericaceae Fraxinus Ilex Pomoideae Prunus Quercus Salix/ Juniperus Picea/ Taxus no. alder birch hazel heather ash holly hawthorn spinosa oak Populus juniper Larix yew spruce/ group blackthorn willow/ larch poplar Phase 3: late 12th/early 13th century to mid-15th century 214

318

445

12 (r)

464

472

3h, 4s

14r

1h, 1s

479

483

488

489

494

4002

4005

509

59r

613

41r

614

616

12r

617

21r

619

623

625

841

3h, 1s

854

880

55r

13r

890

29r

8010

15h

8014

1h, 1s

Phase 4: mid-15th century to late 20th century -

2h, 2s

cf. 1

447

451

452

461

708

800

820

829

276 high island: excavation of an early medieval monastery

included hazel, heather and oak, and peat was also present (Tables 9.18, 9.24; F515). Charcoal extracted from this deposit produced a radiocarbon date range at 2 sigma cal. AD 1176–1276 (UB-6453, 802±34 BP). Hazel, oak and hawthorn (Sorbus group) were identified from a linear stone feature in the trench excavated at the south-west of the monastic enclosure (Table 9.18; F621). A sample of hazel and birch was recovered from the silty clay above the paved surface in this trench (Table 9.18; F628). Within the chamber in the western side of the monastic enclosure wall, oak was obtained from a spread of burnt debris on the floor (Table 9.18; F707). Small quantities of hazel were recovered from other deposits on the chamber floor (Table 9.18; F702, F703). There was no evidence to suggest that the deposits originated from other than domestic use. Small quantities of hazel and oak were identified in the silty clay fill of Grave 11 and in the overlying ash spread outside the north-east entrance to the church enclosure (Table 9.18; F8018, F8020). The overlying extensive layer of grey clay that covered the area to the north of Cell B in the monastic enclosure included a high proportion of narrow twigs and stems of juniper, and also oak and heather (Tables 9.18, 9.21; F889).

Phase 3: late 12th/early 13th to mid-15th century

This phase marked the decline and probable abandonment of the monastery by the religious community, though there is evidence of some activity continuing at the site. The nature of this activity is suggested to include at least one period of reoccupation of Cell A, but this appeared to have been transient. The island continued as a station for pilgrims, and would also have offered temporary shelter for farmers, fishermen and other visitors. The remains of burnt debris, or hearth residues, were recorded across the site and they indicated a similar use of firewood and peat to that of the earlier phases. Twenty-six samples of charcoal were examined that dated to Phase 3, the later medieval period up to the mid-15th century. They were recovered from the church and the monastic enclosure, including the interiors of Cell A and Cell B (Fig. 5.1; Areas 2–6, 8). The charcoal from the lowermost fill in pseudo Grave 10 in the north-western corner of the church included birch, hazel, oak and willow or poplar (Table 9.19; F214). Birch was identified from the ash that was lying on a postabandonment temporary surface inside Cell B (Table 9.19; F318). Samples collected from the layers of fill that had

Table 9.20 Charcoal: roundwood diameters and ages, Phase 1.

Feature Taxa no. Phase 1: 8th century to mid-11th century 75 Ericaceae, heather 93 Corylus avellana, hazel Salix sp./ Populus sp. 93c Ericaceae, heather Juniperus sp., juniper 93d Juniperus sp., juniper 93e Corylus avellana, hazel 93g Ericaceae, heather 93h Corylus avellana, hazel 93i Corylus avellana, hazel

221 250 4037 705

Ericaceae, heather Corylus avellana, hazel Juniperus sp., juniper Corylus avellana, hazel Corylus avellana, hazel Ericaceae, heather

Diameter in mm

Growth rings

2 20 12 1–3 1–4 2–5 25 1–4 15 15

13 6 4 4

<3 10 3 3 10 3

4 5 4 -

Comments

Fast growth; probably coppice Fast growth; probably coppice Fast growth; probably coppice, felled in late summer Fast growth; coppice stem Slow growth Large quantity of stems -

the environmental reports 277

Table 9.21 Charcoal: roundwood diameters and ages, Phase 2.

Feature Taxa no. Phase 1: 8th century to mid-11th century 319 Ericaceae, heather 321 Corylus avellana, hazel Juniperus sp., juniper 335 Salix sp./Populus sp., willow/ poplar 457 Corylus avellana, hazel 478 Juniperus sp., juniper 4016 Corylus avellana, hazel 869 Juniperus sp., juniper 874 Corylus avellana, hazel 889 Juniperus sp., juniper

Diameter in mm

Growth rings

Comments

<4 10 4 7

12 -

Slow growth -

12 3 12–17 2 3 2–3

4 5 -

Fast growth; large quantity of stems Fast growth -

Table 9.22 Charcoal: roundwood diameters and ages, Phase 3 and Phase 4.

Feature Taxa Diameter no. in mm Phase 3: late 12th/ early 13th century to 15th century 445 Corylus avellana, hazel 20 464 Juniperus sp., juniper 3 472 Corylus avellana, hazel 10 Ericaceae, heather 489 Ericaceae, heather 2–3 509 Corylus avellana, hazel Up to 10 Ericaceae, heather 2 Juniperus sp., juniper 2 616 Corylus avellana, hazel Up to 10 617 Corylus avellana, hazel 15 Fraxinus excelsior, ash 20 619 Corylus avellana, hazel Up to 10 841 Corylus avellana, hazel 10 880 Corylus avellana, hazel 10 Fraxinus excelsior, ash 10 Salix sp./Populus sp., willow/ 10 poplar Phase 4: mid-15th century to late 20th century 708 Ericaceae, heather 2 829 Corylus avellana, hazel 8

Growth rings 6 c. 6 8 2 8 5 1–4 10 -

Comments

Slow growth Slow growth Fast growth; probably coppice -

Fast growth; coppice stem

278 high island: excavation of an early medieval monastery

accumulated in the disused water-collection and drainage channel (F8025) in the northern part of the monastic enclosure included hazel, heather, ash, oak, hawthorn (Sorbus group) and juniper (Table 9.19; F445, F464, F483). Heather and oak were identified from a layer of sod that had accumulated within the northern end of the channel (Table 9.19; F489). At the section of the disused channel in the east of the enclosure, charcoal-rich deposits were preserved in the uppermost layers of silt. These consisted mostly of narrow hazel stems, but birch, heather, ash, holly, hawthorn (Sorbus group), oak and willow or poplar were also present (Tables 9.19, 9.22, 9.25; F880, F890). Small amounts of oak charcoal were recorded in the layers of silt and accumulated debris that sealed the rock-cut well and the path leading to it in the east of the enclosure (Table 9.19; F8010, F8014). A clay layer that covered the hearth outside Cell B included birch, hazel, oak and juniper (Table 9.19; F841). The charcoal from the layers relating to the period after the monastery was abandoned in the south-west of the monastic enclosure indicated the predominant use of hazel stems, some of which were fast grown, but birch, ash, holly, hawthorn (Sorbus group), blackthorn, oak, willow or poplar and juniper were also represented (Tables 9.19, 9.22; F613, F614, F616, F617, F619, F623, F625). Hearth debris found in situ on the stone floor in Cell A consisted predominantly of slow-growth hazel stems, but also included heather, juniper and peat (Tables 9.19, 9.22, 9.25; F509). Charred cereal grains were abundant in the hearth debris and they provided a radiocarbon date range of cal. AD 1287–1424 (UB-4988, 580±40 BP). Oak, hazel, ash and peat were present in a deposit of silty clay and stones on which the sub-rectangular structure (F427) to the north of the monastic wall had been built (Tables 9.19, 9.25; F479). A spread of ash, which was cut by the construction trench for this building, contained similar species, with the addition of heather (Tables 9.19, 9.25; F472). Ash and oak were also recorded in the bank of stone and soil defining the eastern side of the access passage (F4028) through the monastic wall to the sub-rectangular building (Table 9.19; F4002, F4005).

Phase 4: mid-15th to late 20th century

During this phase, High Island would have continued to attract pilgrims, those engaged in farming the land on the island and also fishermen seeking temporary shelter. The only historical record for occupancy relates to the copper miners in the 1820s. Eight samples of charcoal were examined, all of which came from within the monastic enclosure and its boundary wall (Fig. 5.1; Areas 4, 7, 8). The charcoal almost certainly originated from domestic fires. Fuel supplies of firewood and peat reflected similar usage to the previous periods. Samples from the deposits of clay and collapsed rub-

ble and from a hearth on the surface in the abandoned interior of the sub-rectangular building (F427) included birch, hazel, oak and peat (Tables 9.19, 9.25; F452, F461). A deposit of clay and stone that had accumulated above the paved surface (F422) within the monastic enclosure included birch, hazel, heather and ash (Table 9.19; F451). The charcoal from the ash, peat and grassy sod layers in the eastern part of the enclosure included hazel and oak (Table 9.19; F829, F820, F800). A localised deposit from outside the entrance to the mural chamber in the south-west of the monastic enclosure included heather and possibly willow or poplar (Table 9.19; F708). A sample from a peaty soil that had accumulated to the south and south-east of Cell A contained ash, hawthorn (Sorbus group), oak and, unusually, spruce or larch (Tables 9.19, 9.25; F447). Spruce and larch are anatomically similar and cannot be distinguished using wood features; neither is native to Ireland or Britain. The single piece of charcoal appeared to originate from a fairly wide branch or plank. The most likely source would be driftwood, though the rugged cliffs defining the coastline of High Island present few areas where such material can be collected, which may suggest that it was brought over from the mainland. Similar finds have occurred off the west coast of Britain, for example at the prehistoric settlements of Halangy Down and Nornour, on St Mary’s, Isles of Scilly, and also at Knockans, Rathlin Island, off the north coast of Ireland (Keepax and Morgan 1978; Gale 1996 and 1999).

Fuel resources

The bulk of the charcoal is attributed to domestic hearth debris and it is clear that, though this was often used in combination with peat, firewood was readily available. The frequency of charcoal in the fuel deposits throughout all of the phases is particularly striking given the (almost total) absence of trees and shrubs on the island today (Section 1.5). The diverse range of species identified is also of great interest. Oak and hazel were dominant and, to a lesser extent, heather (Tables 9.17–9.19). Additional species included birch, alder, ash, holly, the hawthorn group (which may include both hawthorn and rowan), blackthorn, willow or poplar, juniper and yew. Apart from noting that yew was only recorded in Phase 1 and mainly in the vicinity of the pre-church features and that the frequency of ash appeared to be slightly increased in Phases 3 and 4, there was no evidence of temporal or spatial disparity. The only secure evidence of industrial activity recorded was the metal-working in the smithing hearth pit located in the east of the monastic enclosure (Phase 1, F352). The charcoal from all other contexts was interpreted as domestic fuel debris from heating and/or cooking. Occasionally, the domestic hearths still retained the remains of fuel, for

the environmental reports 279

Table 9.23 Features and associated sample numbers that contained charcoal, Phase 1.

Feature no.

Samples examined

Phase 1: 8th to mid-11th century 42 ’96G, ’96E 56 68 75 93 93a 93b 93c 93c 93d 93e 93g 93h 93i 93j 93m 96 97 101 105 216 221 232 246 250 262 265 267 346 347/351 493 4034 4036 4037 4038 4039 4051 4052 4055 4072 705

’96Y ’97E ’97AD

Hand-collected charcoal Hand-collected charcoal Hand-collected charcoal Hand-collected charcoal ’99F Hand-collected charcoal Hand-collected charcoal ’99I Hand-collected charcoal ’00K, ’00Q ’99T Hand-collected charcoal Hand-collected charcoal ’99AD Hand-collected charcoal ’01B ’96I ’97AB, ’98AF, ’97O, ’96Z ’97K, ’98AP ’97L ’97S Hand-collected charcoal Hand-collected charcoal ’98AQ Hand-collected charcoal ’98AR, ’00G ’99Z Hand-collected charcoal ’01T ’01U, ’01W Hand-collected charcoal Hand-collected charcoal Hand-collected charcoal ’02A, ’02FF Hand-collected charcoal Hand-collected charcoal ’98B

Comments Includes ?tubers Includes peat Includes peat Includes peat Includes peat Includes peat and herbaceous stems (monocotyledons) Includes peat Includes peat and herbaceous stems/root (dicotyledons) Includes peat Includes herbaceous stem/root (dicotyledon) Includes peat Includes peat Insufficient charcoal for id Some very degraded Includes herbaceous stems Insufficient charcoal for id Includes peat Charcoal absent but includes ?tuber Includes large chunks of oak, some slow grown Includes peat Includes herbaceous stems (dicotyledons) Narrow stems Includes ?tuber -

280 high island: excavation of an early medieval monastery Table 9.24 Features and associated sample numbers that contained charcoal, Phase 2.

Feature no.

Samples examined

Phase 2: mid-11th to late 12th/early 13th century 10 Hand-collected charcoal 21a Hand-collected charcoal 209 a ’98D + Hand-collected charcoal 209 b Hand-collected charcoal 209 c ’99 N 319 ’97AJ, ’98G, ’98J 321 335 448/486 457 468 478 484/4079 496 4011 4012 4016 4017 4018 515 621 626 702 703 707 803 844 869 874 874 888 889 8018 8020 8036

’01S, ’97AI ’98AH

Hand-collected charcoal ’99B, ’99S Hand-collected charcoal ’99O ’00AJ Hand-collected charcoal Hand-collected charcoal Hand-collected charcoal Hand-collected charcoal ’00AD Hand-collected charcoal ’98C Hand-collected charcoal Hand-collected charcoal ’98AA ’98AV, ’98AD, ’98U ’98AJ ’01M

Hand-collected charcoal ’01A Hand-collected charcoal ’01E Hand-collected charcoal ’01V Hand-collected charcoal ’01Z Hand-collected charcoal

Comments Charcoal absent Very narrow woody stems. Includes herbaceous stem (dicotyledon) and peat. Some charcoal very degraded Includes peat Includes peat Includes peat Includes unidentified conifer Includes peat Includes peat and herbaceous stems (dicotyledon) Includes peat, herbaceous stems (dicotyledon) and tarry material Includes peat and herbaceous stems (dicotyledon) Narrow roundwood Mostly very narrow roundwood Charcoal absent, probably peat

the environmental reports 281

Table 9.25 Features and associated sample numbers that contained charcoal, Phase 3 and Phase 4.

Feature no.

Samples examined

Phase 3: late 12th/early 13th to mid-15th century ’96H 214 214b Hand-collected charcoal 318 ’98AB 445 ’99G + hand-collected charcoal 464 ’99V + hand-collected charcoal 472 ’99S 479 Hand-collected charcoal 483 Hand-collected charcoal 488 ’99Y 489 ’99AB 494 ’99AA 4002 Hand-collected charcoal 4005 Hand-collected charcoal 509 ’97AC 613 Hand-collected charcoal 614 Hand-collected charcoal 616 Hand-collected charcoal 617 Hand-collected charcoal 619 Hand-collected charcoal 623 Hand-collected charcoal 625 Hand-collected charcoal 841 Hand-collected charcoal 854 ’00AC 880 ’01L + hand-collected charcoal 890 ’01Q + hand-collected charcoal 8010 Hand-collected charcoal 8014 ’01Y Phase 4: mid-15th to late 20th century 434 Hand-collected charcoal 443 Hand-collected charcoal 447 Hand-collected charcoal 451 Hand-collected charcoal 452 Hand-collected charcoal 461 Hand-collected charcoal 708 ’98AU 800 Hand-collected charcoal 820 Hand-collected charcoal 829 Hand-collected charcoal

Comments Includes wood from unidentified conifer Includes peat Includes peat and herbaceous stems (monocotyledon) Includes peat and Pteridophyte (?bracken/fern) Includes peat Includes peat Includes peat and herbaceous stems (dicotyledon) Includes peat Mostly narrow roundwood ?Aquatic monocotyledonous root/rhizome Includes peat but no charcoal Includes peat Includes peat Includes peat Includes herbaceous stems (dicotyledon) -

282 high island: excavation of an early medieval monastery

example, on the paved hearth in Cell B (Phase 2, F319), in the chamber at the south-west of the monastic enclosure wall (Phase 2, F707) and within Cell A (Phase 3, F509). Domestic firewood consisted mostly of narrow roundwoods from a range of taxa. In contrast, the industrial fuel consisted of chunks of mature oak from fairly wide, slowgrowth roundwood or trunkwood. In view of the consistency in supply and the frequency of use in all phases, it is suggested that firewood was supplied from sources within the island, probably from managed woodland growing in sheltered valleys. This suggestion is supported by the regular use of roundwood and the fairly high incidence of fast-growth hazel stems. The structure of the hazel roundwood was often consistent with that of coppice growth, where there were wide incremental rings in the initial stages of growth (Tables 9.20–9.22; Morgan 1982). These result from the rapid regeneration of stems (rods) from well-developed root stock growing in stressfree conditions (see below, ‘Environmental Evidence’). Managed coppice is traditionally cut on a cyclical basis, the rotational cycle being determined by the ultimate use of the wood, for example, approximately seven years for hurdles but less for basket-making. However, similar growth patterns could develop in stems of isolated bushes or trees cropped on a more random or irregular (nonmanaged) basis. The presence of fast-growth wood also inferred that woodland grew in sheltered areas or valleys and was protected from grazing stock. When the island came under the auspices of the monastery, it is probable that land use was organised to maximise provisions for the community, particularly with reference to food and fuel. The extant woodland would almost certainly have been controlled and managed, ensuring the survival of this renewable crop. By the 8th century, laws of woodmanship were enforced to stabilise wood shortages across Ireland, with steep penalties incurred for the illegal felling of trees (Mitchell 1986, 165). The similarity of the taxa in the charcoal deposits throughout Phases 1–4, from the 8th century to the 20th century, suggests a common source for the fuel and, by implication, that firewood remained available on the island throughout this time. Once the monastery went into decline, unless management was maintained, local woodlands would have become derelict. It is an interesting thought that the island may then have become a source of fuel for coastal communities on the mainland. So far, it has been assumed that the island was selfsufficient in firewood throughout the entire period, but alternative or supplementary sources of fuel should also be considered. The most obvious source would be the importation of fuel by boat. Until relatively recently, currachs in western Ireland were constructed as sea-going crafts from flat lengths of timber (about 6m long) and

they were capable of carrying large cargoes (Mitchell 1986, 84). Bundles of faggots and brushwood, hurdles and other wooden components could have been transported to the island by currach. Discarded wooden items would almost certainly have been recycled as firewood. The topography of the island would have made the collection of driftwood difficult, but perhaps not impossible, and the presence of spruce or larch in a late context suggested the use of driftwood (Phase 4, F447). A more local source of fuel may have been wood peat: the remains of woodland growing prior to the development of peat bogs, which were subsequently exposed following the removal of overlying layers of peat (see below, ‘Environmental Evidence’). In western Ireland, pine (Pinus) formed a major element of these early woodlands and its presence on High Island was verified by the pollen record in prehistoric levels (Molloy et al. 2000), but since there was no evidence of pine in the charcoal deposits on High Island, the use of wood peat seems unlikely. Peat was an important constituent of the fuel and it cropped up regularly in the fuel debris (Tables 9.23–9.25). In most samples, peat occurred as hard concretions of compacted, black material. This was composed of numerous fragments of plant, including stem, leaf and epidermis. Typically, this consisted of narrow, woody stems (up to 2mm in diameter), with central pithy cores reduced to hollow channels. The dimensions and character of these stems were similar to those identified in the charcoal as heather. There were also dense fibre caps (from vascular bundles) from monocotyledonous plants, for example, grasses and reeds, and epidermal fragments and leaf fragments from both dicotyledons and monocotyledons. Diagnostic features were sparse and it was not possible to name any species. This range of plant material is consistent with that associated with peat formation (Mitchell 1986, 93). The lumps of peat were often difficult to fracture for examination and the surfaces frequently appeared vitrified (glassy), a condition usually associated with burning at high temperatures. Under high magnification, it was evident that the plant tissues had undergone loss of structure, with the partial fusion of some cell walls.

Environmental evidence

Based on pollen records, there is now a considerable corpus of knowledge pertaining to the development of prehistoric woodlands in Ireland (Mitchell 1986, 33–62; 1995; O’Kelly 1989, 7–11). The regional distribution of arboreal species was influenced not only by local soils but by a major east or west (dry or wet) divide. The first development of peat began about 9,000 years ago and continued as a cyclical process, which was greatly affected by the climate, local water table and drainage. The formation of blanket bog effectively drowned existing woodland and overlay the

the environmental reports 283

fallen remains. With ameliorating climates, some areas became drier, thereby enabling the growth of heather and finally the development of woodland. It is estimated that by about 5,000 BC primary woodland, predominantly pine, oak, hazel and elm (Ulmus), covered at least 80% of the land surface. By the Neolithic period, pine was restricted mostly to western Ireland, where it was gradually replaced by oak and hazel until it was totally eradicated (probably about AD 200). The prevailing climatic and edaphic conditions particularly favoured the spread of hazel. The early and extensive land clearance to make way for agriculture reduced the broad swathes of primary woodland to areas of scrub. In the present day, woody species on High Island are confined to creeping willow (Salix repens), a low-growing species, and ling (Calluna vulgaris) and crowberry (Empetrum nigrum), both members of the heather family Ericaceae; the two latter species are sparse and confined to the north-eastern end of the island (Molloy et al. 2000). Palynological studies on High Island by Molloy, Fuller and Conaghan (2000) show that until the late Iron Age, woodland species consisted predominantly of hazel and pine but also included elm, alder and oak. Following occupation of the island there was a marked decrease in tree pollen, which can be attributed largely to land clearance to make way for agriculture. Hazel appears to have survived, however, though less frequently than before. The age at which trees flower (and produce pollen) varies with the species (Rackham 1995), and if coppicing was practiced from this period on a regular, short-term rotation, the amount of pollen in soil deposits would have been significantly reduced. The disappearance of pine can probably be attributed to climatic changes and seems to have occurred well before occupation of the island. It is worth noting that the resinous wood of pine burns fiercely and would have provided an excellent source of firewood but once felled, pine, unlike broadleaf trees, does not regenerate. It is interesting, therefore, that the earliest archaeological evidence of occupation on High Island dates to the Iron Age and associated charcoal, though sparse, provided evidence of hazel, juniper and hawthorn (Sorbus group). Following the establishment of the monastery in the 8th century, the remains of hearth debris verified the use of peat and, to a large extent, firewood obtained from a broad range of species. When the fuel from High Island is considered in the context of the archaeological evidence from other Irish island sites, it seems more likely that firewood was mainly obtained from local (managed) woodland. The taxa identified from the hearth deposits on High Island were predominantly oak and hazel, but also alder, birch, heather, ash, holly, hawthorn (Sorbus group), blackthorn, willow or poplar, juniper, yew. These species correlate closely with those recorded from charcoal from a

Bronze Age settlement at Dún Aonghasa, Inis Mór, Aran Islands, though here oak formed the dominant taxon and there was less evidence of hazel (Gale 2012). The site on Inis Mór presented evidence of a wooded landscape that was probably typical of many islands on the western seaboard, where sheltered valleys provided a stable environment. Only tough species, such as juniper, birch and hawthorn, would have tolerated the harsh salt-laden winds in exposed situations. Today, Inis Mór is barren of trees. If these islands supported woodland until relatively recently, as is suggested by the archaeological evidence, the question remains as to what caused the final transition to their present barren status.

Conclusion

High Island is one of a group of small islands situated on the western seaboard of Ireland that, in the present day, are almost, if not totally, barren of trees and shrubs. Based on pollen records from High Island, however, it seems probable that sometime before the Iron Age these islands were tree-covered to some extent, predominantly with pine and hazel but also oak, alder and elm. Following climate change, pine populations declined and disappeared. Hazel continued to flourish, however, and was still present when the early medieval settlements were established. Alder and oak also show up in the pollen record at this time. In later levels, tree/shrub pollen gradually diminished. Archaeological evidence from the extensive excavations of the monastery on High Island suggests that a rather different vegetal history may be feasible. Although the size of the monastic population remains unknown, it is thought that during the most active period of occupation there may have been as many as 50–70 inhabitants on High Island (Rynne 2000, 212). In view of the tricky access to the island’s rocky coastline and the not inconsiderable distance to the mainland, it is probable that the community strove to be self-sufficient in both food and fuel, and also probably to maintain a degree of seclusion. In addition to peat (easily obtained from the 1m deep layer covering the island), fuel residues from domestic hearths demonstrate the use of a surprisingly diverse range of woodland species, particularly oak, hazel and heather, but also alder, birch, ash, holly, hawthorn (Sorbus group), blackthorn, willow/poplar, juniper and yew, which would infer that a much wider range of trees and shrubs grew in the local environment than previously thought (or indicated by palynology). However, alternative sources could be relevant for sparsely represented taxa. Yew, for example, may relate to discarded artefactual material. Some firewood may have been gathered as driftwood, as indicated by a late (Phase 4) deposit of non-native spruce or larch. Indeed, by this period, the island may have been more or less devoid of woodland resources.

284 high island: excavation of an early medieval monastery Table 9.26 Results of loss on ignition analysis.

Sample number

Predominant constituents

Percentage weight loss on ignition

F8000/’02X/3

Component 1

2.22%

F221/’97O/1

Component 5

13.16%

F221/’97O/2

Component 5

12.36%

F221/’97O/3

Component 5

12.45%

F93a/’99C

Components 2 and 3

39.12%

F93c/’99F

Components 2 and 3

51.50%

F93i/’99Q

Components 2 and 3

49.53%

A smithing hearth, dating to Phase 1, was used for processing iron and other metals. The amount of work undertaken here is unknown, but seems unlikely to have been large-scale. The preferential selection of oak charcoal to fuel this operation correlates with well-established practices elsewhere in Britain and Europe and suggests that oak was either grown especially for this purpose on a longer rotational cycle (to produce heartwood) or taken from cordwood from fairly substantial trees. For only occasional use of the hearth, small quantities of suitable charcoal may have been imported from the mainland. The excavated archaeological evidence suggests that the island and monastery were not permanently occupied from the late 12th/early 13th century. A reason for this is, perhaps, that it became impossible to sustain the island economy. Local woodland resources may have been overused and rundown, or affected by bad weather and storms causing waterlogging or soil erosion. With the cessation of regular management and protection from grazing, extant woodland evidently deteriorated and, at some point, disappeared altogether. In the absence of more conclusive evidence for total self-sufficiency on the island, one cannot rule out the possibility that regular supplies of wood and charcoal were ferried from the mainland by sea-going currachs to supplement home-grown supplies.

9.4 Soil and sediment analysis Stephen Lancaster

Introduction

The archaeological deposits at High Island are complex and variable in nature. This is reflected in the variety of sampling locations, which included: from deposits within and under the church building below Cell B; from deposits infilling walls, abutting cells; and from the fill of the monastic enclosure wall. Samples were taken in order to understand the composition and formation processes of some of the deposits and how these reflected the history

of the site and the activities that had taken place there. Specific issues included determining whether different contexts had the same composition, sources of materials and whether deposits had undergone significant human modification. The principle technique used was thin-section micromorphology, the examination of slides of soil or sediment at the microscopic scale. This allows the examination of the microstructure and composition of deposits to be undertaken, to provide observations that cannot be made in the field. Loss on ignition tests supplemented the micromorphological examination. The original work was done in two phases, for which separate reports exist (Lancaster 2005; 2006). Mortar analysis (Appendix A) and Clayey Soil analysis (Appendix B) were also undertaken.

Methodology Sampling Thin-section sampling Most samples were taken as small, non-orientated blocks or clods by the excavator. Two larger block samples were also taken, and were sub-sampled in the laboratory to give orientated samples. A total of 21 samples were submitted for analysis (Table 9.30). Sample numbers include the feature number, followed by the year collected, followed by a letter of the alphabet denoting the sequence it was collected in, i.e. sample no. F216/’96I was obtained from feature F216 in 1996 and was the ninth sample collected. The samples are: Non-orientated blocks F271/’00T, F8060/’02U, F336/’98AS, F246/’97L, F4030a/’01J, F105/’01B, F232/’98AP, F93a/’99C, F93c/’99F, F93i/’99Q, F205/’00AK, F42/’96G, F262/’00AN, F216/’96I, F270/’00Q, F8033/’02N. F8049/’02P, F803s/’00AE, F850/’01I.

the environmental reports 285

Table 9.27 Main components identified in thin-section.

Component no.

Description

Basic interpretation

Samples

Poorly sorted, randomly arranged, mineral particles with a wide range of sizes. Predominant minerals: quartz grains, feldspars and micas, moderately to heavily weathered condition. Occasional minerals: fragments of granite and other igneous rocks. Rare minerals: fragments of limestone or metamorphic rocks.

Composition consonant with glacial till in this region of Ireland, which would form the basis of any local mineral soils. Mineral parent material/ sub-soil.

F8000/’02X/1-3, F271/’00T, F8060/’02U, F336/’98AS, F205/’00AK, F105/’01B, F42/’96G, F262/’00A, F216/’96I, F8033/’02N, F803s/’00AE

Pale, amorphous, anisotropic material, common remains of altered biogenic silica, sometimes identifiable as diatom frustules.

Composition (see ‘Description’ and ‘Loss on ignition results’) suggests the residue left from burning lacustrine mud. Mineral ash.

F232/’98AP, F246/’97L, F336/’98AS, F93a/’99C, F93c/’99F, F93i/’99Q, F221/’97O/1-3.

Black material, opaque in transmitted light, strongly reflective under incident light, mostly amorphous with occasional small areas of cellular structure.

Optical properties indicate charred material, scarcity of cellular structure suggests derivation from humified peat. Charred material.

F93a/’99C, F93c/’99F, F93i/’99Q F42/’96G F8033/’02N,

Coarse mineral material: unsorted sand and silt-sized particles of quartz, weathered feldspar and micas. Fine organo-mineral material, proportion of organic and mineral in fine material variable. Occasional channels and invertebrate excrement pellets. Occasional fragments of other component types.

Consistent occurrence of features attributable to soil invertebrates suggests this being the A or A/B horizon of a silty soil. Topsoil.

F93a/’99C, F93c/’99F, F93i/’99Q, F221/’97O/1-3, F232/’98AP, F246/’97L, F336/’98AS, F42/’96G, F8033/’02N, F850/’01I.

High birefringence carbonate particles.

Ash, principally derived from wood.

F221/’97O/1-3

Principally clay. Domains seem to vary widely in size, about 15–20% of the area of the slide is covered with material in the siltsized range, quartz grains, very heavily weathered feldspars, mica grains and occasional fragments of granite.

Clay.

F270/’00Q, F4030a/’01J, F8049/’02P

286 high island: excavation of an early medieval monastery Table 9.28 Thin section descriptions.

Sample no.

Principal components

Additional observations

F221/’97O/1-3

3, 4, 5

Components 3 and 5 arranged in fine laminae and lenses. Alternate with thicker bands of Component 4 with areas of Component 5. Also fused and distorted phytoliths, small fragments of bones. Internal arrangement of laminae, lenses and bands chaotic.

F8000/’02X/1-3

Densely packed. Occasional passages infilled with Component 1. Occasional fragments of Component 3. Dusty coatings in voids toward base of sequence.

F216/’96I

Very stony. Rare infillings of Component 5.

F232/’98AP

Frequent fragments of Components 2 and 3. Rare phosphatic infillings of channels.

F42/’96G

1, 3, 4

Component distribution random and chaotic. Bone fragments frequent. Minerals heavily weathered.

F105/’01B

Occasional phosphatic infillings. Comminuted charred material throughout sample.

F246/’97L

Frequent fragments of Component 2.

F262/’00AN

Channels and biologically related fabric patterns.

F93a/’99C, F93c/’99F, F93i/’99Q

2, 3, 4

Rare faecal pellets.

F270/’00Q

Pure kaolinite (Pavía and Bolton, this volume, Appendix B), well sorted, clay domains c. 2µm, parallel orientation.

F4030a/’01J

Clay domains vary widely in size. Silt-sized quartz, feldspar and micas, plus granite fragments: 15–20% of slide area.

F8049/’02P

Clay domains vary widely in size. Silt-sized quartz, feldspar and micas, plus granite fragments: 15–20% of slide area.

F271/’00T

F8060/’02U

Rare channel infills of organic matter and biologically related fabric patterns identified.

F336/’98AS

1, 2, 4

Comminuted charred material throughout sample. Some of Component 2 forms channel infills.

F205/’00AK

F803s/’00AE

Channels and biologically related fabric patterns.

F850/’01I

1, 4

Occasional fragments of Component 1. Abundant ferruginous and organoferruginous impregnative pedofeatures. Rare mineral ash aggregates. Some impregnative pedofeatures fractured, as are some mamillated excrements observed in the sample.

F8033/’02N

1, 3, 4

Arranged in fragmented laminae, different laminae composed of different components. Component 1 is most frequent; Component 4 generally occurs in combination with Component 3. The orientation of the groups of laminae varies across the slide.

the environmental reports 287

Table 9.29 Classification of thin sections into interpretive groups.

Interpretive group

Components

Thin sections

Organo-mineral soils. Semi-natural/natural till-derived mineral sub-soils and organo-mineral soils derived from them, including turf.

1, 4

F205/’00AK, F216/’96I, F105/’01B, F262/’00A, F803s/’00AE, F8000/’02X/1-3, F271/’00T and F8060/’02U, F850/’01I

Organically rich soils. Organo-mineral soils, richer in organic matter, charred material and bone fragments than preceding group.

4, 2

F232/’98A and F246/’97L

Stratified deposits In situ laminated deposits of ash, charred material and organo-mineral soil.

5, 3, 4,

F221/’97O/1-3

Chaotic infills/dumps Mixed, randomly arranged fragments of organo-mineral soil and charred material.

4, 1, 3

F42/’96G, F336/’98AS and F8033/’02N

Fuel residue dumps Deposits of bands and fragments of charred material and lacustrine mud.

3, 2, 4

F93a/’99C, F93c/’99F and F93i/’99Q

Clays

F270/’00Q, F4030a/’01J and F8049/’02P

Orientated sub-sampled blocks F221/’97O/1, F221/’97O/2, F221/’97O/3, F8000/’02X/1, F8000/’02X/2, F8000/’02X/3. Loss on ignition sampling Samples were taken from contexts (F93a/’99C, F93c/’99F and F93i/’99Q) rich in Component 2 (see below) and contexts (F221/’97O/1-3) known to be rich in ash on the basis of thin-section micromorphology, so that the organic content of the different context types could be compared. A sample was taken from the mineral sub-soil (F8000/’02X/1-3) to act as a control. Thin-section manufacture In order to make the thin sections, the samples were impregnated with resin, cured, cut, mounted on a large glass slide and ground to a thickness of 30µm. Resin impregnation and thin-section preparation was undertaken by the School of Biological and Environmental Sciences, University of Stirling, Scotland, and followed standard procedures (Murphy 1986). Thin-section analysis The thin sections were recorded using an Olympus BX50 petrological microscope and using a variety of lighting techniques. Recording followed the scheme recommended by Bullock et al. (1985), as supplemented and modified by Courty et al. (1989).

Loss on ignition tests Loss on ignition samples were air-dried and passed through a 2mm sieve. The <2mm fraction was then combusted at 450°C, this temperature being selected to avoid decomposition of carbonates, which would have distorted the results (Rowell 1994). The percentage mass loss on ignition is taken as a measurement of the organic content of the sample, with higher loss indicating higher organic content. Loss on ignition tests were performed by the School of Biological and Environmental Sciences, University of Stirling.

Results Loss on ignition analysis The mineral sub-soil control sample rendered the expected low organic content (White 1997). The carbonate ash-rich samples can be clearly distinguished from those from the Component 2-rich samples on the basis of loss on ignition measurements. The carbonate ash-rich samples fall in the 12–13% range, the Component 2 samples fall in the 40–50% range.

Thin-section analysis Six basic components and groups of components have been identified that comprise the vast majority of the samples. These are tabulated below, together with an interpretation and list of samples which contain a significant

288 high island: excavation of an early medieval monastery

proportion of that component type (Table 9.27). Individual slide descriptions are given in Table 9.28. The six components are: Component 1, sub-soil; Component 2, mineral ash; Component 3, charred material; Component 4, topsoil; Component 5, carbonate ash; Component 6, clay. The interpretation of Component 2 was dependent on the loss on ignition tests (see below).

Discussion Loss on ignition and the composition of Component 2 If Component 2 was essentially organic in nature, then the loss on ignition results of the Component 2-rich samples should have been 90–95%. If Component 2 was essentially mineral in nature, then the combination with the charred material should give loss on ignition results in the 40–60% range, varying according to the proportion of charred material to Component 2 (Lancaster 2006). The samples taken from the Component 2-rich deposits, F93a/’99C, F93c/’99F, F93i/’99Q, have produced results between 39% and 51% organic content. The source of Component 2 may be suggested on the basis of its composition. It is composed of pale, amorphous, anisotropic material. This very fine mineral ash contains a relatively high frequency of biogenic silica, occasionally identifiable as diatom frustules, interpreted as residue from the burning of an organic sediment. A fine, organically rich sediment containing diatoms suggests a deposit that formed in low energy, aquatic conditions, e.g. lacustrine mud. Component 2 is interpreted as fine mineral ash, derived from organic mud. The use of such deposits as fuel has been described in the Northern Isles of Scotland, an area facing similar fuel constraints to the islands off western Ireland (Carter 1998).

Formation processes The thin sections have been classified into interpretive groups based on composition and formation processes. The classification of the thin sections, together with the components that form the predominant elements of the sections, is tabulated below (Table 9.29). The first four groups all contain significant proportions of the sub-soil (Component 1), which is derived from the local glacial till. The majority of the sampled material is therefore of local derivation.

Organo-mineral soils The largest interpretive group is the organo-mineral soils. In addition to the predominant glacial till-derived mineral content, there is generally evidence of biological reworking, with limited evidence for cultural inputs. This group demonstrates the sequence of soil development over the site, starting with the glacial till and

developing through the input of organic material, biological activity and weathering of igneous minerals. Other common soil processes have occurred, such as translocation of iron and fine material, but only on a minor scale. This group includes both in situ soils and those that have been redeposited. Where it is possible to ascertain through contextual evidence, it is usually the case that the ‘A’ horizon of these soils has been truncated or removed. The frequent observation of comminuted charred material in the redeposited soils and the upper part of the sub-soils (e.g. Sample F8000/’02X/1) indicates the reworking and movement of the soils on-site. The variability of the soils in this group is due to three factors: natural variation in the sub-soil, differential weathering and variations in cultural inputs. Of these, natural variation is probably least significant, with differential weathering of the mineral fraction accounting for most of the variation of texture in this category.

Organically rich soils The organically rich soils are derived from the local subsoils, but have significant cultural inputs: these soils have a high content of the mineral ash component, intimately mixed with charred material (as fragments and finely comminuted particles) and bone fragments. These deposits probably represent the effect of soil-forming processes active during human activity on the site, resulting in the substantial input of cultural material into deposits being homogenised by earthworm reworking.

Stratified deposits The fine lamination of the stratified deposits suggests that these were deposited intermittently in thin spreads over time. There is no evidence that these deposits were produced in situ in terms of evidence of heating of the mineral component: although this implies some degree of transport, it may not necessarily mean anything further than a hearth being raked out. Some of the periods between dumps of material being deposited have been sufficient to allow soil formation to occur, through biological activity. These samples are probably of broadly domestic origin. The wood ash content is notable and contrasts with the other fuel residue types from the site. The survival of the carbonate wood ash component in the stratified dump deposits may reflect the impact of the building of the church. Over the site in general the survival of the carbonate ash would be highly unlikely, given a naturally acidic sub-soil and the wet climate that would assist soil leaching. The construction of the church would keep the underlying deposits relatively dry, preventing the leaching of the wood ash, and perhaps also preventing physical disruption of the deposits, which would account for the surviving fine lamination in the deposits under

the environmental reports 289

Table 9.30 Contextual information of analysed samples (Scally 2013).

Feature no.

Sample no.

Phase

Contextual information

F8000

’02X/1-3

In situ sub-soil north side Cell B, Area 8.

F271

’00T

In situ sub-soil within church, Area 2.

F8060

’02U

In situ sub-soil south side Cell B, Area 8.

F336

’98AS

Disturbed sub-soil surface inside Cell B, Area 3.

F232

’98AP

Charred vegetation, possible ‘old ground surface’, Area 2.

F246

’97L

Burnt residues within F232, Area 2.

F105

’01B

Disturbed sub-soil NE corner church exterior, sub paving (F29), Area 1.

F93a

’99C

Redeposited fuel/hearth residues retained by wall (F91) north of church enclosure, Area 1.

F93c

’99F

As F93a.

F93i

’99G

As F93a.

F221

’97O/1-3

Redeposited fuel/hearth residues sealing ‘paved area’ (F245), sub-church, Area 2.

F205

’00AK

Redeposited sub-soil possibly used a rudimentary mortar, Area 2.

F4030a

’01J

Weathered granite within fill of primary monastic enclosure wall, Area 4.

F42

’96G

Dark grey moist clay in pre-church level of ‘pseudo’ Grave 2, = F50, Area 1.

F216

’96I

1/2

Fragmented schist clay, possibly contemporary with pre-church ‘slotted feature’ or sub floor level of church, Area 2.

F262

’00AN

1/2

Mixed deposit of schist clay, stones, hearth residues, located within excavated section of east wall of church, probably upcast deposit on which wall built, Area 2

F270

’00Q

1/2

Ball of kaolin clay found in deposit (F216/F262) at lowest level of east wall of church, Area 2.

F8033

’02N

Disturbed and redeposited fuel residue, Area 8.

F803s

’00AE

Redeposited sub-soil, packing around Cell B, Area 8.

F850

’01I

Possible sod cover over F803, north side Cell B, Area 8.

F8049

’02P

Weathered granite over paving within north-east passage (F814/F8040) through church enclosure wall, Area 8.

the church, which is not repeated in the samples from the other parts of the site. Unlike the wood ash, charred peat is nearly ubiquitous throughout the samples, either in relatively large fragments or in highly comminuted form, intimately mixed through the fine fraction of many deposits. The relative proportions of the two residues (ash and charred material) are not strictly comparable because of the issue of differential preservation, however it can be noted that peat was being burnt often enough to become a ‘background’ element in the local soils and sediments. This is evidence of

recycling and mixing of materials across the site, reflecting the persistence of human activity and its impact on the local deposits.

Chaotic infills/dumps The chaotic infills/dumps have been formed through the rapid, deliberate, filling or dumping of material. They are composed of a high proportion of fragments of organo-mineral topsoil but, unlike the redeposited soils, are combined with significant quantities of other components in a random and chaotic arrangement. These

290 high island: excavation of an early medieval monastery

deposits may be the result of similar processes to those that elsewhere have removed topsoil (see above). Fragments of laminated sediment, similar to the laminated sediment found in Samples F221/’97O/1–3, were found in Sample F8033/’02N. This demonstrates that this disturbance resulted in the disruption of previous cultural deposits, i.e. the fragmentation and deposition of the laminated sediment. Building associated with the later phases of activity on the site did not, however, destroy all older stratigraphy as shown by the survival of the wood ash-rich laminae in Samples F221/’97O/1–3, which underlay the church.

Fuel residue dumps The fuel residue dumps are predominantly composed of two forms of fuel residue: amorphous charred material, perhaps peat, and fine mineral ash. The simplest interpretation of the samples must be that they come from a dump of ashes and incompletely burned fuel. The presence of these fuel residues in relatively pure forms and as constituents of topsoil fragments suggests that the dumps were used for some time, allowing the beginning of soil formation, hence the incorporation of the fuel residues into the topsoil, which was disrupted by the addition of further fuel residues. Whether the dumps were simply a convenient point of disposal or were a form of storage for the material for further use is uncertain. Such material could have been used as a fertiliser/soil conditioner for the monastic garden.

Clays The clay samples stand as a separate group. The precursor minerals from which the clays developed indicate a source in weathered granite. Samples F4030a/’01J and F8049/’02P appear to be of primary origin, that is they are the result of in situ weathering of granite. Sample F270/’00Q is composed of very well sorted kaolin, the degree of sorting indicating a secondary origin, i.e. that the clay has been washed out from its original weathering point and deposited elsewhere. The difference between the clays in terms of origin indicates knowledge of (or access to) more than one source, but in any case this will be an off-island source. It seems highly unlikely that weathered lumps of clay could have survived transport by ice and none has been identified in the samples of sub-soil examined for this assessment. It is therefore concluded that the lumps of clay must have been imported deliberately.

Discussion Differential survival The pattern of accumulation seen in Samples F221/’97O/1-3, from the pre-church deposits, is one of phases of deposition alternating with phases of no deposi-

tion, during which biological processes rework the deposits to form soils. The time taken for this to occur is difficult to estimate; it could be anywhere between 40 and 200 years. There is no evidence of gross disruption of these deposits. This contrasts with many of the deposits across the rest of the site, where there is evidence of: rapid deposition, e.g. F42/’96G, F336/’98AS and F8033/’02N; disruption of soil profiles, e.g. F205/’00AK, F8000/’02X/1, F216/’96I and perhaps F271/’00T and F8060/’02U, F850/’01I, F105/’01B, F262/’00A, F803s/’00AE; or pre-existing cultural deposits, e.g. the fractured laminae in Sample F8033/’02N. These processes probably relate to the intensive nature of the occupation in the extant monastic area. This level of activity, especially the disruption of the soil profiles, suggests that any earlier occupation deposits may have been destroyed.

Construction impact and use of soils A particular factor in the movement and reworking of those samples categorised in the ‘Organo-mineral soils’ interpretive group is likely to be the removing of topsoil and ground-levelling at various stages of occupation: Samples F205/’00AK, F216/’96I, F803/’00AE and F8033/’02N are all redeposited sub-soils. Samples F105/’01B and F8000/’02X/1 are disturbed sub-soils – probably in situ, but showing signs of disturbance and intrusive material. Although such disturbance need not necessarily be the result of construction activity, it certainly adds to a picture of a site with intensive activity having an impact on the local soils. A possible example of the use of local soils in construction may be seen in the case of deposits F803 and F850. F803 is interpreted as a redeposited sub-soil. The position of the deposit with respect to the northern wall of Cell B suggests the use of the sub-soil to help weather-proof the beehive hut, particularly to help shed water away from the foundations and lower wall. Although the composition of F803 would be suitable for this, the high silt content would make the deposit prone to erosion. Deposit F850 had originally been interpreted in the field as a layer of peat. The deposit is actually a somewhat humic turf or topsoil. The deposition of turf, in the form of F850, might either have been to repair this or to protect F803, in addition to acting as a further weather-proofing layer in its own right.

Fuel economy Fuel residue forms an important element in many of the samples. Three types have been identified through soils analysis: the carbonate fuel ash, associated with burning wood; amorphous charred material, which is probably derived from peat; and the mineral ash derived from burning organic-rich mud. Carbonate ash is only present in Samples F221/’97O/1–3, under the church. Given the presence of wood charcoal from a variety of contexts across the

the environmental reports 291

site, this must reflect differential preservation of the carbonate ash across the site (see above and Gale, this volume, p.269). There appears to have been a change in the fuel regime over the period of occupation of the site. The mineral ash derived from burning organic-rich mud has only been noted in the earlier Phase 1 deposits, having been found in very small quantities in Samples F232/’98AP and F246/’97L and in larger quantities in Samples F221/’97O/1-3 and F93a/’99C, F93c/’99F and F93i/’99Q. There are no significant observations of this material in the later phases. This suggests a change in the pattern of fuel use, with the organic mud not being used as fuel after Phase 1. Wood and peat appear to be used throughout the history of the occupation of the site. Relative proportions of these fuel types used may have varied over this period, but it is not possible to accurately assess this from the available evidence.

Local and imported materials The majority of the material in the samples is probably of local derivation, including not just the local topsoil and sub-soil but probably also the organic-rich mud. Whether the source of such a sediment is on the island or elsewhere cannot be determined from the slides and loss on ignition data. Organic mud would constitute a poor quality and heavy fuel, unlikely to have been imported to the island. Assuming the currently mapped reservoir on the island was originally a natural feature, then it would seem a likely source. Some material could have been imported: any peat sources on High Island would have been exhausted or at least severely depleted early on in the occupation of the site, so peat for fuel may have been brought to the island.

The wood charcoal recovered from the site is thought to have come from trees on the island (Gale, this volume, p.269), and it can be assumed that wood that produced the ash in Samples F221/’97O/1 to F221/’97O/3 is likely to be of the same source. The import of wood either for fuel or for other purposes cannot, however, be ruled out. As noted above, the clay deposits may also have been imports. The clays (samples F270/’00Q, F4030A/’01J, F8049/’02P) may have been imported for any of a range of uses. In particular, the clays should be considered with respect to the crucible fragments found on the site. These appear to be composed of pipe clay with fragments of micaceous schist and an organic temper (Young, this volume, p.169). The schist fragments are consistent with a local origin, though that would include the coastal area near the island. The mineral element of the crucible probably most closely resembles samples F4030A/’01J and F8049/’02P, although the addition of schist to kaolin as a temper cannot be ruled out. The purity of the kaolin (sample F270/’00Q) is worth noting in this respect. Kaolin is known to have refractory properties, and it is possible that sample F270/’00Q represents the clay prior to preparation for the manufacture of crucibles (Rosenthal 1954). Another function for the pure and very well-sorted kaolin sample might be for making the fine, inner part of a mould (T. Young, pers. comm.). It is also worth noting that the use of fine white clays for whitening surfaces, cloth and textiles is well attested in the later historical period, although no contemporary parallels to the find at High Island are known.

292

10 discussion and conclusions 10.1 Introduction The archaeological excavations at the early medieval monastery on High Island uncovered a complex sequence of occupation that is not reflected in the scant historical documentation. The monastery and other archaeological monuments on the island have survived largely due to its inaccessibility and the fact that the island was long uninhabited. This has meant that the structural remains have escaped successive rebuilding, alteration and expansion, which occurred on many of the more accessible sites throughout the late medieval and subsequent periods, and they appear to have remained almost intact until the early decades of the 19th century. Since then, the monastery has fallen into a ruinous state, a process that continued virtually unimpeded until excavations and conservation works began in 1995. A little less than half the area inside the monastic enclosure was excavated. Any discussion of the results and the conclusions reached must be tempered with an understanding that the investigations were confined to the structures that required conservation and their immediate surrounds. Furthermore, the depth of excavation was limited by the requirements of conservation and in many areas the lowest levels were not revealed. Clearly, these restrictions led to fewer early features being found and a bias in favour of those from the later phases. This lack of information is compounded by the fact that much of the dating has had to rely on radiocarbon determinations that provide only broad date ranges, which frequently has made it difficult to link possible contemporary features. Whilst today High Island might be viewed as distant and remote, the sea in early medieval Ireland served as a highway and no doubt contacts would have been forged between neighbouring islands and coastal mainland communities. High Island lies within sight of Inishbofin, to which Colmán, abbot of Lindisfarne in Northumbria, travelled via Iona and established a monastery there in 665 (Robinson 1990, 23). Inishbofin continued as a seat of learning for three centuries. It is likely that influences and developments in architecture, sculpture and spiritual beliefs spread from Iona to Inishbofin and its neighbouring islands. A recent reappraisal of Françoise Henry’s excavations on Inishkea North, Co. Mayo, has identified a possible ‘zone or axis of contact’ between north-west Con-

nacht and north-west Scotland (Greene 2009, 350). These contacts are thought to have been sufficiently strong to suggest that this was a ‘region’ in itself, and it is likely that the monks on High Island were part of this common cultural zone.

10.2 The Iron Age

The earliest dated evidence for occupation uncovered in the excavations was from the Iron Age. These dates (360– 102 cal. BC, UB-6452 and 387–170 cal. BC, UB-4986) were obtained from a small deposit of peaty silt overlying the undisturbed boulder clay located outside the north wall of the church enclosure (Fig. 5.32). As this deposit was the only one excavated from this stratigraphic level, it is not known if other material of a similar date remains elsewhere on the site. It is conceivable that more extensive prehistoric settlement exists on High Island, particularly as the possible remains of a defended headland or promontory fort (GA021-033) have been recorded at the extreme eastern end of the island. At least two phases of prehistoric activity were uncovered in the excavations on nearby Omey Island (O’Keeffe 1994b, 17). Here, there were structures and associated shell middens where pottery, bone pins, antler and stone tools were dated by radiocarbon analysis to the early Bronze Age (1800–1500 BC). Further structures of possible later prehistoric date were also uncovered (Scott 2006, 27). As these excavations have not been fully published, however, and without access to the excavation archive, it has not been possible to consider in any great detail the relationships between the monastic settlement on High Island and its mother-house on Omey.

10.3 The origins of the monastery

The monastery on High Island is reputed to have been founded by St Féichín, whose death is recorded in 665 (Moran, this volume, p.16). The cult of St Féichín appears to have been widespread in west Connacht and he founded an important monastery at Cong. He is also credited with the establishment of the churches on Omey Island and on High Island, and a considerable number of holy wells were dedicated to him in this region. The main foundation at

discussion and conclusions 293

Monastic wall? Monastic wall?

Monastic wall? 6 8 10

5 11 1 4 3 2

Later structures

6 Wall (F91)

Lintels

1 Paved area (F245) and post-hole (F250)

7 Retained masonry (F882)

Paving

2 Grave 3

8 Rock-cut well (F8068)

Bedrock

3 Grave 4

9 Retaining wall (F863)

Obscured

4 Grave 5

10 Path (F8028)

Conjectural

5 Grave 8

11 Smithing pit (F352)

Limit of excavation

Fig. 10.1 The monastery showing pre-church paved area (no. 1), the earliest Graves (nos. 2–5) and other features; Phase 1a.

Fore was the chief monastery of a federation dedicated to the saint. It is recorded that Omey was obliged to pay tribute to the mother-house of Fore, and it is likely that High Island too was subject to its authority. There are only scant documentary sources relating to Féichín’s work in Connemara and the dates of his foundations are not known, but it is likely that Omey may have been established during the last decade of Féichín’s life (655–665). The sole source recording the foundation of High Island by Féichín is John Colgan’s life of the saint. It would seem that, particularly in the early period of occupation, the monastery on High Island was for those who withdrew from the community

on Omey to seek solitude as hermits. These were very holy men who imposed exile on themselves and were possibly comparable with the peregrini (‘pilgrims’) of the documentary sources (O’Sullivan and Ó Carragáin 2008, 318). The foundation of High Island is later than that of Omey, and most likely after the latter was well established. A date some time after the mid-7th century is therefore postulated for the establishment of a monastic community on High Island. No evidence of 7th-century occupation was recovered during the excavation and the earliest radiocarbon dates for the monastic settlement range from the third decade of the 8th to the latter part of the 10th century.

294â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Monastic wall Monastic wall (F4040)

Paving?

Monastic wall 6 8 10

5 11 1 4 3 2

Paving? 6 Wall (F91) - retained from primary enclosure?

Later structures Lintels

1 Slotted features (F260, F263, F269a-c)

7 Masonry (F882)

Paving

2 Grave 3

8 Rock-cut well (F8068)

Bedrock

3 Grave 4

9 Retaining wall (F863)

Obscured

4 Grave 5

10 Path (F8028)

Conjectural

5 Grave 8

11 Smithing pit (F352)

Limit of excavation

Fig. 10.2â&#x20AC;&#x201A; The monastery showing pre-church slotted features (no. 1), the earliest Graves (nos. 2â&#x20AC;&#x201C;5) and other features; Phase 1b.

The Atlantic seaboard of Ireland has many islands that bear witness to an ecclesiastical past, and High Island is just one of a large number of church sites along the west coast (Fig. 1.1). Modern perception of the remoteness of islands does not necessarily tally with how they were seen in the past when sea travel was an essential means of communication. Island and coastal settlements were particularly important during the early medieval period. It is likely that High Island would have been cut off from the mainland for considerable periods of time, at least in winter, because of its location. The practicalities of life on the island led to significant physical hardships, which no doubt suited the ethos of an eremetical foundation.

10.4 The monastery during Phase 1: 8th to mid-11th century Two separate building episodes were identified as belonging to this phase. The first episode comprised an area of paving that had possibly belonged to a table altar or leacht around which burials were interred (Fig. 10.1). These features were contained within a rectangular enclosure that demarcated the most sanctified space within the early settlement. Domestic and industrial activities were located in the outer monastic enclosure, including metal-working in the eastern area close to the well, which was accessed via a stone path. After a period of use, the primary structure relating to the paved area fell out of use and there was an

discussion and conclusions 295

(a)

(b) Pl. 10.1 (a) and (b) Trahanareear, Inishmurray, Co. Sligo: (a) view of the excavation site showing the primary paving that was discovered beneath the leacht; and (b) a detail of the central stone socket for an altar post or wooden cross (after O’Sullivan and Ó Carragáin 2008, pl. 65).

296 high island: excavation of an early medieval monastery

apparent hiatus in the ecclesiastical occupation. However, the site does not appear to have been totally abandoned as debris from the remains of domestic hearths accumulated over the paving indicate nearby activity, although no structures were uncovered. Following this possible hiatus, this area regained its ecclesiastical focus and a structure, possibly a leacht or perhaps a more substantial building, was erected within the enclosure in close proximity to the burials (Fig. 10.2). The rectangular enclosure wall was rebuilt to a trapezoidal plan also at this time. These features were enclosed within the monastic enclosure wall, one of the enduring symbols of early medieval ecclesiastical sites. The date when the outer enclosure wall was erected is unclear, but it is likely to have been in place by the end of the 10th or the beginning of the 11th century at the latest.

10.4.1 The church enclosure The paved area It is likely that the paved area and its associated features, most notably the post-hole, formed the core of the earliest phase of the monastery (Figs 10.1, 10.5, 10.9; Pl. 5.1). The proximity of the graves, coupled with their location beneath the later church, strongly suggests that these remains were ecclesiastical. Unfortunately, the limited extent of the surviving remains makes it difficult to establish their nature with certainty. A small area of paving and a single post-hole are insufficient evidence upon which to reconstruct a ground plan of a potential building, if indeed these features belonged to a building. The posthole may have held a post for a wooden table altar. There is literary evidence for wooden churches with table altars in Ireland in the mid-7th century (Herren 1974, 187). There is limited archaeological evidence for timber structures at excavated sites, however, because these leave little trace and frequently later burial has disturbed or removed the remains (Ó Carragáin 2010, 19). On Inishmurray, excavation beneath one leacht (Trahanareear) revealed the remains of a cell, in front of which was a small area of paving with a central post-hole. This post-hole has been interpreted as having held a timber upright for a wooden table altar (Pl. 10.1(a); O’Sullivan and Ó Carragáin 2008, 218–19, 236). It is postulated that this example was probably erected some time before the end of the 10th century. Another example has been found at Caherlehillan, Co. Kerry. Whilst acknowledging that the High Island evidence is considerably more difficult to interpret, the possibility exists that the paved area and associated post-hole at this earliest level may represent the partial remains of a table altar. No clear dating evidence was retrieved from these early levels on High Island, but radiocarbon analysis of samples extracted from the deposit of burnt debris (see below) that sealed the paved

area provides a terminus ante quem for this feature of late 8th to early 11th century.

The apparent hiatus in ecclesiastical use Analysis of the redeposited material found overlying the paved area indicated that it had been formed by intermittent dumping of domestic waste, predominantly hearth residues (Figs 5.4, 10.5; F221). Analysis of this debris established that it had built up over a considerable period of time, estimated at between 40 and 200 years, and it was further concluded that the material was not burnt in situ. This evidence is significant as it implies that after the paved area, or the structure with which it was associated, fell into disuse, settlement continued close by or within the abandoned structures. Radiocarbon dates from the debris produced the range of cal. AD 770–980 (OxA-8946, 1182±36 BP) and cal. AD 860–1020 (OxA8918, 1110±40 BP), indicating that this activity took place between the late 8th and the early 11th century. This date range corresponds closely with the radiocarbon determination obtained from similar material recovered at the northern side of the church enclosure wall. Insights into the economy of the people who deposited this material were provided by analysis of the faunal and fuel remains. The faunal report produced evidence for a range of species, evidence that was also present in other deposits and in later phases. However, the sample also included five pig bones (McCarthy, this volume, p.244) that, with the exception of a single example from a Phase 2 deposit (F478), are the only pig bones recovered from the site. The largest assemblage of pig bones (24) was recovered from the hearth deposits (F93a–i) abutting the outer face of the northern side of the church enclosure wall, and one bone was recovered from a fill within the primary monastic enclosure wall (F93a–i, F4030). The fuel analysis report reflected a somewhat similar trend. Whilst a significant amount of wood ash was identified within the redeposited debris (F221), a small but recognisable amount of fine mineral ash, the result of burning lacustrine mud, was also present (Lancaster, this volume, p.284). The use of lacustrine mud as a fuel was recognised only in Phase 1 levels. Small amounts of lacustrine mud were also found within a potential old ground surface and in one of the small hearths therein (Fig. 5.2, F232, F246). Quite significant quantities were identified in the deposits (F93a–i) outside the northern side of the church enclosure wall, the same deposit where the pig bones were found. The radiocarbon determination of cal. AD 728–971 (UB-4522, 1171±41 BP) from these deposits produced dates ranging from the early 8th to late 10th century, which sits comfortably with that obtained from the redeposited debris (F221). Comparable examples of mud used as fuel have been found in the northern Isles of

discussion and conclusions 297

Scotland, where there were similar environmental conditions (Carter 1998, 99–103). The presence of pigs and the use of lacustrine mud as a fuel appear to be restricted to Phase 1 activities in the period between the early 8th and late 10th century. This may indicate the temporary presence of a secular population at this time or, alternatively, perhaps the monastic community had adopted a slightly different economic regime from that of the later developed monastery. Parallels for a possible hiatus in other ecclesiastical sites are not easily found. Excavations at St Féichín’s foundation on Omey Island revealed a sequence of burials that extended over several centuries, possibly beginning in the 7th century (O’Keeffe 1992a, 3; 1994b, 16–17). The burial activity at this site was not continuous, perhaps reflecting a shifting pattern of ecclesiastical stability in the area. It is not unreasonable to suggest, therefore, that this situation was mirrored at St Féichín’s other monastery, on High Island. No radiocarbon dates have been published for the early medieval period from Omey Island, which means close comparisons to particular phases of activity are not possible. O’Keeffe has suggested that a possible reason for one of the periods of abandonment of Omey Island was the presence of Viking activity in the area (O’Keeffe 1992a, 3). Viking activity There are no references to Viking raids on the monastic site on High Island, however raids in the 9th century are known to have occurred on nearby islands and on the coastal mainland to the north and south (Gibbons and Kelly 2003, 31). In 807, the monasteries on Inishmurray and at Roscam, Co. Galway, was attacked and five years later the Conmaicne Mara were defeated in battle (Moran, this volume, p.16). A Viking burial was discovered in 1947 at Eyrephort, near Clifden (Fig. 1.1; Raftery 1960) and its subsequent re-evaluation indicates that Scandinavian activity in this area may have been more extensive than previously thought (Sheehan 1988). Limerick became the chief base for Viking operations in Ireland in the 9th century (Chadwick 1975, 21–2). Vessels en route from Scandinavia to Limerick would have passed Slyne Head, a prominent landmark lying less than 20km south of High Island (Fig. 1.1). The Old Irish name for Slyne Head was Léim Lára (‘the mare’s leap’) (Ó Tuathail 1948– 52, 155–6). This placename also occurs in the 12th-century Icelandic text Landnámabók as Jölduhlaup (‘the mare’s leap’), frá Reykanesi á sunnarverðü Islandi er fimm dœgra haf til Jölduhlaups á Írlandi (‘from Reykjans in southern Iceland there is a sailing of five days to Jölduhlaup in Ireland’) (Benediktsson 1968, 32–4), suggesting trading and cultural relations between Scandinavia and this area of the west coast (Sheehan et al. 2001, 112).

Soapstone (steatite or talc schist) was a commodity much favoured by the Vikings, and deposits occur in a band from Inishshark, through Inishbofin, to Croagh Patrick (Fig. 1.1; Gibbons and Gibbons 2006, 26). In this region, evidence for the working of soapstone and for the manufacture of spindle-whorls and buttons was found on Inishkea North (Henry 1951, 75). Scandinavian activity along Ireland’s western seaboard in the 10th century has been further suggested by the reassessment of the excavation results from Beginish, Co. Kerry, where a soapstone bowl was recovered (Sheehan et al. 2001). The absence of soapstone in the finds assemblage from High Island could be interpreted as an indication that this material was not valued or exploited. Significantly, research into soapstone quarries in Scandinavia suggests there is little correlation between the location of quarries and the concentration of finds. Whilst it may be expected to find soapstone artefacts in the vicinity of a quarry site, this was not normally the case. It has been suggested in regard to the Scandinavian finds that most of the soapstone vessels that have been found were traded (Resi 1987, 96). Viking activity along the western seaboard could account for the abandonment of High Island by the religious community for a period between the late 8th or late 9th century and the late 10th century. Whilst it would be unwise to draw any definite conclusion from this, the people who occupied the monastery during this time may have had a different economy from the monks, as indicated by aspects of their diet and their use of fuel. The communities on Inishmurray and Inishglora, Co. Mayo, fled from Viking raids to mainland monasteries (Moran, this volume, p.16) and it is possible that the monks on High Island also moved temporarily to the mother-house on Omey Island. Renewed building activity A second episode of building activity at the focus of the settlement on High Island is represented by two stonelined slots and a number of in situ upright stones within a possible third slot (Fig. 10.2; Pl. 5.2). A possible interpretation is that the slots are the remains of an altar belonging to a small church. This altar would probably have been centrally placed along the east wall, which was the norm in small, unicameral pre-Romanesque churches, implying that the building to which the slots belonged had the same layout and orientation as the stone church that presently stands on its site. If this was the case, the original building would appear to have respected the graves to the east, though the relationship between the two is uncertain. Another possible interpretation is that the slots belonged to a free-standing leacht. The terms leacht and altar have in the past been interchangeable, but this is no longer accepted as not all leachta were altars. Leacht is

298 high island: excavation of an early medieval monastery

used in this case to indicate an open-air structure where resident monks and visiting pilgrims may have gathered to pray, and where clergy performed the sacrifice of the Eucharist in proximity to earlier burials. If these features belonged to a leacht, its function would have been primarily as a focus at the spiritual core of the monastery, around which people would have gathered to pray and honour the memory of the dead. The potential leacht was located less than 1m west of the graves (Fig. 10.2, nos. 1–4) and could conceivably be interpreted as having been built to honour the person or persons interred. A close parallel is found on Omey Island, where a number of lintelled burials were uncovered at an early level adjacent to and, in one case, beneath a drystone leacht. The leacht was orientated somewhat differently from the enclosure, suggesting that they may not be contemporary. O’Keeffe compares the leacht and enclosure on Omey with the features on High Island (O’Keeffe 1992a, 2, 4). Comparison can be made between the leacht at Trahanareear on Inishmurray and the slots found on High Island. At Trahanareear, the second phase in the development of the monument saw the erection of a leacht directly over the earlier paving and central post-hole (Pl. 10.1(b)). The authors suggest that the leacht was erected immediately after the post was removed because the exact position would have been forgotten if a substantial time lapse had been allowed to occur (O’Sullivan and Ó Carragáin 2008, 236). If the slots on High Island were part of a leacht, the fact that it was located a short distance from the earlier paved area may be explained by the hiatus that occurred between its functional period and the erection of the leacht. During this time, the exact position of the earlier structure may have been forgotten and when the site was reoccupied, the graves may have been the only visible symbols marking the focus of the site. Analysis of the leachta on Inishmurray showed that a characteristic feature was a basal course of thin, edge-set slabs, which formed the outer faces of the leachta at foundation level (ibid. 297). Due to the limited nature of the evidence from High Island it is difficult to draw comparisons, but edge-set stones (F269a– c) survived in situ while the slots clearly would have held some form of upright stones. Most leachta were marked by a cross or cross-inscribed slabs (Thomas 1971b, 169). In this context, it is noteworthy that the later altar built over the slots incorporated within its basal course a cross-inscribed stone (Section 7.1, Cross-slab 23) that could originally have marked a leacht. It is also possible that this stone was a reused grave-marker, as was found incorporated in the altar of the 8th-century stone chapel at Ardwall Isle, Kirkcudbright, south-west Scotland (Thomas 1971b, 181).

The graves Three burials were uncovered located to the east of the

paved area and the slotted features (Fig. 10.2, Graves 3–5). To the north, a decorated recumbent grave slab presumably also covered a burial, but was not excavated (Fig. 10.2, Grave 8). It is likely that Graves 3, 4 and 5 also had recumbent slabs as at this date graves normally had a covering of stone (E. O’Brien, pers. comm.). The alignment of the burials was somewhat varied and it is possible that the graves were laid out relative to one or more structures with different orientations. The sequence of burial established at the early medieval cemetery on Ardwall Isle suggests a somewhat similar scenario. On this site, the earliest phase of burial related to a focus, possibly the slabshrine (Thomas 1967, 127). Burials aligned on the axis of a later small timber structure, possible an oratory or chapel, were revealed together. Other burials laid out on the axis of a possible 8th-century and larger stone chapel formed the latest phase. The earliest burial on High Island was that of an older male, aged over 55 years, placed in a simple earth-cut grave (Fig. 5.16, Grave 5; Pl. 5.22). This individual was dated to cal. AD 881–977 (UB-3999) and presumably was a distinguished member of the community, having been granted burial at the core of the monastery. His grave may have been the sole focus of veneration for approximately a century. Just to the south, two stone-lined graves (3 and 4) were constructed to receive the remains of two men. An older man aged over 55 years was dated to cal. AD 980–1025 (Grave 3) and the individual in Grave 4 was dated to cal. AD 980–1023. Graves 3 and 4 were exceptionally well made, to the extent that the skill and craftsmanship employed set them apart from others. Both were stone-lined and had a paved floor (Fig. 10.3(a); Pls 5.15, 5.17). It is tempting to suggest that because two grave types were identified in the same cemetery, this might have some chronological significance. However, this diversity of burial rite has been identified in other excavated sites, for example at Reask, Co. Kerry, where it was indicated that this was not the case (Fanning 1981, 151). This is supported by a recent and more widespread appraisal of burial rites in the early medieval period, which concludes that although patterns can certainly be discerned, it would be unwise to date excavated burials by burial rite alone (O’Sullivan et al. 2008, 158). Notwithstanding this, on Omey Island two sequential but separate cemeteries were identified (Scott 2006, 28, 33). In the earlier cemetery, all but two of approximately 300 burials were located in simple earth-dug graves, whilst many of those in the later cemetery were in slab-lined graves, thus suggesting that the grave type seems to reflect a change over time in mortuary practice rather than a differing status among the dead. There were similarities in the burial rite of Graves 3 and 4. Quartz pebbles were found in the two graves and

discussion and conclusions â&#x20AC;&#x192; 299

4 7

1 1 Grave 5 with burial (F37) in situ beneath recumbent slab (missing) 2 Grave 4 with burial (F40) in situ 2

3 Grave 3 with burial (F70) in situ

4 Headstone (CS 5) for Grave 3 8

(a)

5 Footstone (CS 4) 6 Footstone (CS 7)

7 Headstone (CS 9) for Grave 5

8 Footstone (CS 11) 9 Slabs (F67) covering Grave 3 10 Headstone (CS 8), within church wall, west end Grave 4 11 Grave 2, recumbent slab (CS 3) 9

12 Headstone (CS 2)

13 Recess (F51) for headstone (CS 2) above pseudo Grave 2 14 Grave 3, recumbent slab (CS 6) 15 Grave 4 (recumbent slab and footstone missing) 16 Grave 5 (recumbent slab F23)

(b)

17 Skull of burial (F70), Grave 3 18 Skull of burial (F40), Grave 4

19 Skull of burial (F37), Grave 5

Suggested position

Upright stones Quartz CS

Cross-slab Conjectural

Obscured

Bedrock

16 5

(c)

10 13 4 7

(F62)

(d)

Fig. 10.3â&#x20AC;&#x201A; Conjectural development of pseudo Grave 2 and Graves 3, 4 and 5: (a) suggested layout of Graves 3, 4 and 5, Phase 1; (b) Graves 3 and 4 during restructuring, Phase 2; (c) pseudo Grave 2 and Graves 3, 4 and 5, Phase 2; (d) elevation of exterior east wall of church and earlier features in area of pseudo Grave 2 and Graves 3, 4, and 5.

300 high island: excavation of an early medieval monastery

both burials had been laid on and were covered by a shallow layer of sand. There is no known source of sand on High Island, suggesting it was imported. The sand, with its distinctive dark-coloured flecks, probably originated from Omey. If this was the case, it must be asked why sand would have been imported for this purpose and what spiritual belief could lie behind it. It has been suggested that the holy men buried in these graves on High Island may have belonged to or had a particular association with Omey (E. O’Brien, pers. comm.). Perhaps these men had requested burial in ‘their own land’, but as circumstances may have prevented this, sand was brought from Omey to High Island to allow fulfilment of their wishes. This practice is recorded in Irish hagiographical texts, ‘in which the saint brings soil from the holy sites of Christendom to spread in his cemetery’ (Ó Carragáin 2010, 83). Parallels for burials laid in sand are rare. A burial uncovered at a possible early medieval farmstead at False Bay in the townland of Truska, south Connemara, is thought to have been placed above, though not directly upon, a whitish sandy layer, whilst the surface of both interments, which had been placed together in a single shallow grave, was visible as a roughly semicircular area of white sand (Gibbons and Kelly 2003, 28). The directors of the excavations suggest that the white sand at the surface may have been deposited deliberately to mark the position of the grave. In this case, however, sand was a readily available commodity. It is hardly a coincidence that the dates of these two graves, which were elaborately built and embellished and located in the most prominent position at the core of the monastery, correspond with the death of Gormgal (AD 1018). It is possible that one or other grave contains his remains. His obit, Gormgal in Ardailean, prim-anmchara Erenn, in Christo quieuit (‘Gormgal of High Island, chief anmchara of Ireland, rested in Christ’), is found under the year 1018 in the annals. The title, anmcharae, means ‘spiritual advisor’ or ‘confessor’, and prim-anmcharae Erenn relates to his high reputation throughout Ireland and might suggest that Gormgal was abbot of High Island. It also suggests that the monastery was accepting penitents under his authority and his esteem could have attracted a substantial number. This reference to Gormgal is the sole mention of High Island in the annals, and the text of the entry indicates that it was he, rather than the monastery, that was of significance. The identity of the second individual is less clear. A reference to the death of Muirchertach Mac Líacc, ard-ollam Éireann (‘leading chief poet of Ireland’), recorded at 1014 while he was under the monastic rule on High Island is significant (Moran, this volume, p.26). His date of death corresponds closely with that of Gormgal and falls within the date range for the burials in Graves 3 and 4.

The church enclosure wall There were two phases of construction in the church enclosure wall. The primary enclosure was probably rectangular with rounded external corners and is estimated to have enclosed an area 9m east–west by 8m north–south (Fig. 10.1). The northern side of this enclosure was rebuilt inside the line of the earlier wall (Fig. 10.2). The foundations of the primary wall face (F91) were retained and the space between the two structures was used for the disposal, or possibly the storage, of the residues from domestic hearths (Pl. 5.30). The slight kink in the later wall can be explained by the need to accommodate the retention of the foundations of the earlier wall. The rebuilt wall surrounded a reduced area measuring 9m east–west by approximately 7m north–south, resulting in a trapezoidal plan. The dating of both phases of construction of the church enclosure wall is unclear, but the date range of cal. AD 728–971 from the hearth debris provides a terminus ante quem for the primary wall. The High Island monastery had three enclosing walls: the church enclosure wall, surrounded by a monastic enclosure wall and a wall extending to the east and west from the monastic enclosure to the cliff edges (Figs 4.1, 4.3, 5.1; Pls 4.2, 4.8). This latter wall essentially cut off the south-western part of the island, including the monastery, from the remainder of the island. Multiple enclosure walls were a feature of many early ecclesiastical sites and their function has been interpreted as a way of demarcating the most holy areas (Edwards 1990, 116). Enclosure walls were also used to block out the outside world. Bede’s account of Cuthbert’s hermitage on the island of Lindisfarne, Northumbria, makes it clear that the function of the enclosing wall served to direct the hermit’s gaze away from the world. … the wall is higher than a man standing upright, but inside he made it much higher by cutting away the living rock so that the pious inhabitant could see nothing except the sky from his dwelling, thus restraining both the lust of his eyes and the thoughts and lifting the whole bent of his mind to higher things … (cited in Herity 1995a, 21) Whilst the church enclosure wall on High Island has not survived to its full height, mounds of rubble on both sides indicate that originally it would have stood considerably higher. This height, estimated to have been approximately 1.5m, would have been sufficient to block out the world from the view of anyone within. Excavations on Omey Island revealed at least three enclosures (O’Keeffe 1992b; 1994b, 15). The earliest enclosure, tentatively suggested to have been of wood, contained a substantial cemetery for laypeople. The second enclosure, rectangular in shape, was surrounded by a stone wall. The third enclosure, also with a stone wall,

discussion and conclusions 301

Fig. 10.4 Early ecclesiastical site at Cappanagroun, Co. Kerry (after O’Sullivan and Sheehan 1996, Fig. 164).

Pl. 10.2 The church and enclosure, Caher Island, Co. Mayo (G. Scally).

was smaller in size and it was probably trapezoidal in plan, with rounded corners both externally and internally. Similarities exist, as the shape of the first church enclosure on High Island was rectangular and the second was trapezoidal (Figs 10.1, 10.2). High Island is a rare example of a monastery where the church is surrounded by a freestanding inner enclosure that is entirely separate from the larger monastic enclosure (Fig. 5.1). Large mainland sites, while having evidence of concentric enclosing walls, do not seem to have had smaller enclosures around churches. The sites that compare closest, Beginish, Cappanagroun and Kildreelig, all

in Co. Kerry, enclose not oratories or small churches but platforms of stone or shrines (Fig. 10.4; O’Sullivan and Sheehan 1996, 261, 265–6, 293–4; Ó Carragáin 2003a, 141). A number of island sites display evidence for an inner enclosure surrounding a church, for example, Caher Island, Co. Mayo, Rathlin O’Birne, Co. Donegal, and Inistooskert, Co. Kerry (Herity 1995b; Cuppage et al. 1986, 296–8). Caher Island compares closely in layout with High Island in that the inner rectangular enclosure surrounds a late medieval church on the site of an earlier one, three leachta and a larger platform, and a grave, known as Leaba Phádraig (Pl. 10.2; Herity 1995b, 108). As on High Island,

302 high island: excavation of an early medieval monastery

South Z2

North Z4

Z3 12 9

2 6

7 4

3 5

14 10

1 Paving (F80)

13 Wall-face (F4025 SS) of extended monastic wall

2 South wall of church enclosure

14 Wall-face (F4040 S) of primary monastic wall

3 Paved surface (F29)

15 Wall-face (F4025 S) of extended monastic wall

4 Boulder clay (F271)

16 Wall-face (F4040 N) of primary monastic wall

5 'Paved area' (F245)

17 Wall-face (F4025 N) of extended monastic wall

6 Burnt debris (F221)

18 Interior of sub-rectangular building (F427)

7 Paving (F29)

19 North wall of sub-rectangular building (F427)

8 Church

Structures

9 North wall of church enclosure

Paving

10 Wall (F91)

Conjectural

11 Paving (F422)

Limit of excavation

12 Stepped paving of passage (F4028) to sub-rectangular building (F427)

Fig. 10.5 North–south section through the excavated area of the monastery. Location of section Z2–Z3 on Fig. 10.7; Z3–Z4 on Fig. 10.10.

sufficient space for unrestricted passage of a person to walk between the church and the enclosure wall was apparent. The provision of space to walk around the church seems to have been important at these sites. On Caher Island, this space has been interpreted as leaving room for pilgrims to circulate freely in the performance of the pilgrimage round or turas (Rourke 2009, 134). The oratory enclosure (10m x 8m max.) on Rathlin O’Birne is also comparable to High Island in that the oratory was surrounded by a wall enclosing a polygonal-shaped area (Herity 1995b, 87). Similarly, on Inishtooskert, the oratory, together with stone crosses and a leacht, are contained within a wedgeshaped enclosure.

10.4.2 The monastic enclosure The rock-cut well The rock-cut well, its retaining wall and a paved path, all located in the extreme east of the monastic enclosure, appear to be contemporary features (Fig. 10.1; Pls 5.34– 5.36). The well would have been essential to provide the community with a source of fresh drinking water and, as such, may have been one of the earliest features constructed. It was finely built, indicating some knowledge of water management. It is possible that parts of the path remained in use for the entire period of the monastic settlement and possibly for centuries afterwards. The full extent of the path is unclear, however its south-westerly direction indicates that it may have extended as far as the metal-working area revealed beneath Cell B.

Metal-working Evidence for the working of iron and other metals was uncovered in the east of the monastic enclosure underneath Cell B. The remains have been interpreted as a smithing hearth pit along with a series of contemporary working surfaces in the vicinity (Fig. 10.1, F352). A radiocarbon date of cal. AD 870–1030 (OxA-8917) from hazel charcoal in the pit showed that the smithing hearth pit was operating some time between the late 9th and the early 11th century. Smithing is the process of refining the bloom produced by smelting, by reheating and hammering it to remove excess slag and impurities followed by the final phase of making tools and utensils. No evidence for smelting was found on High Island and it is probable that the bloom was transported to the island in a semi-processed condition. Less archaeological evidence has been identified for smithing than for smelting (Edwards 1990, 88). On High Island, the fact that smithing of iron was carried out together with the working of copper and copper alloy may indicate that it was done by a specialist smith. The crucible found in the smithing hearth pit on High Island has been tentatively identified as belonging to a bell-shaped, possibly lidded form, with a suggested 7th to 10th century date (Young, this volume, p.169). Documentary sources demonstrate that the early medieval Irish blacksmith was held in high regard. The archaeological evidence, which mainly consists of slag and sometimes smelting furnaces or smithing hearths

discussion and conclusions 303

and other debris, shows that some iron-working was carried out on almost every site, both secular and ecclesiastic (Edwards 1990, 86). It is probable that some monks may also have been skilled metal-workers considering that in the Vita Columbae (VC11.29), Adomnán refers specifically to ‘the monks who knew the blacksmith’s craft’ (McCormick 1997, 60). Evidence for metal-working is routinely found at large mainland monastic sites, such as at Clonmacnoise, at Kilpatrick, Co. Westmeath, and at Armagh and large island sites, such as Iona in Scotland (O’Sullivan et al. 2008, 153). Evidence for pre-10th-century metalworking has also been uncovered at small island sites, such as Beginish, Church Island, Illaunloughan, all in Co. Kerry, and Inishkea North, Co. Mayo (O’Kelly 1956, 182; 1958, 59–60; White Marshall and Walsh 2005, 19–22; Henry 1945, 141). The iron-working evidence on these small Atlantic islands is significant as it indicates that, unless the raw material was present on the island, iron ore had to be obtained from the mainland as well as fuel for the furnace. O’Kelly (1958, 117) remarked that since almost every site that has been excavated has produced evidence of iron-working, this suggests that smelting was carried out even on remote islands by the people living on them. On Church Island, O’Kelly (ibid. 116) suggested that iron was smelted within the early circular wooden house. However, more recent excavations have indicated that the iron-working pre-dates this structure (Hayden, forthcoming). This is a similar situation to the iron-working evidence found at High Island, where the metal-working pre-dated the construction of Cell B. The monastic enclosure wall When it was complete and at its original height, the monastic enclosure wall on High Island was a substantial, imposing structure that would have dominated the site. This would have been no different from other contemporary ecclesiastical enclosures. Their primary function was to define the limits of the settlement and the cemetery, demarcating an area of sanctuary (MacDonald 1997, 42; McErlean and Crothers 2007, 8). It has been demonstrated that Irish ecclesiastical settlements were designed to a set pattern, with a sacred core around which were areas of sanctuary that decreased in importance the further they were from the centre (Doherty 1985, 57). The monastic enclosure wall at High Island defined a spiritual boundary that may also have been a legal boundary. The monastic enclosure wall encompassed a roughly oval area of approximately 27m north–south by 32m east–west (Fig. 5.1). There were two phases of construction evident in the northern section of the wall. The primary wall was extended to form a wider and more substantial wall during Phase 2 (Figs 10.1, 10.5, 10.7; Pl. 5.47). Both phases of the wall construction comprised a bank of

soil and small stones, which was revetted on both faces with stone. The original height of the wall is unknown, but where the southern wall ran alongside the lake it is estimated that it may have stood approximately 1.5m in height. The main entrance to the monastery was located at the south-east and two other entrances, at the south-west and north-east, probably also belong to the primary phase of construction, but as excavation was not undertaken in these locations, this could not be verified (Fig. 5.1). The wall enclosed all the main features of the monastery and two buildings abutted it at the south-east entrance, one of which was probably a guesthouse and the other a porter’s lodge. The southern wall of the enclosure flanked the edge of the lake, which appears to have been artificially extended up to the edge of the wall to restrict access to the western part of the island. No evidence was obtained with which to date unequivocally the primary phase of the wall. Charcoal extracted from deposits that were later than its construction has provided a date range of cal. AD 778–1000 (UB-4987), suggesting a terminus ante quem of the late 10th century. The monastic enclosure wall contained two chambers in its western side and a third at the main south-east entrance (Fig. 5.1). The largest chamber, that at the south-west, was investigated and it appears to have been constructed as an integral part of the primary wall, but no evidence was retrieved to indicate its function. The low entrance and difficult access could indicate its use as a place of refuge, but its cool, dry environmental conditions would also have been suitable for the storage of foodstuffs, such as grain, dairy products, etc. Chambers within enclosure walls are known only from a small number of early medieval monastic sites. The substantial wall at Illauntannig (Co. Kerry) has a wall chamber that is linked to a hut inside the enclosure by an underground passage (Cuppage et al. 1986, 294). On Inishmurray, the monastic enclosure wall houses eight chambers, while on Caher Island the wall incorporates a single chamber (O’Sullivan and Ó Carragáin 2008, 68; Herity 1995b, 106).

10.5 The monastery during Phase 2: mid11th to late 12th/early 13th century

There were at least two periods of intense building activity at the monastery between the mid-11th and late 12th/ early 13th century. The structural works, particularly in the earlier period, were extensive and may have been carried out in response to a significant event. Perhaps the death of Gormgal, along with that of another esteemed member of the community, in the early 11th century was the catalyst for these developments. The beginning of the first phase was marked by the construction of the stone church at the centre of the monastery (Figs 5.1, 10.6). The continuity

304â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Monastic wall Monastic wall (F4049)

Paving? Paving? 9

Monastic wall

10 2 3

12 11

15a Church

16 13

15c

15d

15b

1 Graves 1â&#x20AC;&#x201C;8

10 Grave 11 with burial (F8026) in situ

2 Well (F8068)

11 Trough (F845)

16 Buttress (F817) Later structures/features

3 Path (F8028)

12 Channel (F8052)

Lintels

4 Retaining wall (F863)

13 Soakaway (F846)

Paving

5 Paved surface (F29)

14 Wall-face (F832)

Bedrock

6 Cell B

15a Paving (F835)

Upright stones

7 Church enclosure wall

15b Paving (F80)

Conjectural

8 Wall (F91)

15c Paving (F833)

Obscured

9 Passage (F414/F8040)

15d Paving (F833a)

Limit of excavation

Fig. 10.6â&#x20AC;&#x201A; Plan of the monastery showing features associated with the primary period of expansion in the 11th or early 12th century; Phase 2a.

of structural activity in this location testifies to its enduring sanctity as the spiritual focus of the settlement. At the same time as the church was built, the special graves were restructured both to facilitate their enhancement and to permit further burial outside the east wall. The surface between the church and its enclosure was paved at this time. Cell B was also built during this time and the adja-

cent part of the church enclosure wall was dismantled and rebuilt to provide a new entrance at the north-east corner (Fig. 10.6). A rock-cut trough was constructed in order to store water for those using Cell B and the rock-cut well continued in use. The second period included the dismantling of the northern side of the monastic enclosure wall, which was extended in width when Cell A was erected

discussion and conclusions 305

G G

13 28

2 24

25a 26

29 Z

12a 26

Church

10 27 11

12c 12d

20 12b

12e

Z2 0

5m 1 Graves 1–8

12e Paving (F615a)

26 Clay (F803)

2 Well (F8068)

13 Extended monastic wall (F4025)

3 Path (F8028)

14 Cell A

28 Passage (F437)

4 Retaining wall (F863)

15 Annulus (F428)

29 Grave 9 with burial (F208) in situ

5 Paved surface (F29)

16 Water-collection and drainage channel (F8025)

Later structures

6 Cell B

17 Lintelled drain (F8007)

Lintels

7 Church enclosure wall

18 Leacht (F887)

Paving

8 Retained wall (F91)

19 Leacht (F843)

9 Trough (F852)

20 Boundary wall (F813)

10 Soakaway (F846)

21 Paved path (F875)

Upright stones

11 Wall-face (F832)

22 Paving (F422)

Conjectural

12a Paving (F835)

23 Blocked passage (F414/F8040)

Obscured

12b Paving (F80)

24 Wall (F848)

Limit of excavation

12c Paving (F833)

25 Grave 11 with burial (F8026) in situ and stone-lined grave (F856) above

12d Paving (F833a)

27 Buttress (F817)

Annulus G

Granite

25a Sod (F850)

Fig. 10.7 Plan of the monastery showing features associated with the later period of expansion in the 11th or early 12th century; Phase 2b. Section Z–Z1, see Fig. 10.9; Z2–Z3, see Fig. 10.5.

306 high island: excavation of an early medieval monastery

within its thickness (Fig. 10.7). A sophisticated water management system was constructed along the eastern side of the monastic enclosure, consisting of a drainage channel that connected with the earlier rock-cut trough.

10.5.1 The church enclosure The church The church is a small, plain, pre-Romanesque church without antae. It has been argued that such churches can be dated to the 11th or very early 12th century (Manning 2009, 277). Archaeological dating of the church relies partly on its relationship to the graves outside its east end. It was constructed after the latest of the Phase 1 burials, but before the burial in Grave 1. The Phase 1 burials had a date range from the late 9th to the early 11th century, and the individual in Grave 1 was interred sometime between the mid-11th to mid-12th century. This is not at variance with the dating suggested above for such churches. The construction of a stone church on the site of an earlier structure, as occurred on High Island, was commonplace. Such continuity has been revealed at Illaunloughan, Church Island and Caherlehillan (White Marshall and Walsh 2005, 32–6; Hayden, forthcoming; Sheehan 2009, 196). Continuity of occupation was also apparent at Skeam West, Co. Cork, where burials of the mid-5th to mid-8th century pre-date the construction of the mortared church (Cotter 1995, 72–3), and similar continuity was also evident on Church Island and at Reask

(Fig. 10.8; O’Kelly 1958, 91–3; Fanning 1981, 79–84). On Iona, St Ronan’s church overlay the walls of an earlier building, which in turn overlay earlier burials (O’Sullivan 1994, 327). A series of burials associated with a primary shrine at Whithorn evolved through seven stages of development and remained the focus of the site (Hill 1997, 89). The implication of such continuity is that the location of the stone oratory or church was usually determined by the site chosen as the focus for the primary cemetery (Thomas 1971b, 48–90). Continuity of the foci of ecclesiastical settlements frequently had the effect of influencing the orientation of later structures, for example, at Teach Molaise on Inishmurray (O’Sullivan and Ó Carragáin 2008, 320). On High Island, the orientation of the church was east– west, which diverged up to 20° from the graves outside the east wall that pre-date it. This shows that the orientation of the earlier features was not in all cases perpetuated in that of the later stone church. Another example is the drystone oratory on Church Island, which differed in orientation from the buildings over which it had been built (O’Kelly 1958, 61). The church on High island, which is amongst the smallest of the Atlantic island mortared stone churches, has an internal ratio of 1:1.1, which differs considerably from the most common ratio of about 1:1.5 for early medieval Irish churches, as noted by Leask (1955, 49). A more comprehensive study of 91 pre-Romanesque churches showed that they ranged from 1:1.1 for High Island to 1:1.86 as at

Fig. 10.8 Plan of the oratory, Reask, Co. Kerry (after Fanning 1981, Fig. 5).

discussion and conclusions 307

Lynally, Co. Offaly (Ó Carragáin 2010, 110–13). The closest parallel to the church on High Island is the second phase of St. Michael’s Church on Skellig Michael, with a ratio of about 1:1.3 (O’Sullivan and Sheehan 1996, 283). Reasons why the dimensions of the High Island church are so much at odds with others can be partly attributed to the constraints imposed by its location within a preexisting enclosure together with the desire of the builders to incorporate the headstones from special graves into its foundations (Figs 10.3(d), 10.6, 10.9). The church was altered at least twice and possibly three times (Figs 5.5, 5.8). The walls and the doorway were raised and a decorated cross-slab was reused as the lintel above the doorway. Later, the walls were again heightened and the roof replaced. This is indicated by the large stones that survive on the upper levels of the side walls and the change in the masonry evident in the west gable. Earlier writers have speculated on the date of the final phase of rebuilding. Herity suggests that the masonry in the west gable dates to the later medieval period, but cites no evidence (Herity 1990a, 76). White Marshall and Rourke believe that the west gable is somewhat earlier, which they base on the results of radiocarbon analysis of the charcoal in the mortar from the exterior of the gable, which was dated to cal. AD 690–1210 (UCLA-2792) (Berger 1995, 166; White Marshall and Rourke 2000, 74, 121). This, they conclude, indicates that the last construction phase took place no later than the beginning of the 13th century. The radiocarbon sample was in fact taken from the internal face of the west wall, just below the level of the lintel (Berger 1995, 166–7). The masonry here belonged to the first construction phase and therefore it cannot be used to date the latest phase in the upper part of the west gable. Another sample, taken from the west gable above the lintel, yielded a date of cal. AD 240–1020 (UCLA-2570) (Berger 1995, 167). These very wide date ranges and the questionable reliability of the sampling render these dates of limited value in attempting to date the construction of the church. The east wall of the church is wider (0.82–1m) than the others (Fig. 10.6), leading White Marshall and Rourke (2000, 82) to suggest that it belonged to an earlier structure. No evidence was uncovered during the excavation, however, to suggest the existence of an earlier stone church. The west wall is also wider (0.8–0.9m) than the side walls (0.7–0.75m). The use of large thin slabs on edge in church walls is commonly referred to as cyclopean masonry. On High Island, the use of large slender slabs in three of the internal corners of the church may be related to this masonry type (Fig. 5.5 (b), (d)). Manning notes that this was a sophisticated rather than a primitive style of masonry, requiring a strong mortared core for stability, and may be largely a phenomenon of the later 10th and 11th centuries and other writers, such as Ó Carragáin, broadly

agree (Manning 1995a, 16; Ó Carragáin 2005, 100).There is no evidence to suggest that the primary masonry of the west wall of the church on High Island is not contemporary with the side walls. The presence of the doorway, an area of obvious structural weakness, is the most likely explanation as to why the west wall is of greater thickness than the side walls. The same argument can be made in relation to the east wall, which needed to be of sufficient thickness to house the aumbry and to accommodate the recesses and the incorporation of headstones in its external façade (Fig. 5.5 (b), no. 10, Fig. 5.8 (e)). It has been suggested that walls of the same construction phase were commonly built to the same width (White Marshall and Rourke 2000, 83), however this is not always the case. An example is the drystone oratory on Church Island, where the wall widths vary considerably (O’Kelly 1958, 63). Here, the east wall was built on rock, whereas the wider west wall had insecure foundations as there were existing graves beneath that had been cut into a layer of wind-blown sand. It seems that walls can be wider for structural reasons because either they house features that require greater thicknesses or they were built on foundations that were not secure. From its earliest phase, the church was accessed via a trabeate doorway in the west wall, which was slightly offcentre to the north (Fig. 10.6; Pl. 5.4). A restricted site was postulated as the reason for the off-centre doorway in the drystone church on Inishvickallane, Co. Kerry (Ó Carragáin et al. 2005, 35–6). The interior of the church on High Island was lit by a single window in the east gable above the altar. Nineteenth-century antiquarian accounts record this small, round-headed window (Fig. 3.1) and its head and sill were found in the excavations (Pls 5.16, 5.22). The exact form of the window embrasure is not known, but its base was possibly stepped. This was a common type in pre-Romanesque churches and an almost identical example occurs in St Molua’s Oratory, Killaloe, Co. Clare, which had originally stood on Friar’s Island (White Marshall and Rourke 2000, 79–80). It is unlikely that the High Island window was glazed, as there is no evidence for the use of glass in churches of this date in Ireland. There is, however, evidence for the use of external shutters on the windows of many late 11th-/12th-century churches and this may have been the case at High Island. The structure of the roof on the High Island church was most probably of timber, which would have been covered with sod or thatch. The side walls were not of sufficient width to support the lateral thrust of a stone roof. The interior of the High Island church was plastered as traces were evident on all of the walls, with the bestpreserved sections adhering to the east wall behind the altar. This is a rare survival in pre-Romanesque churches. Some examples survive from 12th-century Ireland, such as

308 high island: excavation of an early medieval monastery

St Peter’s church, Waterford, and St Michael’s church on Skellig Michael (Murtagh 1997, 239; Rourke 2009, 134). A small area of stone paving survived intact just inside the doorway and it is possible that the entire interior was originally paved (Figs 10.5, 10.6; Pl. 5.1). The only surviving furnishing was the altar (Figs 10.6, 10.9; Pl. 5.5). Altars are rare survivors in churches of the early medieval period and Ó Carragáin has estimated that there are just ten masonry altars in pre-Romanesque churches in Ireland, of which High Island is the only excavated one (Ó Carragáin 2010, 186). The aumbry in the north-east corner of the church on High Island is unusual as there is an additional concealed cavity below it, within the east wall (Figs 10.6, 5.10 (b); Pl. 5.5). Also, the upper level of the aumbry widens and steps back beyond the line of the north wall, and its interior displayed a slight corbelling (Fig. 5.10; Pl. 5.5). One of the arguments put forward by White Marshall and Rourke (2000, 80) to support their view that the east wall pre-dates the north wall is that the north wall conceals c. 6cm of the opening of the aumbry. It is suggested here, however, that both walls are contemporary and that the aumbry was deliberately built in this manner. It is possible that the opening of the aumbry was concealed from view by a single stone, which could be removed and replaced at will. This stone would have had to be slightly longer than the size of the opening, so that it could be placed in position to abut the stone concealed behind the north wall. This would have created the appearance of a tie stone extending from one wall into the other, thus obscuring the aumbry. However, if, as the evidence suggests, the wall was plastered, then such a stone would have been highly visible, thus negating its concealment function. The purpose of an aumbry was to store precious liturgical vessels, such as the chalice and paten, as well as items such as candles, candlesticks and liturgical books. The wine may also have been kept in this recess. As the High Island aumbry had a lower, concealed cavity, it is likely that valuable items were hidden for safekeeping in this section, although retrieving them from such a confined space would have been awkward. Ó Carragáin has noted that aumbries occur in only six pre-Romanesque churches, all of them small churches, mostly in the west of Ireland (Ó Carragáin 2010, 191). All of these aumbries are small. He draws an interesting correlation between aumbries and stone altars by pointing out that five of the churches with aumbries also have stone altars, a striking coincidence considering the rarity of both features (Ó Carragáin 2010, 191, note 195). In conclusion, it can be stated that the presence of the aumbry on High Island is not out of context in churches of similar date located along the Atlantic seaboard. What is unusual about the High Island aumbry is its complexity of construction, which could perhaps be explained if it had

Pl. 10.3 The carved stone in the south wall of the church, Inchagoill, Lough Corrib, Co. Galway (G. Scally).

functioned as a repository where venerated bones may have been held. The external façade of the east wall of the church incorporated within its foundations the headstones of the Phase 1 Graves 3 and 5 (Figs 5.8 (e), 10.3 (d)). Centrally placed between these, the largest and most elaborately decorated headstone was set in position above Grave 4 when the church wall was built. Two niches were built into the façade, each to accommodate a headstone. It would seem that aesthetics played an important part in the design of the east façade. The effect created by the well-balanced niches, the prominent central placement of the most highly decorated headstone, the well-proportioned layout and appearance of the graves beneath all indicate that these features were intended to have a strong visual impact. It would appear that this work was part of a substantial reorganisation of the ecclesiastical focus of the monastery, carried out to a formalised plan. Elements of design in pre-Romanesque ecclesiastical architecture were not unusual. Formal planning has been demonstrated in relation to the leachta, enclosures and burials on Inishmurray (O’Sullivan and Ó Carragáin 2008, 263, 297, 313). It is suggested that, from the outset, the desire to incorporate the grave-markers into the wall dictated the design and alignment of the High Island church. The incorporation

discussion and conclusions 309

of headstones into a church in this manner is without known parallel. A decorated cross-slab almost identical in style to that placed above Grave 4 on High Island was incorporated into the fabric of the south wall of the nave of St Patrick’s church Inchagoill, Lough Corrib (Fig. 10.3 (d); Pl. 10.3). Further afield, but with known Irish associations, a cross-inscribed stone was incorporated into the south wall of the chancel of the small 12th-century church on Inchmarnock (Scotland) and other carved stones found loose inside the building may likewise have been built into the structure (Lowe 2008, 87). However, none of these examples compares closely with the elaborate features of the east façade of the High Island church. High Island’s association with pilgrimage and the cult of relics may have been the impetus for building the church to incorporate the headstones of the special graves. At the same time as the church was constructed, the enclosure surrounding it received an impressive and carefully laid paved surface, which facilitated its use as an ambulatory (Fig. 10.6). The graves Eleven graves were revealed which relate to the occupation of the monastery between the mid-11th to late 12th/ early 13th century. Eight were located outside the east wall of the church, two (one above the other) outside the north-east entrance to the church enclosure and one inside the church (Fig. 10.6). Burials had been interred in just six of the graves. The location of a burial ground for the ordinary members of the community on High Island was not found, suggesting perhaps that the extensive burial ground at the mother-house at Omey (O’Keeffe 1994b, 15–16) served both communities. The graves outside the church must represent the burials of only the most venerated and were constructed in tandem with and shortly after the construction of the east wall of the church, as part of a larger reorganisation of this focal area of the monastery. When the church was built, the earlier graves (3, 4, 5 and 8) were restructured and one (Grave 3) was realigned along the same orientation as the church. The pseudo graves (2, 6 and 7), which did not contain burials, were added at this time. At the southern end of this row of graves, an individual was subsequently buried in Grave 1 (Fig. 10.6). Clearly the preservation and the enhancement of the earlier graves were paramount, involving the investment of a significant amount of thought, time, physical effort and resources. It would seem that this work was carried out as part of a cult of relics and as a focus for pilgrimage (Ó Carragáin 2010, 57–85). The main work involved in restructuring was to raise the upper surface of the pre-church graves. This involved

the addition of sidestones, which in the cases of Graves 3 and 4 were set above the earlier sidestones resulting in two-tiered graves (Fig. 10.3 (a)–(c); Pls 5.16, 5.22). Tiered construction of other early medieval graves has not been recorded. After the upper (Phase 2) sidestones had been put in place, layers of fill were placed in the graves and Graves 3 and 5 were covered either by a decorated or a plain recumbent slab (Fig. 10.6; Pls 5.16, 5.22).The fill in Grave 4 did not extend to the top of the sidestones, suggesting that originally a recumbent slab may have been laid over the surface of the grave but was removed from the grave shortly before the church gable collapsed in the 19th century. There was no evidence to indicate that the headstone contemporary with the pre-church phase of Grave 5 was altered in any way during the restructuring phase (Fig. 10.3 (a)–(d)). In order to construct the pseudo Graves 2, 6, 7 and the restructured Grave 8 to a comparable level, deposits of soil and stones were laid down. That these graves were deliberately laid out during this phase is indicated by the fact that Grave 6 would have been marked by a headstone (now missing) set within the niche built into the church wall directly above it (Fig. 5.8 (e) Q). Another niche was located in the church wall above Grave 2 in which the headstone was still in place (Fig. 5.8 (e) C). Grave 2 was covered by a decorated slab and Grave 6 by an undecorated slab (Fig. 10.7; Pls 5.16, 5.22). It is possible that the remaining graves (7 and 8) had also been covered. Footstones were either moved from their Phase 1 positions or were added during the restructuring that was completed by kerbing set at both the northern and southern ends of the graves in line with the church walls. It is conjectured that these were symbolic graves, perhaps representing people whose remains were lost at sea. In instances such as these, a grave may have been made in absentia and was marked by the traditional symbols of decorated grave-markers, thereby creating an enduring focus for prayer and veneration of the individual. Parallels for this practice have been found in the early medieval cemetery on Ardwall Isle in Scotland (Thomas 1967, 135, 152). The possibility that a body had been interred but was removed or translated can be ruled out. If this did occur, however, it would be highly unlikely that the skeletal evidence would have been removed in its entirety (E. O’Brien, pers. comm.). Graves of holy men increased the sanctity and prestige of a monastery and they were viewed as a means of attracting pilgrims and increasing revenue through patronage. These were probably the main reasons for restructuring, enhancing and enlarging this venerated cemetery on High Island. The construction of the church and the restructuring and enhancement of this special cemetery was undertaken before the individual in Grave 1 was buried in

310 high island: excavation of an early medieval monastery

cal. AD 1033–1169 (UB-4156). This elderly male was buried with his knees flexed, his legs crossed at the ankles and his right leg weighed down by a stone (Fig. 5.28; Pl. 5.23). His head rested up against the church wall and his feet were touching the footstone. These factors, especially the flexed knees, suggest that the grave was of insufficient length to accommodate this individual. It would appear that the option of lengthening the grave and moving the footstone further east was not availed of, perhaps indicating that the regular appearance of the line of graves took precedence. Eighteen cross-slabs were associated with the graves outside the east end of the church, fifteen of which were decorated (Figs 5.22, 10.7; Pls 5.16, 5.22). It is probable that another six grave-markers were originally associated with these graves, giving a total of at least 22 grave-markers in this cemetery. It is noteworthy that the decoration on some of the stones was obscured from view by the grave fills. The headstones of Graves 3 and 5 and the headstone in the niche above pseudo Grave 2 were decorated on both broad faces, and the concealed faces of these stones were not visible once the east wall of the church had been built around them (Fig. 10.3 (d)). Many of the High Island cross-slabs and crosses appeared to have been reused and some must originally have been free-standing and visible from both sides. They exhibit decorative motifs comparable to the recumbent slabs at Clonmacnoise (Maddern, this volume, p.176). The majority of the cross-slabs and crosses were of mica-schist, which was sourced on High Island or its vicinity. The footstone at Grave 3 was, like that of the recumbent slab on the same grave, of imported limestone. These were the only non-local cross-slabs identified from the entire corpus of slabs on High Island and they were imported to mark a venerated individual’s grave. Both displayed a distinctive fossil imprint, which suggests a provenance on the Aran Islands, some 70km to the south (White Marshall and Rourke 2000, 151, note 174). These imported stones may demonstrate élite patronage. On St Mac Dara’s Island, the local stone is a hard, metamorphic granite that is not suitable for carving. The crosses on this island, with the exception of one, are of ‘blue limestone’ (De Breffny and Mott 1976, 12), possibly also sourced on the Aran Islands, which are in close proximity. The isolated burial in an earth-cut grave (Grave 11), uncovered outside the north-east entrance to the church enclosure, was probably associated with pilgrimage (Fig. 10.6; Pl. 5.28). The remains of the middle-aged to older male were dated to cal. AD 1021–1159 (OxA-13665). Burials at entryways, thresholds or boundaries could signify high status and there was a belief that interments in these 1 Thanks to E. O’Brien for bringing this reference to my attention.

locations might ward off enemies or offer some kind of protection (Fry 1999, 170–71). The small hearth overlying the fill of Grave 11 may represent the remains of a funerary meal that took place shortly after burial. Excavated evidence for such rituals is rare, but a somewhat similar relationship between a possible 13th-century hearth and a burial was revealed in the cemetery at Ballyshannon, Co. Donegal (Ó Donnchadha 2007, 8). Thomas suggests that such rituals derive from Mediterranean customs and evidence for such has been uncovered at Eny’s Brachan, the Isle of Lundy in the Bristol Channel and other locations (Thomas 1994, 205). A stone-lined grave that contained twenty quartz pebbles but no human remains was later built above Grave 11 (Fig. 10.7; Pl. 5.29). The date of the restructured grave is not known, but it cut through the leg bones of the burial below, suggesting that the exact position of the earlier grave had been lost from memory by the time of its construction. Parallels for pseudo graves built on top of earlier burials are rare. One example is a burial at Cannington Cemetery, Somerset (burial FT26), in a slab-marked grave mound excavated in the 1960s (Rahtz et al. 2000, 54, 413). In this case, there seems no doubt that the slabs were set in the surface of the mound to mark the grave below. The remains were those of a juvenile of approximately thirteen years, dating to some time between the 7th and early 11th century. It was suggested that this grave provided positive evidence for marking of a special grave in connection with pilgrimage.1 The burial of the young male inside the church on High Island took place in the period cal. AD 1163–1230, at a time when there is little archaeological evidence for burial within churches in early medieval Ireland. However, the situation had changed by the later medieval period when burial within churches was commonplace (Ó Carragáin 2010, 302). White Marshall and Rourke (2000, 93) concluded that this interment must have taken place after the monks were no longer in full-time residence on the island and they argue that the monks were unlikely to sanction a burial practice they avoided for themselves. However, practices may already have been changing at this time, as Toirdelbach and Ruaidrí Ua Conchobair were buried within a functioning cathedral at Clonmacnoise in the 12th century (Manning 1998a, 78). Grave goods, comprising a hone and a line-sinker or plumb line, were recovered from the High Island burial. The artefacts had been placed one on each shoulder in a clearly symbolic gesture. Grave goods are highly unusual in Christian burials, but they are frequently found in burials of Scandinavian tradition in Britain and Ireland from the mid-9th to mid-10th century. Soapstone pots and whetstones are common burial equip-

discussion and conclusions 311

ment on Viking sites in parts of Norway (Resi 1987, 95–6). The artefacts usually consist of a pair of objects, comprising a set of one light and one dark schist whetstone or hone, each from a different geological region. Whilst the goods accompanying the male buried in the church on High Island are not an exact parallel, similarities do exist in that one was a hone and the two items were of different geological provenance, one local and the other imported. Isotope analysis was carried out in an attempt to identify the man’s cultural background. Whilst the results of the analysis have ruled out a Scandinavian origin, they indicate that he came from a place with similar geology within Ireland or Britain, but it cannot be confirmed if this was from a HibernoScandinavian community. It is known that the culture of the settlement and its inhabitants on Beginish remained Scandinavian or Hiberno-Scandinavian throughout the period of occupation, including into the 11th and 12th centuries (Ó Corráin 2009, 147; Sheehan et al. 2001, 93). Similarly, the community on High Island may have had earlier Hiberno-Scandinavian links in the mid-11th century as also indicated by the finding of the silver penny in Cell B. Quartz, which is in plentiful supply on High Island, was recovered from graves of all periods. Even though there is no specific reference to white quartz in the early literary sources, it is suggested that its significance related to its colour because to Christians white has traditionally symbolised purity, innocence and a holy life (White Marshall and Rourke 2000, 111). In the medieval period, the light-emitting properties of quartz may have had added connotations of purity and salvation through the cleansing fires of purgatory, symbolising rebirth through the processes of baptism, death and resurrection (Gilchrist 2008, 139, 151). Gilchrist notes that transparent stones, such as quartz, developed associations with the Apocalypse, making them particularly relevant as grave goods for the Christian dead. The occurrence of quartz pebbles in mortuary contexts is known from the prehistoric to the later medieval period and was concentrated in the Isle of Man, Scotland and Ireland (Gilchrist 2008, 139). Inclusion of quartz pebbles is particularly common in funerary settings along the Irish Atlantic seaboard. They were found on and in some of the graves on Omey Island (O’Keeffe 1992a, 2) and also on Inishmurray, especially in the large leacht in the cemetery enclosure (O’Sullivan and Ó Carragáin 2008, 275). They were found scattered in and on a number of the Illaunloughan structures (White Marshall and Walsh 2005, 87). A vast number of quartz pebbles (6,800) were also found on Church Island (O’Kelly 1958, 94) where, in the light of more recent excavations, evidence was found for their abundant use in 7th- and 8th-century burial contexts and around the tomb shrine of that date, as well as with some

of the 10th-/early 11th-century burials around the large stone church (Hayden 2013). Small fragments of quartz crystal were recovered from approximately two-thirds of the 1,275 burials of possibly 13th-century date from the cemetery at Ballyshannon (Ó Donnchadha 2007, 10). In Scotland, significant concentrations of quartz pebbles were found at Whithorn in contexts dating from the second half of the 7th or early 8th century to the late 9th century (Hill 1997, 473). Quartz pebbles occurred in graves pre-dating St Ronan’s church on Iona (O’Sullivan 1994, 334).

10.5.2 The monastic enclosure The cells Date and function Descriptions of the monastery in the 19th century combined with the excavation findings indicate that there are five, or possibly six, beehive huts located within the monastic enclosure. In addition to Cell A and Cell B, the remains of three others are preserved as mounds of collapsed rubble on the western side and it is likely that these represent the accommodation huts for the monks (Fig. 5.1). The outline of the rubble suggests that they were small, probably single occupancy buildings of comparable size to Cell A. Their location in the western area of the monastery corresponds to the evidence from sites such as Reask, Illaunloughan and Inishmurray, where it has been suggested that there was a spatial separation between the ecclesiastical and the domestic buildings, with the latter located in the western part of the monastery (O’Sullivan et al. 2008, 124). It is suggested that the construction of Cell B was carried out as part of the initial period of expansion, when the church was constructed, in the mid- to late 11th century (Fig. 10.6; Pls 5.31, 5.33). Stratigraphically, it was demonstrated that Cell A was constructed at the same time as the monastic enclosure wall was extended (Fig. 10.7; Pl. 5.52). Burnt debris (F4079) sealed by this wall was dated to cal. AD 1042–1217 (UB-6454), which provides a terminus post quem. Another date of cal. AD 1176–1276 (UB-6453) was obtained from a sub-floor deposit within Cell A, indicating that it was built in the late 12th century at the earliest. It is postulated that Cell B was a communal building where food was prepared and probably consumed. This is suggested by the presence of a substantial hearth and the recovery of food waste from within. The main source of fresh water was directly in front of this cell, which was also easily accessible from all areas within the monastery. Cell A, on the other hand, was considerably smaller and was likely to have accommodated no more than one person and because of its location, at some distance from those in the western part of the enclosure, it was likely to have accommodated the monk in charge.

312 high island: excavation of an early medieval monastery

Methods of construction The excavated beehive huts exhibited varying levels of sophistication, with Cell B clearly more skilfully built than Cell A. The masonry of Cell B was extremely well crafted in contrast with that of Cell A, which was of inferior quality. It was noted during conservation works that long ‘through’ stones were used in the corbelling of Cell B. Such stones, which provided structural stability to the corbelling, are probably responsible for the survival of the roof of Cell B (J. McDonagh, pers. comm.). Externally, Cell B was roughly rectangular with rounded corners and internally it was square in plan (Fig. 5.37; Pl. 5.31). Cell A was smaller, more rounded in shape and rectangular in plan internally (Fig. 10.7; Pl. 5.52). According to Leask, the normal clochán was circular, or nearly so, on plan, and it developed towards the oblong with rounded corners and a rectangular internal plan (Leask 1955, 18–19). The evolution from round to rectangular plans for domestic structures of the early medieval period is generally accepted (Lynn 1978). The plans of the cells at High Island, where both displayed short rectangular or nearly square interiors and were relatively late in date, fit neatly into this scenario. Beehive huts with plans similar to Cell B were common along the western seaboard, for example on Skellig Michael and in the case of the rectangular house on Church Island (O’Sullivan and Sheehan 1996, 284–5; O’Kelly 1958, 73, 128). The church enclosure wall was restructured primarily to facilitate the bonding of its masonry with that of Cell B, constructed on the outside (Fig. 10.6). The entrance at the north-east of the church enclosure was redesigned with a curved passage permitting convenient access from the church enclosure to Cell B. There were unusual features in the construction of Cell B where the foundations were formed by large boulders on the southern side and by a stone plinth on the northern side (Fig. 10.6; Pls 5.31, 5.32). These wide foundations were built with a specific purpose in mind. The lower parts of the walls and the foundations were covered by redeposited boulder clay, which was most likely covered with layers of sod to form a weatherproof cladding. It would appear that the foundations acted as a key for the clay, which was probably applied shortly after the cell was constructed. If this was the case, Cell B can be compared with the cells at Reask and with the circular stone house on Church Island, which appeared to have been deliberately built in this manner in order to provide ‘a good key’ to help to secure the turves in place (Fanning 1981, 90; O’Kelly 1958, 71). The north and east sides of Cell A had an additional low wall-face that retained small stones and clay in a similar manner to features known as annuli found at Reask and Church Island. The annulus of Cell A was an extension of the north face of the enclosure wall. Annuli provided

a layer of extra protection against wind and rain and also strengthened the base of the cells in a manner similar to the feature described above for Cell B. Water management Water supply was controlled by the extensive collection and drainage system that was constructed in the eastern side of the monastery (Figs 10.6, 10.7). The system appears to have originated from an initial requirement to remove excess water from the Phase 1 rock-cut well. The water-management system developed in two stages during Phase 2. The first stage, contemporary with the construction of Cell B and possibly relating to its function as the refectory and kitchen, saw the construction of the rock-cut trough and the soakaway in the mid- to late 11th century (Fig. 10.6). The system was extended when Cell A was constructed and when the north side of the monastic enclosure wall was rebuilt, suggesting a mid-/ late 12th-century date. The most significant alteration was the construction of the channel that connected the newly developed north-eastern part of the monastery with the rock-cut trough to the south (Figs 10.7, 10.9; Pl. 5.40). At the junction of the newly built channel and the trough, the earlier stone-lined channel was roofed with lintels so that it formed a covered drain at the head of the trough (Figs 10.7, 10.9). This channel would have considerably enhanced the existing water system and after its construction, significant quantities of water could be collected. The work involved to create this system was extensive and it was probably carried out to provide more water by more efficient means to the point of collection in the trough. Sophisticated water-collection and/or drainage channels are well attested at other early medieval sites, most notably Skellig Michael, Church Island, Reask and Killelton, all in Co. Kerry (Rourke 2009, 134; O’Kelly 1958, 74; Fanning 1981, 104; Manning 1988). The monastic enclosure wall The northern side of the monastic enclosure wall underwent significant alterations during this period. It was rebuilt and widened on the inside after Cell A was built, making it a much more substantial structure than its predecessor (Figs 10.5, 10.7; Pl. 5.50). On the inner side of the wall there is a double facing, one sitting directly on the primary foundations and the other constructed to the south. A date obtained from charcoal extracted from a deposit of burnt debris sealed by the latter wall-face provides a terminus post quem of the mid-11th to the early 13th century for the alterations to the monastic enclosure. At the same time, a stepped passage was created over this stretch of the wall in the north-west corner (Figs 10.5, 10.7; Pl. 5.51). This meant that during the latest phase of activity associated with the monastery, the enclosure wall had four entrances, one in each corner (Fig. 5.1).

discussion and conclusions 313

West Z

East Z1

8 7

1 West wall of church

13 Kerb for trough (F845)

2 Paved surface (F29)

14 Channel (F8052)

3 The 'paved area' (F245)

15 Leacht (F887)

4 Church

16 Lintelled drain (F8007)

5 Altar (F209)

17 Monastic enclosure wall

6 Headstone (CS 9), Grave 5

Structures

7 Surface (F54), Grave 6

Paving

8 Footstone (CS 12), Grave 6

Lintels

9 East wall of church enclosure

Upright stones

10 Cell B

Conjectural

11 Paved floor (F326), Cell B 12 Paving (F835)

Fig. 10.9 West–east section through the excavated area of the monastery. Location of section Z–Z1 on Fig. 10.7.

10.5.3 Lifestyle and economy

The impression of the lifestyle and economy during the most active phase of the monastery is one of a simple, frugal and, in many respects, self-sufficient community. The assemblage of small finds was dominated by stone artefacts, many of which could have been used for a variety of purposes and their types would have had a long period of usage. The most numerous artefact was the hone or whetstone, suggesting that a significant number of iron tools must have been in use, although the soil conditions militated against their survival. Such tools would have been necessary for quarrying the stone for building as well as for carving the cross-slabs and crosses. Iron tools were also required when working wood, bone and leather. Some of the flint or chert scrapers may have been for leatherworking. The rubbing stones were probably used when preparing hides as well as for smoothing objects manufactured from bone, wood, soft stone and metal. Sheep/goat were prevalent amongst the domesticated animals in the faunal assemblage, followed by cattle. Local environmental conditions favoured the rearing of sheep, which not only would have provided a source of meat, but also milk, wool and woolfells. The evidence from the butchery marks suggests that animals were slaughtered on the island. Fish also formed part of the diet and the most frequent species were whiting, ballan wrasse, scad, plaice, cod and possibly dogfish. As on other monastic coastal sites, shellfish, predominantly limpet, was used to supplement the diet. Seabirds were killed, probably whilst

nesting on the island. Storm petrels were dominant, but whether they represent food remains in all cases could not be determined. However, shag bones displayed evidence of having been roasted and a cormorant bone had knife marks consistent with meat removal. Barley was the predominant cereal and oat was also present, but it could not be established whether this occurred as a weed in the barley crop or whether it was being cultivated in its own right. There was a notable lack of evidence for wheat or rye. In early medieval Ireland, barley and oat were ranked as low status cereals (Kelly 1998, 226–7). The presence of barley is consistent with the findings from other early medieval monastic sites in Ireland, such as Skellig Michael, where both barley and oats were recorded (Allen 2011, 393). The island had a horizontal mill for grinding grain and its presence indicates a substantial investment. The mill was capable of meeting the needs of up to 100 individuals, though it is likely that the island could only have sustained considerably less than this, perhaps 50–70 people (Rynne 2000, 212). It is possible the mill on High Island served both the Omey and the High Island communities. Rynne has suggested that grain could have been grown on High Island itself and he speculates that this was possible in the area adjacent to the lake (Rynne 2000, 206). Pollen analysis from the island supports this view, indicating a date of about AD 650 for a surge in arable activity, including an increase in rye pollen (Molloy et al. 2000, 236). The pollen record from Church Lough, on nearby Inishbofin, has confirmed that

314 high island: excavation of an early medieval monastery

at the time the monastery was founded there in AD 665 there was a dramatic change in the vegetation cover, with a decline in birch and hazel giving way to grass, plantain and cereal, mainly wheat (O’Connell 1994, 56). This site provided evidence of a farming economy in Connemara, where cereal production played an important role. Gardens were probably an integral part of early medieval life. Many of the artificially created terraces on Skellig Michael have been tentatively identified as garden terraces used for the cultivation of various food crops (Horn et al. 1990; Bourke et al. 2011). Communities needed to grow vegetables and they were also likely to have kept herb and medicinal gardens. The fragments of quern stones and possibly the rubbing stone found on-site could have been used for the processing of cereals and other foods. Peat, or turf, was the most commonly used fuel in the monastery followed by wood and heather. Much of the wood consisted of narrow roundwoods and in view of the consistency in supply and the frequency of use, it is most likely that firewood was supplied from sources within the island, which were probably managed woodland growing in sheltered valleys. This may have been supplemented by other means, such as the collection of driftwood. Seven different tree species were identified: hazel, oak, birch, ash, hawthorn, willow/poplar and yew, all of which could have been native to the island. This range of species is similar to that identified on Skellig Michael, where the wood must have been sourced on the mainland (O’Donnell 2011, 399). There is little evidence for long-distance trade and external contacts. The most notable is the Scandinavian or Hiberno-Scandinavian silver penny dating to the second half of the 11th or early 12th century. The economy of High Island appears to have been more frugal and materially poorer than other Atlantic island monasteries. It resembles the evidence uncovered on Church Island, but on Church Island the finds assemblage was more diverse (O’Kelly 1958; Hayden 2013). The lifestyle on High Island seems to have been very basic, lacking luxury items, in contrast to Illaunloughan, where considerable quantities of personal decorative objects of copper alloy, bone and antler were found, as well as pottery and fragments of clay moulds (White Marshall and Walsh 2005).

10.5.4 The monastery and pilgrimage

The excavations have thrown some light on the role that pilgrimage played in the development of the monastery on High Island, and it is significantly greater than had previously been thought. Pilgrimage played an important role in its physical and spiritual development, particularly during the period from the mid-11th to the late 12th/early 13th century. High Island was probably part of an established pilgrim route that included other island and mainland

sites in north-west Connacht at this time, when pilgrimage appears to have been on the increase (O’Sullivan and Ó Carragáin 2008, 257–8). Penitential stations The term leacht is interchangeable with that of penitential station and indicates a low, open air, rectangular or square drystone structure where the ritual of the pilgrimage round (an turas) was probably performed. The OS Fair Plan of 1839 marks the positions of five stations on the island, which were located near the south and south-east landings, between Brian Boru’s well and the south landing, on the north-eastern shoreline of the lake and another at the south-east entrance to the monastery (Fig. 3.2). The antiquarian reports of the 19th century also noted the presence of these stations. These were striking features, sufficiently visible and noteworthy to be commented upon. Over a century later, Herity’s account of the island’s antiquities describes eight stations, which included the five already mentioned (Fig. 3.7; Herity 1990a, 83–5). The first of the additional stations is located in the monastic enclosure just outside the north-west entrance to the church enclosure (Fig. 3.6). The second station is a relatively substantial circular structure located north of the small lake (Fig. 3.7). Herity compares the form of this structure to the enclosure at Station 12 at Gleanncholmchille turas, Co. Donegal. The third additional station is Brian Boru’s well, located on high ground close to the centre of the island (Figs 3.7, 4.7; Pl. 4.11). This was the first time the well had been mentioned in the context of a station and Herity remarks that it would be a highly unusual omission from an early medieval turas, especially in view of its unusually elevated situation. He further suggests that the recovery of the head of a small decorated cross in the vicinity of the well substantiates the view that this was a station. Herity outlines a probable route taken by pilgrims on a turas performed deiseal (i.e. moving clockwise from one station to the next) (Herity 1990a, 97). Excavation revealed evidence for a further two leachta in the east and south-east of the site which were contemporary with the later occupation (Fig. 10.7; Pls 5.38, 5.39). Both of these leachta were unusual in that they abutted other features, preventing pilgrims from walking around them. It is suggested that they fulfilled a different function and that they were used as locations for prayer associated with the occupants of the monastery, as distinct from stations that were part of a turas. Significantly, a decorated cross was found in rubble overlying the leacht outside the entrance to Cell B (Fig. 10.7). It is likely that it had stood upright on the leacht in the same manner as the crosses remarked upon by Wakeman in relation to the other stations on the island. The discovery of these additional

discussion and conclusions 315

leachta within the monastic enclosure shows that there were at least ten penitential stations on the island. Leachta were an important component of other wellestablished pilgrimage destinations. On Omey Island, though only two leachta were located during the excavations, the relationship of these to earlier burials led the excavator to suggest that pilgrimage was the reason for the continuity of the structures over time, probably lasting several centuries (O’Keeffe 1992b; 1994b, 17). Eleven leachta (or sites thereof) were identified around the perimeter of Inishmurray, with a further four inside the monastic enclosure (O’Sullivan and Ó Carragáin 2008, 7). Seventeen pilgrimage stations have been recorded on Caher Island, where a pilgrimage still takes place each year on 15 August, the feast of the Assumption, and the suggested route taken by early medieval pilgrims has been set out by Herity (1995b, 106–7, 110–22). On St Mac Dara’s Island, stone altars were located in the vicinity of the church and all were surmounted by crosses or fragments of crosses (Westropp 1905, 57). This island is famed as a site of pilgrimage and to this day it is visited by a substantial number of people on the saint’s feast day, 16 July. Fifteen stations have been identified on Turas Columb Cille at Gleanncholmcille, where many are marked by early medieval cross-slabs or cross pillar stones (Herity 1993a, 9). On High Island, there is no tradition of modern pilgrimage. Accessibility Certain aspects of the layout of the monastery on High Island indicate that access was actively controlled. The main entrance at the south-east of the monastic enclosure was flanked on each side by buildings, one probably a guesthouse and the other a possible porter’s lodge (Fig. 5.1). By the mid-/late 12th century the enclosure had an entrance in each corner. The provision of four entrances within a wall of relatively short circumference (approximately 100m) at a small monastic site was of significance. The entrances were surely designed to allow access from all directions, regardless of which landing place was used by visitors and pilgrims. On Inishmurray, where the enclosure wall was approximately 50% greater in length than on High Island, there were four entrances. Inishmurray has been regarded as a monastic establishment first and a pilgrimage destination second (O’Sullivan and Ó Carragáin 2008, 346). The number of entrances could signify the importance of pilgrimage as other monasteries appear not to have been provided with additional means of access. On Church Island and on Illauntannig, where the enclosure walls were longer, just one entrance was extant, though both walls were not well preserved. At Kildreenagh/Loher, there was just one entrance in the enclosure wall that compared closely in length to that on High Island.

The cult of relics and architectural innovation The formative period in the development of corporeal relic cults in Ireland possibly dates to the 8th or 9th century, and it went though a revitalisation in the late 11th and 12th centuries (Ó Carragáin 2010, 85). Ó Carragáin’s belief that the cult of relics was a driving force for architectural change during its revitalisation appears to be particularly relevant in relation to the church on High Island. It is also relevant that pilgrimage from Ireland to places further afield, such as Santiago, Rome and other destinations in Britain and Europe, was on the increase during the first half of the 12th century, which led to innovation in architecture, in particular the introduction of Romanesque, which was common along pilgrim routes (Ó Riain-Raedel 2001, 25–6). This is of significance in the context of the development of the architecture on High Island in that the church has unusual elements that are unparalleled in other contemporary churches in Ireland. These features can be best explained by the interlinked traditions of corporeal relic cults and pilgrimage, both of which were experiencing resurgence at the time the church was built. Although the monastery on High Island was part of the cult of St Féichín from its foundation, the saint was not buried there. Nevertheless, the island may have been attractive as a place of pilgrimage from its earliest period. Remote hermitages of saints were important destinations for pilgrims at sites such as Inishmurray (Ó Carragáin 2010, 273). The architectural and archaeological evidence show that the prestige of the monastery increased from the mid-11th century. The impetus for this was most likely associated with Gormgal, whose death occurred in 1018 and whose remains are likely to lie either in Grave 3 or Grave 4 at the eastern end of the church (Figs 10.2, 10.3a; Pls 5.13, 5.15). Both graves were exceptionally well made, which set them apart from others. The headstone and footstone of imported limestone, both carved with elaborate crosses, marked Grave 3. It is possible that the inclusion of a piece of porphyry with the individual in Grave 4 indicates that he had been on pilgrimage and that this is a pilgrim souvenir, rather than a relic of trade (E. O’Brien, pers. comm.). However, the possibility that it resulted from Hiberno-Scandinavian contacts cannot be discounted. The interpretation of porphyry sherds in post-Roman contexts is that they were for use in portable altars or that they represent pilgrims’ keepsakes, and in both cases pilgrimage is invoked as the chief mechanism for their dissemination (Lynn 1984, 27). Gormgal’s fame would not only have enhanced the reputation of the monastery and increased the number of pilgrims travelling to the site, it would also have attracted wealthy patronage. His status as anmchara (‘spiritual advisor’ or ‘confessor’) would have brought penitents to his monastery. High Island would certainly have been an ideal location for austere penitential practice and

316 high island: excavation of an early medieval monastery

the esteemed reputation of Gormgal most likely led to the promotion of his grave as a place of pilgrimage, itself a form of penance. Pilgrimage would also have been a source of revenue. The work invested in the preservation, enhancement and maintenance of the graves outside the church on High Island and the grave outside the entrance to the enclosure suggests that the monks were facilitating access for pilgrims to the burial places of holy men. Even though the east wall of the church partially overlay the special graves, their headstones were incorporated within the wall and enough of the graves was left exposed for visitation and prayer (Figs 10.3 (d), 10.6; Pls 5.16, 5.22). The physical manifestation of the eastern façade, replete with its symbols of early medieval Christian devotion, would probably have reflected elements of contemporary spiritual belief and practice. The incorporation of the headstones reflects the veneration in which these graves were held. Shrine chapels were particularly common in the 8th and 9th centuries and it seems there is both archaeological and documentary evidence for a suggested revival of the practice in the late 11th and 12th centuries (Ó Carragáin 2003b, 134). The treatment of the east wall of the High Island church could belong to this revival. The practice of incorporating spolia, or the reuse of building materials and occasionally funerary monuments in churches, is well attested (Ó Carragáin 2010, 156–67). Ó Carragáin argues that in Ireland, the use of spolia served as continuity with the age of saints and was a tangible physical link that was clearly believed to create and to reinforce the sense that the building was in itself almost a relic. There seems little doubt that there was a symbolic dimension to the retention of the headstones from one building episode to the next. The incorporation in the altar of the church of the upright stones associated with the Phase 1 slots could be seen in the same light.

10.5.5 The decline of the monastery Little is known about the decline of the monastery. The reforming measures of the 12th-century synods saw the transformation of the Irish Church, which led to the gradual demise of native monastic institutions (Ó Riain-Raedal 2001, 25, 27). Churches were reorganised on diocesan and parochial systems. New monastic orders, such as the Augustinians and Cistercians, were introduced from outside the country (Horn et al. 1990, 11; Bourke et al. 2011). The importance and the number of these reformed monasteries increased with the Anglo-Norman conquest of Ireland in the 12th and 13th centuries. Giraldus Cambrensis refers to Skellig Michael in the late 12th century, stating that, by this time, the monks had moved to Ballinskelligs (Bourke 2005, 125). Colder weather and the increasing frequency and severity of sea storms appear to have forced the monks to withdraw to the mainland site, yet they con-

tinued to maintain and use the island monastery (Horn et al. 1990, 11). Skellig Michael lies 14.5km (nine miles) offshore and the community may have felt the impact of the worsening climate. Perhaps this was also the case on High Island where, even though it was considerably closer to the mainland, sea conditions would have played a significant role in its viability as a settlement. It is likely that a combination of factors contributed to the end of monastic life on High Island in the 13th/14th century. It may be that the number in the community reduced slowly over time, so that at some point, life on the island became unsustainable. It is possible that the monks moved initially to the mother-house on Omey Island, which was occupied into the 14th century, in the same manner as the community on Skellig Michael had transferred to the mainland foundation in Ballinskelligs.

10.6 Phase 3: late 12th/early 13th to mid-15th century

The excavation revealed evidence to suggest that some activity continued at the monastery after full-time occupation had ceased. The nature of the activity was difficult to ascertain, and it was unclear whether it related to permanent or more transitory, perhaps seasonal, settlement. The monks on Omey may have retained an interest in their outlying island monastery during this time and pilgrims would probably have continued to come to the island. The main structural activity during this period was the erection of the sub-rectangular building abutting the outside of the northern side of the monastic enclosure (Figs 10.5, 10.10; Pl. 5.54). The fact that this building was constructed on deposits that had accumulated above the in-filled water drainage channel indicated that this latter may have been one of the earliest features to fall into disuse after the monastery was no longer permanently occupied (Fig. 10.10). Without regular maintenance, the channel would have silted-up quickly. The recovery of a silver halfpenny of early 13th-century date from the fill in the channel suggests this may have occurred in the 13th century. Access to the sub-rectangular building was by means of a paved and stepped passage that was cut through the northern side of the monastic enclosure wall (Figs 10.5, 10.10; Pls 5.56, 5.57). Construction of the passage would have necessitated the removal of a substantial stretch of the wall and no evidence was uncovered to indicate that it was ever rebuilt. No contemporary floor surface was identified within the sub-rectangular building, and no evidence was uncovered to suggest its function. This could imply that it was a temporary shelter for visitors to the island. Cell A was the only building known to have been reoccupied on High Island in the high medieval period (Fig. 10.10; Pl. 5.52). The evidence derives from a date of cal. AD

discussion and conclusions 317

1287–1424 (UB-4988) from the substantial deposit of barley grains recovered from a hearth on the paved floor of the cell (Figs 5.63 (b), 5.64). Whilst this evidence could relate to a single incidence of occupancy, the presence of a later level of paving, with ash and charred peat on its surface, indicated a more protracted period of use. Although no evidence was uncovered to date the later deposits uncovered within Cell B, it is likely that it too was reused during this period. Its considerable size and substantial walls would have provided an ideal place of shelter (Fig. 10.10; Pls 5.31, 5.33). The evidence for this later occupation comprises quite substantial deposits of clay and stone within which there were sporadic food remains. Further evidence for activity in the monastic enclosure was recovered in the form of small mounds of burnt debris, probably the remains of hearths (Fig. 10.10). A notable feature of these was that almost all were located outside buildings, but in positions where they would have been sheltered by the monastic buildings, suggesting that the structures were abandoned when they were made. Whilst no dating evidence was obtained for these features, one of the hearths was located immediately above the deposit where the 13th-century coin had been found (Fig. 10.10). This would imply that the hearth had been in use not long after the coin had been lost. The other hearths were located on stony clay deposits that had accumulated above the paved surfaces south of Cell B, suggesting that these too probably date to this period (Fig. 10.10). Reoccupation of monastic sites in the high medieval period was common. What have been termed ‘post-monastic squatters’ were living at the monastery on Church Island in the late 13th or 14th century, after the religious community had abandoned the island (O’Kelly 1958, 77, 115). These people caused considerable disturbance to the monastic buildings in their search for material for their own shelters. One of these (Shelter G) was similar to the sub-rectangular building on High Island in that it too was roughly rectangular and it abutted the enclosure wall. At Reask, one of the buildings (Structure E), of unknown date, also bore similarities to the sub-rectangular structure at High Island in that it was a small, rectangular, drystone building that had been erected as a lean-to against the inner face of the enclosure wall (Fanning 1981, 96). Unlike the High Island building, however, an abundance of occupation material was found in the Reask structure. The church may also have continued in use, albeit irregularly. The pseudo grave was built and its associated marker erected in the north-west corner in connection with continuing pilgrimage to the site (Figs 5.13, 10.10). The diminutive grave was filled with quartz pebbles and crushed shell, but no date for its construction was obtained. Whilst it is possible that it may be earlier in date, the fact

that its side walls extended above the level of the church floor and that its surface had remained intact suggest that most likely it was built after the church was no longer in use. The paved area around the church became covered over by a deposit of clay and stone that, it is suggested, would have taken no more than a decade or two to build up. Three small ironstone nodules were found embedded into the clay at the foot of the graves to the east of the church (Fig. 10.10; Pl. 5.22). These were revered objects, which were left as offerings by pilgrims. The stones were particularly distinctive and two were carved, one with an inscribed cross (Cross-slab 25). The provenance of the ironstone was probably the Burren or the Aran Islands and the closest parallel for the incised nodules is one from Temple Brecan, Inis Mór, Aran Islands, which bears the inscription ‘OR DO BRAN N-AILETHER’ (‘pray for Bran the pilgrim’). Thus, it seems that the monastery, even after it was no longer fully active, continued to attract pilgrims to venerate these graves.

10.7 Phase 4: mid-15th to late 20th century Little is known about the site during the 15th and 16th centuries and no finds were associated with this period. The island provided exceptionally good summer grazing, which would have attracted a small number of inhabitants as indicated by the fish, bird and mammal bones found. Ownership of the island during the late medieval period is unclear. It is probable that it was in the hands of the secular rulers of Iarchonnacht, the O’Flahertys, until they were dispossessed of their lands in the Cromwellian settlement. From the late 17th century, the island was owned by the Martin family of Ballynahinch Castle and in the late 18th century it was leased to a tenant, John Bodkin. In 1837 an entry in a survey of the estates of Thomas Martin reads: ‘Omey and High Islands are reclaimable and at present produce excellent feed for all kinds of stock’ (White Marshall and Rourke 2000, 225, note 229). George Petrie’s account of 1820 indicates that the monastic buildings were then largely undamaged. Other antiquarians who visited the island throughout the 19th century recorded the gradual decline and eventual collapse of many of the monastic buildings. Shortly after Petrie’s visit, the short-lived mining campaign took place. The miners lived on the island and the first real destruction of the monastic buildings can be attributed to them. The graves outside the wall of the church remained largely unencumbered by rubble, and the manner in which the remains of the east window of the church were found would suggest that the graves were maintained until the window’s collapse in the later 19th century.

318 high island: excavation of an early medieval monastery

46 Z4

52 47 50

25 12

G G

27 56 45

31 16

37 29

34 35

6 7

40 9b

5m 48 Post-hole (F471)

1 Graves 1–8

16 Wall (F848)

32 Stones (F8022)

2 Retaining wall (F863)

17 Clay (F803)

3 Cell B

18 Burial (F208)

33 Gravels (F8003, F8004, F8006, F8008, F8030)

4 Church enclosure wall

19 Cist-grave (F214)

35 Clay (F841)

5 Wall (F91)

20 Soil and stone (F4)

36 Clay and stone (F891/F893)

6 Trough (F845) filled with silt (F852) and rubble

21 Find spot of Pilgrim stones 4:1, 4:4 & 4:5

37 Cobbles (F894)

7 Soakaway (F846)

22 Cross base (F415)

38 Natural gravel (F830 w)

8 Wall-face (F832)

23 Silty clay (F479)

39 Clay and rubble (F834)

9a Paving (F833a)

24 Clay and stones (F459)

40 Possible surface (F838)

9b Paving (F833a)

25 Silty clay (F458)

41 Ash (F48a)

10 Extended monastic enclosure wall (F4025)

26 Silty clay (F463)

42 Rubble and clay (F78)

27 Ash (F445)

43 Schist (F847) sealed in fine soil (F837)

Annulus

44 Ash (F849) above Grave 11

Ash

45 Paving (F467)

Obscured

46 Building (F427)

Conjectural

47 Construction trench (F470)

Limit of excavation

11 Cell A 12 Annulus (F428) 13 Boundary wall (F813) 14 Paved path (F875) with upper layer of paving stones (F875a) 15 Paving (F422)

28 Ash (F472) 29 Silt (F8013) 30 Silt (F880/F881F/890) 31 Ash (F879a/b)

34 Burnt debris (F841)

49 Passage (F437) 50 Boulders (F427a) 51 Stony clay (F469) 52 Post-holes (F495a/b) 53 Passage (F4028) 54 Bank of stone and soil (F472) 55 Ash (F4005) 56 Schist clay (F4071) Lintels Upright stones G

Granite

Fig. 10.10 Plan of the monastery at High Island showing features associated with the later medieval period; Phase 3. Section Z3–Z4, see Fig. 10.5.

319

appendix a

A nal y s i s o f m o r ta r sa mpl e s: vi sua l a nd p e t ro gra ph i c a n al ysi s of a ggr e ga te, b i nd e r a n d a ddi ti ons Sara Pavía

Summary

The aim of this report is to study the type and composition of mortars from the monastic site on High Island, Co. Galway. This report was prepared with the objectives of: establishing different phases of building and rebuilding by comparing the different types of mortar used; determining the origin of the raw materials; and recording any features of the mortars that may illustrate the technology used to build the monastery. Eight mortars were analysed, using both physical and chemical methods. A number of conclusions were established concerning origin of the aggregate and the type and origin of lime binders. Building phases could not be clearly determined, however, as seven out of the eight mortars studied are similar in texture and composition and were made with local raw materials. However, variations in the composition of the aggregate tend to suggest that two of these seven mortars may belong to a separate building phase, the others being made with similar raw materials and technology at a later building stage. The mortars analysed are traditional lime-based mixes consisting of varying amounts of aggregate bound with carbonated lime. The petrographic features of the binder suggest that the lime used to make the mortars does not have any hydraulic properties. Most of the lime mortars analysed contain abundant shell and are heavily weathered. They are rich in lime binder and were made with a local aggregate where schist predominates (bedrock on High Island). Petrographic analysis suggests that bivalve shells could have been used to produce the lime binder for these mortars. However, one of the mortars clearly differs from the others in the nature of the aggregate and the aggregate/binder proportions. This mortar was made with raw materials from the mainland, or from another island location with this geology; it does not contain shell and is only slightly weathered, maintaining cohesion. The lime used to make this mortar was probably obtained by burning carbonifer-

ous limestone, and the aggregate was probably gathered from a stream. Whether this mortar was made on the island or brought as such from the mainland/other island location cannot be determined. Most mortars were heavily weathered, suggesting that they have been exposed to wet and dry episodes. They showed decay by salt crystallisation and calcite dissolution. Calcite dissolution is typically due to long-term action of moisture and is often recorded in historical lime mortar.

Introduction

This report includes the analysis of eight mortar samples gathered through the course of archaeological excavation. A detailed scientific study of ancient mortars can provide information on the different phases of building and rebuilding, on the origin of the raw materials used to make the mortars and on the level of building technology and the processing of building materials. Most historical mortars are lime-based and consist of varying amounts of aggregate of different origin and composition, bound with lime. Early lime mortars in Ireland tend to show certain common features. Due to long-term exposure, these mortars are often weathered. The lime binders tend to dissolve and the aggregate may detach and fall. Historical lime-based mortar is typically rich in binder and quite porous. The composition and morphology of the aggregate vary significantly, and the proportions of aggregate:binder typically range between 1.5:1 and 3:1. The mortars from High Island were analysed with both chemical and physical methods. ‘Materials and methods’, below, includes a detailed description of the techniques used for analysis. All technical terms in this report are explained in the Glossary at the end of the Appendix.

Sampling

Eight samples of mortars were gathered through the course of archaeological excavation (Table 1). All features that produced mortar are listed in the tables on pages 321–4.

320 high island: excavation of an early medieval monastery Table 1 Mortar samples studied.

Sample and Feature no. ’98 22) F262 ’98 18) F209 ’96 2) F10 ’98 20) F10 ’96 3) F209 ’96 4) F25a ’97 6) F66 ’97 7) F69

Supposed building phase Phase 1, 2 Phase 2 Phase 2 Phase 2 Phase 2 Phase 2 Phase 2 Phase 1

Type

Location

Mortar Mortar Mortar Mortar Mortar Mortar Mortar Mortar

Beneath east wall of church Behind Cross-slab 15 in altar Core of east wall of church Interior face east wall of church South side altar and east wall of church Sealed beneath recumbent slab, pseudo Grave 6 Uppermost fill beneath recumbent slab, Grave 3 Lowermost fill, Grave 3.

Some samples contained loose material as well as compact fragments of mortar and detached aggregate. In these cases, the compact fragments and loose matter were studied separately.

Materials and methods

Eight mortar samples were analysed using both physical and chemical methods. The samples were first cleaned and examined with the naked eye, and then with the help of a magnifying glass and a stereo microscope reaching up to 100 magnifications. Several fragments of mortar were selected for standard thin section preparation and petrographic analysis. Similarly, the loose aggregate was brushed and examined with the naked eye, and then with the help of a magnifying glass and a stereo microscope. The loose matter generally consisted of aggregate, mainly schist (occasional quartzite) of varying size and shape, lime lumps, fragments of shells, rests of lime binder, mud, charcoal and organic debris. In order to preserve the original features of the mortars, the samples were impregnated in a deep blue-coloured resin under vacuum before thin sectioning. This resin appears with a deep blue colour in the petrographic microscope plates below. The impregnation with resin pre-consolidates the samples, preserving their minerals and micro structure during thin section preparation. The thin sections were fabricated by cutting and polishing the impregnated fragments with oils. The use of water was avoided for sample cleaning as well as for cutting and polishing in order to preserve soluble minerals and minerals that are susceptible to hydration, such as the lime or gypsum binders that may be present in the mortars. The thin sections were polished to the standard thickness of 20 microns, covered with a glass slip and examined with the petrographic microscope. The petrographic examination was carried out by using transmitted parallel and polarised light.

Results

The loose matter accompanying the mortar generally consisted of fragments of schist (occasionally gneiss), of varying size and shape, lime lumps, fragments of shells, traces of lime, mud, charcoal and organic debris. The tables in the following pages include the summarised results.

Aggregate Coarse Lime largely dissolved. Varies from fractured- Abundant schist white to weathered-dark. Binder rich. Uniform, finely grained, with 10% shell shrinkage fractures. Recrystallised shell. Schist

Binder

Petrographic microscopy

Visual and stereo microscopy

’98 18 (F209)

Aggregate Coarse Very white lime, showing typical dissolution. Green schist (< 10mm) predominates Occ. gneissic schist Recrystallised shell transformed within 25% shell binder. Binder-replaced shell. Gneissic schist Dissolution textures. Possible portlandite in ghost of shell and sev- Schist eral lumps.

Binder

Coarse white-lime mortar with abundant schist

Petrographic microscopy

Visual and stereo microscopy

’98 22 (F262)

Medium to coarse lime mortar with abundant schist

Fine Schist mica quartz Shell mica quartz

Schist shell chlorite

Fine Abundant mica Abundant mica quartz shell chlorite

Medium

Abundant schist shell quartz

Medium

Wood fragments <1mm

1–2%

Aggregate oxidation Wood fragments <1mm

Good cohesion

Charcoal Other

3–4%

Charcoal Other

White 4–5mm shell

Shells

appendix a 321

Abundant portlandite.

Petrographic microscopy

Lime lumps. Lime affected by dissolution.

White, non-carbonated lime showing portlandite. Dissolution textures.

Petrographic microscopy

Binder

Visual and stereo microscopy

’98 20 (F10)

Medium-grain lime mortar with mussels

Scarce white lime lumps recovered, some could have been shells.

Binder

Visual and stereo microscopy

’96 2 (F10)

Coarse lime mortar with abundant schist aggregate

Medium Gneiss quartzite

Schist mica chlorite calcite

Quartzite 25% schist/gneiss predominates shell Occ. gneiss

Schist

Medium Schist

Aggregate Coarse Scarce gneiss (<15 mm)

Scarce coarse on thin section

Occ. quartzite

Schist

Occ. quartzite

Aggregate Coarse Abundant schist

Mica mica chlorite quartz

Fine

Mica quartz

Fine Abundant mica quartz schist

12% Two types

Good cohesion. Bond with masonry is still preserved.

Charcoal Other

Very abundant coarse schist. Large fragment recovered for thin section (N side of E wall).

Very weathered.

Charcoal Other

Mussel Abunshell dant (10mm) (< 5mm)

Shells

7–10%

Shells

322 high island: excavation of an early medieval monastery

Binder rich. Varies from uniform, fine grained to fractured with transformed shell. Remains of shell inside lime lumps. Portlandite. Significant dissolution.

Petrographic microscopy

Very white lime showing dissolution.

Binder rich. White non-carbonated lime lumps, including portlandite. Transformed shell. Lime dissolution. Fractures.

Petrographic microscopy

Binder

Visual and stereo microscopy

’96 4 (F25a)

Medium-grain white-lime mortar

Darker lime. Weathered.

Binder

Visual and stereo microscopy

’96 3 (F209)

Medium to coarse lime mortar with abundant schist

Quartzite Gneissic schist

Aggregate Coarse Schist <12mm

Occ. granite (1 fragment only)

Abundant schist grenats quartzite

Medium Shell schist

Abundant schist shell

Shell

Quartzite 10% schist

Medium Schist

Aggregate Coarse Schist

Abundant mica quartz shell

Fine Shell mica

Abundant quartz mica schist/ gneiss

Fine Quartz

White shell

Shells

White shell

Shells

Brown staining by weathered schist aggregate. 3% ash Lime and aggreand char- gate similar to coal ’98 18. Amount of charcoal also similar.

(4–5 mm)

Charcoal Other

AbunLarge mortar dant fragments sur(5–2mm) vived. Very weathered. 4%

Charcoal Other

appendix a 323

Uniform micrite. Fractured. Recrystallised shell. Iron oxides. Typical lime-dissolution textures.

Petrographic microscopy

Shell

Gneissic schist

70% schist

Aggregate Coarse Abundant quartzite. Up to 15x20mm. Black angular aggregate.

Darker lime. Homogeneous texture. Lime lumps 1–5mm.

Ghosts of fossils in lime binder. Well-sorted aggregate, well-distributed within binder. Even texture.

Petrographic microscopy

Binder

Visual and stereo microscopy

’97 7 (F69)

Chert Slate

Aggregate Coarse Very scarce. Angular 5mm. Black slate and rounded black limestone 10x5mm. 3%

Medium to fine lime mortar with chert, limestone and slate aggregate

Very white lime showing extensive dissolution.

Binder

Visual and stereo microscopy

’97 6 (F66)

Coarse white-lime mortar with large shells

Mica chlorite quartz shell

Fine

Abundant limestone quartz feldspar quartzite mica shale

Abundant quartz feldspar mica chert limestone

Medium Fine Abundant Abundant Mica 2–3mm

Schist

Medium

Shells

10%

Large white shells 20x15mm 15x10mm

Shells

Traces of wood.

Traces

1% red ceramic fragments.

Possible ceramics. (4–5mm) Green algae coating.

Charcoal Other

Snails (probably contamination).

Charcoal Other

324 high island: excavation of an early medieval monastery

appendix a 325

Discussion The mortar samples analysed are traditional lime-based mixes. Traditional lime mixes consist of varying amounts of aggregate of different type and nature, bound with lime. Even though all samples fall into this general classification, the mortars show variations in the amount and nature of aggregate and lime binder. Most mortars were weathered, suggesting that they have been exposed to wet and dry episodes for some time. They were affected by binder dissolution, having lost their internal cohesion and showing a strong tendency to aggregate detachment and disintegration. They generally displayed rough surfaces and showed dissolution textures. Dissolution and recrystallisation processes were evident under microscopic examination. Dissolution and recrystallisation of lime binders by the action of moisture typically occur in historical lime mortar due to long-term exposure to wet and dry episodes. According to their composition and features, the eight mortars studied were classified into two different groups.

Group 1 The first group includes all mortars except F69. Common features of these mortars are: • they all contain a significant proportion of aggregate of crushed schist; their main aggregate is of angular schist; • the fine aggregate in these mortars is mainly angular quartz, mica and chlorite; • they contain charcoal; • they include abundant shell fragments; • they are rich in lime binder; • the lime binder is uneven, showing shrinkage cracks and partially transformed shell. Schist is locally found on High Island. The main aggregate used to make these mortars was therefore crushed local stone. The fine aggregate of quartz, mica and chlorite present in these mortars is probably derived from crushed or weathered schist and gneissic schist. The presence of charcoal is probably due to contamination arising from the burning of shell/limestone to produce lime. Lime was traditionally produced by burning alternate layers of fuel (wood, peat or coal) and limestone or shell in a kiln. These were left to burn and the lime unloaded from the top when cool. The simplest and earliest type of lime kilns tend to be stone-built. These were often temporary structures built to meet an immediate necessity. They were often allowed to collapse and perhaps rebuilt for the next firing. From the examination of the charcoal under the microscope, it can be concluded that, on High Island, wood was used as a fuel to produce the lime to make the mortar. The presence of wood in two

mortars (F262 (’98 22) and F25a (’96 4)) also leads to this conclusion. The lime binder consists of carbonated lime. It lacks cohesion, showing large areas of binder with shrinkage cracks, fine aggregate and shell. Several shell fragments were found to have reacted with the binder, others were fractured and partially transformed. The transformation and petrographic features of these shells could be taken as evidence of the crushing and burning of shells to provide lime. The lime used to make these mortars was probably calcium lime (non-hydraulic) (also known as aerial lime). This conclusion was gathered from the petrographic study of the mortars. The following observations provide evidence of the use of aerial lime in the mortars analysed. • No evidence of hydraulic minerals was found in the lime binder. • The abundant shrinkage fractures derived from the hardening of lime. The release of water and intake of CO2 during carbonation of aerial lime cause shrinkage, which leads to typical fracturing patterns in the lime binder. • As mentioned before, the degree of transformation and petrographic features of the shells found in the mortars suggests that shells were crushed and burned to provide lime. The burning of shell produces aerial lime, the use of shell-derived lime would therefore be consistent with the use of aerial lime in these mortars. There are slight differences between the mortars in this group with regard to the nature of the aggregate. Five out of the seven mortars in this group contain, always in very small amounts, aggregate of other rocks apart from schist: quartzite, gneiss and granite. Mortars F262 (’98 22) and F209 (’98 18) contain aggregate of schist only, whereas mortars F10 (’96 2), F25a (’96 4) and F66 (’97 6) contain quartzite; and F10 (’98 20) and F209 (’96 3) contain gneiss and granite, respectively. Quartzite, gneiss and granite are also locally found on High Island. It is therefore possible that, even though all aggregate in the mortars of Group 1 is local, those containing aggregate different than schist were made during another building phase, with similar raw materials and similar technology. The mortar ’98 20 (F10) differs slightly from the others in the fact that it contains mussel shell and the lime is darker than on the other samples, however no conclusions could be withdrawn from these observations. Mussel shell may have been easily collected in the monastic period. Crystals of portlandite (lime -Ca (OH)2) were observed in several mortar samples in this group (see ‘Results’). As mentioned before, portlandite hardens by taking CO2 in the atmosphere (process known as carbonation) forming carbonated lime. The time needed for aerial lime to become a hard binder by carbonation varies from days to

326 high island: excavation of an early medieval monastery

centuries, depending mainly on the porosity of the mortar and the environmental conditions of the area where the mortar was placed. For example, in a sealed atmosphere, ancient mortar can be found where the lime has not yet finished carbonation. The mortars containing portlandite could have remained in a sealed atmosphere.

Group 2 The second group is formed by a single mortar sample, F69 (’97 7). It contains siliceous aggregate of angular and sub-rounded quartz and chert, fragments of shale, sandstone and limestone. This mortar also contains burnt clay.

The aggregate is not local, but was probably gathered from a stream sediment on the mainland or another island location. The limestone aggregate contains fossils typical of carboniferous limestone. This mortar is better preserved than the others. It is also more finely textured and a more elaborated mix, containing well sorted aggregate evenly distributed within the lime binder. The lime used to make this mortar is also different from the lime used in the other mortars. This lime is more even, its grain size is consistent and has uniformly carbonated. Occasionally, it forms rounded lime lumps. A large fragment of fossiliferous limestone showed a reaction rim with

Table 2 Mortar samples by phase and feature number. (*) sample analysed with petrographic microscopy. (S = small; M= medium; L = large)

Phase 1 1 1 1

Feature no. 57 38 *69 216

Sample no. ’00AF ’96 (1) ’97 (7) ’96 (i)

Area 1 1 1 2

1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 4 4 4 4

259 *262 265 10 10 10 10 *10 *10 11 *25a *66 209 209 209 *209 *209 319 202 202 206 214a/b 200 201 243 311

’98 (23) ’98 (22) ’98 (27) ’97 (8) ’97 (ii) ’98 (19) ’98 (21) ’96 (2) ’98 (20) ’96 ’96AE (4) ’97 (6) ’97AA (9) ’98D (25) ’98M (13-17) ’98N (18) ’98 (AC) ’98F+G ’95A ’95E ’96AF (5) ’96H ’95D ’97 (12) ’98 (26) ’97 (10)

2 2 2 1 1 1 1 1 1 1 1 1 2 2 2 2 2 3 2 2 2 2 2 2 2 3

Description Fill over burial (F37), Grave 5 Fill over skull area burial (F40), Grave 4 Fill over burial (F70), Grave 3 Surface associated with the pre-church ‘slotted features’ Fill of slot (F260) Deposit sub east wall of church Fill of slot (F260) Church wall-face behind altar Rubble fill of east wall of church South-east corner church interior Church wall-face; north side Church wall core Church wall-face behind altar Lower fill, Grave 1 Fill under recumbent slab, pseudo Grave 6 Fill above slabs (F67), Grave 3 West face of altar (F209) Altar core level: upper level Altar core level: middle level Altar core level: lower level South side altar, attached also east wall of church Burnt debris within Cell B Trampled floor church interior Trampled floor church interior Fill over burial (F208), Grave 9 Pseudo Grave 10 within church Pre-excavation surface, church interior Post-medieval rubble within church Rabbit disturbance church interior Rubble surface Cell B

Quantity S S S S S S S L L S M L L S S M M L L S L S L L S L L M S S

appendix a â&#x20AC;&#x192; 327

the surrounding lime binder. Internally, this fragment was also partially transformed. This could be evidence of the use of limestone as a source of lime.

Glossary

Conclusions

Binder: material filling interstices between aggregate and holding the aggregate in a mortar.

Seven of the eight mortars analysed are similar. They are rich in lime binder and contain a significant proportion of crushed schist; charcoal; abundant shell fragments; and fine aggregate of angular quartz, mica and chlorite. The main aggregate used to make these mortars was crushed local stone. There are slight differences between the mortars in this group with regard to the nature of the aggregate. Five of the seven mortars in this group contain, always in very small amounts, aggregate of other rocks as well as schist, namely quartzite, gneiss and granite. These rocks are also found locally on High Island. It is therefore possible that even though the aggregate in the seven mortars is local, those five containing aggregate in addition to schist were made during another building phase, with similar raw materials and similar technology. Petrographic features, including the lack of hydraulic phases and the abundant shrinkage fractures in the lime binder, suggest that these mortars were made with calcium lime (non-hydraulic). Transformation and petrographic features of shells found in the lime binder suggest that shells were used to provide lime. The presence of charcoal is probably due to the burning of the shell/limestone to produce lime. From the examination of the charcoal under the microscope, it can be concluded that wood was used as a fuel to produce the lime. The lime was probably obtained by burning alternate layers of fuel and shell/limestone in kilns. The kilns were probably stone-built, temporary structures, built to meet an immediate necessity. Mortar F69 (â&#x20AC;&#x2122;97 7) largely differs from the others in texture and composition, as well as in the amount and origin of aggregate and binder. This mortar is better preserved than the others. It is more finely textured and a more elaborated mix, containing well sorted aggregate evenly distributed within the lime binder. Several petrographic findings link this mortar to the mainland or to another island site with similar geology. The composition and morphology of the aggregate suggest that it was probably gathered from a stream sediment. The limestone aggregate contains carboniferous fossils typically found in the limestone from the mainland. The lime used to make this mortar is also different from that of the other mortars. It is even and consistent and has uniformly carbonated. Petrographic observations provided evidence of the use of limestone as a source of lime to make this mortar.

Mortar: mixture of binder, fine aggregate and water that hardens.

Aggregate: sand and stone mixed with binder and water to make mortar. It is the solid material that provides a supporting rigid frame for a binder. Traditional lime-based mixes: mortars consisting of varying amounts of aggregate of different type and nature, bound with lime. Petrographic analysis: petrography is the systematic description and interpretation of rock textures and mineralogy in thin section and as hand specimens. The petrographic microscope allows the examination of thin sections of stones (brick, mortar or other materials). The minerals composing these materials are identified by studying their optical properties. The nature, composition, size and shape of the mineral grains; their relationships and arrangement; the presence of pores, cracks, decay processes and textures can be studied and the presence and composition of matrix and cements can be determined. The microscope is fitted with two lenses to allow specimens to be examined in transmitted polarised light from a light source below the stage. It can be used to identify sources of raw materials and to attribute stone artefacts to their geological source. Calcium lime/hydrated lime/aerial lime /slaked lime: pure lime of chemical composition Ca(OH)2, obtained from the burning of calcitic limestone, chalk or other calcareous materials such as shell. This lime hardens in contact with the atmosphere by losing water and incorporating CO2 (process known as carbonation) and has no hydraulic properties. Pure calcitic limestone: calcareous material containing at least 95% of calcium carbonate (CaCO3). Contraction coefficient: amount of shortening and tensing of solids due to changes of physical state, e.g water loss by dehydration, responsible for solids in a plastic state to change into hard solids, is accompanied by a certain amount of shortening and tensing. Portlandite: mineral of chemical composition Ca(OH)2forming calcium lime.

328 high island: excavation of an early medieval monastery

APPENDIX A – Petrographic microscope plates

Mortar F262. General view, including coarse shell (right), charcoal (left), fine quartz and mica and fragments of schist. Parallel polars. 2X. Field of view 4.2mm.

Mortar F262. Detail of fine aggregate of quartz, schist and mica and fine charcoal in a carbonated lime binder, heavily weathered. Parallel polars. 2X. Field of view 4.2mm.

Mortar F209. Fractured lime binder partially missing and coarse schist aggregate. Crossed polars. 2X. Field of view 4.2mm.

Mortar F209. Charcoal, fine shell and fine quartz aggregate in a carbonated lime binder partially missing. Coarse gneiss fragment to upper left. Parallel polars. 2X. Field of view 4.2mm.

Mortar F10 (’98 20). Detail of binder dissolution and fracturing, and fine charcoal, shell, schist and quartz. Parallel polars. 2X. Field of view 4.2mm.

Mortar F10 (’98 20). Coarse schist aggregate and finer shell, quartz and schist in a fractured lime binder. Crossed polars. 2X. Field of view 4.2mm.

appendix a â&#x20AC;&#x192; 329

Mortar F10 (â&#x20AC;&#x2122;93 2). Large shell and fine aggregate in a Mortar F10 (â&#x20AC;&#x2122;93 2). Detail of fine charcoal and fine weathered lime binder. The shell appears fractured and aggregate of shell, quartz and chlorite. Parallel polars. partially filled with lime. Parallel polars. 2X. Field of 20X. Field of view 0.6mm. view 4.2mm.

Mortar F209. General view of binder-rich area, showing string weathering by dissolution and salt crystallisation. Parallel polars. 2X. Field of view 4.2mm.

Mortar F209. Detail of salt minerals in photograph above partially replacing the carbonated lime binder. Parallel polars. 20X. Field of view 0.6mm.

330â&#x20AC;&#x192; high island: excavation of an early medieval monastery

Mortar F25a. General texture of mortar, showing coarse aggregate of mica-schist with grenats, charcoal and fine aggregate of mica, angular quartz and shell. Crossed polars. 2X. Field of view 4.2mm.

Mortar F66. General texture of mortar showing coarse aggregate of mica-schist and shell (lower left corner), charcoal and fine angular quartz, mica and shell. Crossed polars. 2X. Field of view 4.2mm.

Mortar F69. General composition of mortar with abundant aggregate, including coarse slate and shale, abundant quartz, feldspar (twinned) and mica. Crossed polars. 2X. Field of view 4.2mm.

Mortar F69. General composition, including rock fragments such as limestone, chert and shale; and finer abundant quartz, calcite, feldspar and mica. Parallel polars. 2X. Field of view 4.2mm.

Mortar F69. General view of mortar, including coarse quartzite and finer quartz, feldspar and mica. Crossed polars. 2X. Field of view 4.2mm.

331

appendix b

A nal y s i s o f s a m pl e s of cl a yey soi l Sara Pavía and Jason Bolton

Introduction

During the excavations carried out at the early medieval monastery on High Island, a distinctive clay deposit (appearing sometimes as a more or less continuous layer and, occasionally, discontinuous) was identified by excavation director Georgina Scally. The clay has been found in places that would suggest it was being purposefully used as a bedding material (under structures) and as a bonding material, possibly an early type of mortar. According to Scally, similar type natural clay also occurs beneath the sand along the shore of Omey Island in Co. Galway. This clay was used in the past as a fire-proof material at the back of fireplaces and in the foundations of buildings (F. Mulkerrin, pers. comm.).

Objective

The aim of this report is to analyse the mineralogical composition of samples of clayey soil taken from High Island and Omey Island and discuss the origin of such deposits. To this end, the composition of natural, undisturbed boulder clay will be compared with that of the clay samples taken from archaeological levels at the monastery.

Sampling

Seven samples of soil were gathered for analysis. Six samples were taken from the archaeological excavation on High Island, and one from the shore on Omey Island. Three out of the six High Island samples were natural boulder clay, gathered from the lowest level of the excavation. The Omey sample was also gathered from a natural clay deposit. Details of the samples are outlined in Table 1.

Materials and methods

The samples were exposed to dry. They were visually examined with the naked eye, which allowed an initial approximation as to the composition of the samples. Several fragments were selected for further analysis by X-ray diffraction. A number of fragments were selected from each sample, with the exception of F270. In order to avoid contamination from organic mate-

Table 1 Details of clayey soil samples analysed.

Feature no. 205

Sample no. ’00AK

803

’00AE

171

’00AH

271

’00T

270

’00Q

336

’00AM

Omey

’00AL

Description Clay layer contemporary with the paved area (F245) beneath the church. Packing around foundations of beehive hut, Cell B. Natural boulder clay sub graves outside the church. Natural boulder clay inside the church. Ball of clay found at base of core of east wall of church. Natural silt/clay from inside beehive hut, Cell B. Natural clay collected from below sand along eastern shore of Omey Island.

rial or other layers, the selected fragments were opened and the core material in the fresh cut taken for analysis. The ball of clay (sample F270) was opened and clay taken from its interior. The amount of material taken for analysis from each sample was approximately 1/3 oz. The samples were prepared for mineralogical analysis by X-ray diffraction. The samples were dried for 48 hours at 28ºC. They were ground with a mortar and pestle and the resultant dust mounted in a metal support. The metal support was introduced in an X-ray diffraction chamber to carry out the analysis. The X-ray diffraction method used was the powder method. An X-ray beam of known wavelength is directed to the powdered sample. The beam is then diffracted by the crystalline minerals in the powdered sample. By recording the angular positions of diffracted beams (which appear as peaks in Table 4) the internal structure of the

332 high island: excavation of an early medieval monastery Table 2 Results from X-ray diffraction analysis.

Feature no. Mineralogical composition 205 Quartz, muscovite and feldspar. 803 171 271 270 336 Omey

Quartz, muscovite and feldspar. Clay minerals, probably chlorite and kaolinite. Quartz, muscovite and feldspar. Clay minerals, probably chlorite and kaolinite. Quartz, muscovite and feldspar. Kaolinite. Quartz, muscovite and feldspar. Quartz, muscovite and feldspar.

minerals is determined and the minerals are identified. The samples were analysed using a diffractometer Philips using Cu α radiation. The area was scanned between 3 and 55 degrees 2θ.

Results

Results from visual analysis revealed that the samples contained fragments of schist and variable amounts of organic matter. The results obtained from the interpretation of the X-ray diffraction diagrams and peak lists can be found in Table 4.

Discussion

The ball of clay found at base of core of east wall is of a very particular nature, while the remaining samples showed a similar mineralogical composition.

The ball of clay (F270) The ball of clay found at base of core of east wall is composed of very pure clay, mostly kaolinite. Alkali feldspar, which is an abundant mineral in granite, is decomposed by weathering into kaolinite. Alkali feldspar is widely distributed in igneous rocks, gneisses and sedimentary rocks. In these rocks, kaolinite usually occurs in small amounts and is generated by weathering. When kaolinisation in granites is very strong, the rock is reduced to a rotten, friable condition, quartz being the only mineral to survive the process. In this condition, the altered granite can be easily broken down by a highpressure water-jet or mechanical means, with the kaolinite settled out from suspension in setting pools to produce china clay (kaolin). Granite is found on High Island as boulders. In a granite sample previously analysed (Pavía and Bolton 2001), feldspar was very abundant, accounting for 70–80% of the total of the minerals in the sample. However, natural deposits of kaolinite have never been found on the island, so it is therefore unlikely that the kaolinite originated from

Description Clay layer contemporary with the paved area beneath the church. Packing around beehive hut, Cell B. Natural boulder clay below graves outside the church. Natural boulder clay inside the church. Ball of clay found at base of core of east wall of church. Silt/clay from inside beehive hut, Cell B. Natural clay collected from below sand along eastern shore, Omey Island. the weathering of the granite deposits on High Island. Rather, the kaolinite analysed in this report could have been taken onto the island, to be used as a raw material, for ceramic production or paper-making. Kaolinite has been used in the ceramic industry and in the production of paper for centuries. It was also used in the production of common pipes in Europe and Asia for centuries.

The remaining samples The remaining samples show very similar mineralogical composition. The minerals making up the samples analysed are consistent with those found in the local rocks in this area of west Galway. (See Table 4 for the X-ray diffraction diagram showing the mineralogical composition of the samples.) The Omey samples consist mainly of quartz, muscovite and feldspar. These are the main components of the Omey Adamellite, which is the local bedrock in Omey Island. The six samples from the archaeological excavation on High Island display similar mineralogical composition, consisting of mainly quartz, muscovite and feldspar. The relative amounts of these minerals are similar in all samples studied. The main difference among the High Island samples is that, in addition to the minerals above, two of them (F803; F171) contain clay minerals (kaolinite) and chlorite. The first sample was the packing around the beehive hut and the second natural, undisturbed clay from beneath the graves. Sample F205 (the clay layer contemporary with the pre-church paved area) consists mainly of quartz, muscovite and feldspar and does not contain clay minerals (kaolinite) and chlorite, a composition consistent with the natural clay soil. The minerals present in the clay samples from High Island are consistent with those found in the local bedrock. Metamorphic rocks form the major part of High Island. These are mainly schists, gneisses and quartzites grading into impure marbles and granulites. These

appendix a 333

rocks are composed of quartz, muscovite, chlorite and clay minerals, including sericite. Other minerals, including metamorphic aluminosilicates and grenats, are also present in smaller amounts. These minerals are consistent with those found making up the clayey soil analysed. The minerals present in the clayey soil samples from High Island are therefore consistent with those found in the local bedrock. The differences in the clay mineral content mentioned above are probably due to natural variations in the mineralogical composition of the bedrock.

Conclusion With the exception of the ball of clay (F270), all samples analysed are of a similar mineral composition and therefore are of similar origin. The minerals present in the samples analysed (from both High Island and Omey Island) are consistent with those found in the local bedrock. The samples probably originated naturally as a result of the weathering of the upper bedrock. Slight mineralogical variations recorded in the samples studied reflect natural variations in the mineralogical composition of the bedrock and in the dynamics of the process that generated the deposits (ice could have contributed, to a certain extent, to form these clayey deposits).

Table 3 Description of samples analysed on XRD diagrams.

Sample

205

Notation in XRD diagrams 205

803

803 S

171

271

270

336

Omey

Description

Clay layer contemporary with sub church paved area. Packing around foundations of beehive hut, Cell B. Natural boulder clay below graves. Natural boulder clay below church. Ball of clay found at base of core of east wall of church. Natural silt/clay from inside beehive hut, Cell B. Natural clay collected from below sand along eastern shore of Omey Island.

Table 4 X-ray diffraction diagrams comparing the mineralogical composition of the natural soil deposits from Omey and High Island with that of the clay samples recovered during the archaeological excavation at High Island’s monastic site (note: 8035 cpi refers to sample 803s).

334

bibliography Abbreviations AI

Annals of Inisfallen (ed. Mac Airt, 1951).

Ancient Laws and Institutes of Ireland (see Hancock et al. 2000).

Annals of Tigernach (ed. Stokes, 1895–1897).

Annals of Ulster (ed. Mac Airt & Mac Niocaill, 1983).

(Betha Féichín), first Irish life of Féichín (ed. Stokes, 1891, 320–38).

(Betha Féichín), second Irish life of Féichín (ed. Stokes, 1891, 338–53).

Colgan’s conflate life of Féichín (Colgan 1645, 133–9).

CGH

Corpus genealogiarum Hiberniae (ed. M. A. O’Brien. Dublin, 1962).

CGSH

Corpus genealogorum sanctorum Hiberniae (ed. P. Ó Riain. Dublin, 1985).

Chronicon Scottorum (ed. Hennessy, 1866).

DIL

Royal Irish Academy, Dictionary of the Irish language based mainly on Old and Middle Irish materials. Dublin, 1913–75. Repr. in compact ed., Dublin, 1983.

Annals of the Four Masters (ed. O’Donovan, 1848–51).

Thes. Pal. Stokes and Strachan, ed. 1903–1910. V

(Vita Fechini), Latin life of Féichín (ed. Plummer, 1910, vol. 2, 76–86).

JRSAI

Journal of the Royal Society of Antiquaries of Ireland.

JGAHS

Journal of the Galway Archaeological and Historical Society.

JWHS

Journal of the Westport Historical Society.

PRIA

Proceedings of the Royal Irish Academy.

NLI MS

National Library of Ireland, manuscript.

Rawl. MS

Oxford, Bodleian Library, Rawlinson collection, manuscript.

RCAHMS Royal Commission on the Ancient and Historical Monuments of Scotland.

Abbott, T.K. and Gwynn, E.J. 1921 Catalogue of the Irish manuscripts in the library of Trinity College, Dublin. Dublin. Ahlstrom, B., Brekke, B.F. and Hemmingsson, B. 1976 The coinage of Norway. Stockholm. Allen, J.R. and Anderson, J. 1903 The Early Christian monuments of Scotland, 2 vols. Facsimile Balgavies 1993. Allen, J.R. and Anderson, J. 1992 [Facsimile reprint] The high crosses of Ireland. Felinfach. Allen, R. 2011 Plant macrofossils. In E. Bourke, A. Hayden and A. Lynch, Skellig Michael, Co. Kerry: the monastery and South Peak. Archaeological stratigraphic report: excavations 1986–2010. (http://www.worldheritageireland.ie/skellig-michael/publications). Anderberg, A.L. 1994 Atlas of seeds part 4: ResedaceaeUmbelliferae. Stockholm: Swedish Museum of Natural History. Anderson, G. 2008 Birds of Ireland: facts, folklore and history. Cork. Anderson, A.O. and Anderson, M.O. (eds) 1991 Adomnán’s Life of Columba. Oxford. Andrews, J.H. 2003 Sir Richard Bingham and the mapping of western Ireland. PRIA 103C, 61–95. Armstrong, L. 1978 Woodcolliers and charcoal burning. Horsham. Bailey, R. 1989 St Cuthbert’s relics: some neglected evidence. In G. Bonner et al. (eds), St Cuthbert, his cult and his community to A.D. 1200, 231–46. Woodbridge. Bailey, R. and Cramp, R. 1988 Corpus of Anglo-Saxon stone sculpture II: Cumberland, Westmorland and Lancashire North-of-the-Sands. Oxford. Bakels, C.C. 2000 Pollen diagrams and prehistoric fields: the case of Bronze Age Haarlem, the Netherlands. Review of Palaeobotany and Palynology 109, 205–18. Becker, U. 2000 Continuum encyclopaedia of symbols, trans. by L.W. Garner. London. Behre, K.E. 1981 The interpretation of anthropogenic indicators in pollen diagrams. Pollen et spores 23(2), 225–45.

bibliography 335

Beijerinck, W. 1947 Zadenatlas der Nederlandsche flora. Wageningen. Benediktsson, J. 1968 Íslendingabók Landnámabók, Íslenzk Fornrit 1, Reykjavik. Bentley, R.A. and Knipper, C. 2005 Geographical patterns in biologically available strontium, carbon and oxygen isotope signatures in prehistoric SW Germany. Archaeometry 47, 629–44. Berger, R. 1995 Radiocarbon dating of early medieval Irish monuments. PRIA 95C, 4, 159–74. Berggren, G. 1969 Atlas of seeds part 2: Cyperaceae. Stockholm. Beiler, L. 1963 The Irish penitentials. Scriptores Latini Hiberniae 5. Dublin. Bischoff, B. 1954 Wendepunkte in der Geschichte der lateinische Exegese im Frühmittelalter. Sacris Erudiri 6, 189–279. Boardman, S. and Jones, G. 1990 Experiments on the effects of charring on cereal plant components. Journal of Archaeological Science 17, 1–11. Boessneck, J. 1969 Osteological differences between sheep and goat. In D.R. Brothwell and E. Higgs (eds), Science in archaeology, a survey of progress and research. Bristol. Bollandists [Societé des Bollandistes] (ed.) 1643 Acta sanctorum… ianuarius. 2 vols. Antwerp and Brussels. Bourke, E. 2005 A preliminary analysis of the inner enclosure of Skellig Michael, Co. Kerry. In T. Condit and C. Corlett (eds), Above and Beyond: essays in memory of L. Swan, 121–38. Bray. Bourke, E., Hayden, A.R. and Lynch, A. 2011 Skellig Michael, Co. Kerry: the monastery and South Peak. Archaeological stratigraphic report: excavations 1986–2010. (http://www.worldheritageireland.ie/ skellig-michael/publications). Bracken, D., Hawkes, J. and Ó Riain-Raedald (eds) (forthcoming) Pilgrimage. Dublin. Bradley, J. 1988 The interpretation of Scandinavian settlement in Ireland. In J. Bradley (ed.), Settlement and society in medieval Ireland: studies presented to F.X. Martin, O.S.A. 49–78. Kilkenny. Brady, N. 1977 De oratorio: Hisperica famina and church building. Peritia 11, 327–35. Brash, R.R. 1872 Notes on the ecclesiastical architecture of Ireland. The Irish Builder 19, 19. Breen, C. and Forsythe, W. 2004 Boats and shipwrecks of Ireland. Stroud.

Brewer, J.S., Dimock, J.F., Freeman, E.A. and Warner, G.F. (eds) 1861–91 Giraldi Cambrensis opera. Rolls Series 21. 8 vols. London. Bronk Ramsey, C. 1995 Radiocarbon calibration and analysis of stratigraphy. Radiocarbon 37(2), 425–30. Buckley, L. 2005 Skeletal report. In J.W. Marshall and C. Walsh (eds), Illaunloughan Island: an early medieval monastery in County Kerry, Appendix 5. Bray. Buikstra, J.E and Ubelaker, D.H. 1994 Standards for data collection from human skeletal remains. In D. Aftandilian (ed.), Arkansas Archeological Survey Research Series No 44. Fayetteville, Arkansas. Bullock, P., Federoff, N., Jongerius, A., Stoops, G. and Tursina, T. 1985 Handbook for soil thin section description. Waine Research Publications. Byrne, F.J. 1973 Irish kings and high-kings. London. 2nd edn: Dublin, 2001. Campbell, A. (ed.) 1967 Æthelwulf. Oxford. Carter, S. 1998 The use of peat and other organic sediments as fuel in northern Scotland: identifications derived from soil thin sections. In C.M. Mills and G. Coles (eds), Life on the Edge: human settlement and marginality. Symposia of the Association for Environmental Archaeology No. 13. Oxford. Chadburn, R. and Hill, P. 1997 Exotic stones and gems. In P. Hill, Whithorn and St. Ninian: the excavation of a monastic town 1984-91. Stroud. Chadwick, N.K. 1975 The Vikings and the western world. In B. Ó Cuív (ed.), The impact of the Scandinavian invasions on the Celtic-speaking peoples c. 800-1100 AD. Introductory papers read at Plenary sessions of the International Congress of Celtic Studies held in Dublin, 6-10 July, 1959, 13–42. Dublin. Charles-Edwards, T.M. 2000 Early Christian Ireland. Cambridge. Charles-Edwards, T.M. 2003 Érlam: The patron-saint of an Irish church. In R. Sharpe and A. Thacker (eds), Local saints and local churches in the early medieval west, 267–90. Oxford. Charles-Edwards, T.M. 2004a Early Irish saints’ cults and their constituencies. Ériu 54, 79–102. Charles-Edwards, T.M. 2004b Ua Cerbaill, Máel Suthain (d. 1010). In Oxford dictionary of national biography. Oxford. [http://www.oxforddnb.com/ view/article/20485, accessed 26 February 2008]. Chenery, C., Lamb, A., Evans, J., Sloane, H. and Stewart, C. 2014 Appendix 3: Isotope Analysis of

336 high island: excavation of an early medieval monastery

individuals from the Ridgeway Hill Mass Grave. In L. Loe, A. Boyle, H. Webb and D. Score (eds), Given to the Ground: a Viking Age mass grave on Ridgeway Hill, Weymouth. Dorset Natural History and Archaeological Society Monograph Series no. 22, 259–84. Oxford. Clancy, T.O. 2001 The real St Ninian. Innes Review 52, 1–28. Clarke, A. 2012a Coarse stone artefacts. In C. Cotter, The Western Stone Forts Project, excavations at Dún Aonghasa and Dún Eoghanachta, Vol. 2, 57–91. Dublin. Clarke, A. 2012b Flaked lithics and coarse stone. In C. Cotter, The Western Stone Forts Project, excavations at Dún Aonghasa and Dún Eoghanachta, Vol. 2, 329–33. Dublin. Cochrane, R. 1892 Notes on the Ancient Monuments Protection (Ireland) Act 1892 and previous legislation connected therewith. JRSAI 22, 411–29 Colgan, J. (ed.) 1645 Acta sanctorum veteris et majoris Scotiae seu Hiberniae... sanctorum insulae. Louvain. Colgrave, B. and Mynors, R.A.B. (eds) 1969 Bede’s ecclesiastical history of the English People. Oxford. Collins, A.E.P. 1955 Excavations in Lough Faughan crannog, Co. Down, 1951–52. UJA 18, 45–82. Comber, M. 2006 Tom Fanning’s excavations at Rinnaraw Cashel, Portnablagh, Co. Donegal. PRIA 106C(1), 67–124.

Cotter, C. 1990 Excavations at Gragan West, Co. Clare. Unpublished report submitted to the National Monuments Service Archive, Department of Arts, Heritage and the Gaeltacht. Dublin. Cotter, C. 1995 Archaeological excavations at Skeam West. Mizen Journal (3), 71–8. Courty, M.A., Goldberg, P. and MacPhail, R. 1989 Soils and micromorphology in archaeology. Cambridge. Cowgill, J. 2003 The iron production industry and its extensive demand on woodland and resources: a case study from Creeton Quarry, Lincolnshire. In M. Murphy and P.E.J. Wiltshire, The environmental archaeology of industry, 48–57. Oxford. Coyle, J.B. 1915 The life of Saint Fechin of Fore, the apostle of Connemara. Dublin. Craig, D. 1997 The sculptured stones and Appendix 1. In P. Hill, Whithorn and St Ninian. The excavation of a monastic town, 433–41; 615–19. Stroud. Craig, D. 2006 The later carved stone assemblage. In C. Lowe, Excavations at Hoddom, Dumfriesshire: an early ecclesiastical site in South-West Scotland, 123–33. Edinburgh. Craig, H. 1957 Isotopic standards for carbon and oxygen and correction factors for mass-spectrometric analysis of carbon dioxide. Geochimica et Cosmochimica Acta 12, 133–49.

Connolly, A. 1994a Saddle querns in Ireland. UJA 57, 26–36.

Cramp, R. 1984 Corpus of Anglo-Saxon stone sculpture I: Co. Durham and Northumberland, 2 vols. Oxford.

Connolly, A. 1994b Castles and abbeys in Wales: refugia for ‘Mediaeval’ medicinal plants. Botanical Journal of Scotland 46(4), 628–36.

Cubitt, C. 2000 Monastic memory and identity. In W. Frazer and A. Tyrell (eds), Social identity in early medieval Britain, 253–76. Leicester.

Conway, M. 1999 Director’s first findings from excavations in Cabinteely. Dublin.

Cuppage, J. 1986 Archaeological survey of the Dingle Peninsula. Ballyferriter.

Coplen, T.B. 1988 Normalization of oxygen and hydrogen isotope data. Chemical Geology (Isotope Geosciences) 72, 293–7.

Curle, C.L. 1982 Pictish and Norse finds from the Brough of Birsay 1934-74. Society of Antiquaries of Scotland.

Corlett, C. 1999 Castledermot hogback. Archaeology Ireland 13 (4), 11.

Darling, W.G., Bath, A.H. and Talbot, J.C. 2003 The O and H stable isotopic content of fresh waters in the British Isles, 2: ground and surface waters. Hydrology and Earth System Sciences 7, 183–95.

Corlett, C. 2002 A cursed miracle. Archaeology Ireland 16(2), 13. Cormack, W.F. 1989 Two recent finds of exotic porphyry in Galloway. Transactions of the Dumfriesshire and Galloway History and Antiquarian Society 52, 149–50.

Davies, O. 1942 Contributions to the study of crannógs in south Ulster. UJA 5, 14–30.

Cotter, C. 1988 Gragan West. In I. Bennett (ed.), Excavations 1988, 9–10. Bray.

Dickson, C.A. and Mitchell, G.F. 1984 Appendix VI: botanical report on four samples of organic material, 219–22. In P.D. Sweetman, Archaeological �� excava-

De Breffny, B. and Mott, G. 1976 The churches and abbeys of Ireland. London.

bibliography 337

tions at Shop Street, Drogheda, Co. Louth. PRIA 84C, 171–224.

Farr, C. 1997 The Book of Kells: its functions and audience. London.

Diefendorf, A.F. and Patterson, W.P. 2005 Survey of stable isotope values in Irish surface waters. Journal of Paleolimnology 34, 257–69.

Farrell, R.T., Kelly, E.P. and Gowen, M. 1989 The Crannóg Archaeological Project (CAP), Republic of Ireland I: a pre-preliminary report. International Journal of Nautical Archaeology and Underwater Exploration 18, 123–36.

Doherty, C. 1985 The monastic town in early medieval Ireland. In H.B. Clarke and A. Simms (eds), The comparative history of urban origins in non-Roman Europe, Part I. British Archaeological Reports: International Series 255 (1), 45–75. Dolley, R.H.M. 1966 The Hiberno-Norse coins in the British Museum. London. Dumville, D. 1973 Biblical apocrypha and the Early Irish: A preliminary investigation. PRIA 73C, 299–338. Dwyer, P. 1981 Célí Dé: Spiritual reform in Ireland 750900. Dublin. Edward Milner, J. 1992 The tree book. London. Edwards, N. 1990 The archaeology of early medieval Ireland. London. Edwards, N. 2007 Corpus of early medieval inscribed stones and stone sculpture in Wales, Vol. 2. Cardiff. Etchingham, C. 1996 Viking raids on Irish church settlements in the ninth century. Maynooth. Etchingham, C. 1999a The idea of monastic authority in early Ireland. In J. Hill and C. Lennon (eds), Luxury and austerity: papers read before the 23rd Irish conference of historians held at St Patrick’s College, Maynooth, 16–18 May 1997, 14–29. Maynooth. Etchingham, C. 1999b Church organisation in Ireland AD 650–1000. Maynooth. Evans, E. 1957 Irish folk ways. London. Evans, J.A., Montgomery, J., Wildman, G. and Boulton, N. 2010 Spatial variations in biosphere 87Sr/86Sr in Britain. Journal of the Geological Society 5, 167, 1–4. Everson, P. and Stocker, D. 1999 Corpus of Anglo-Saxon stone sculpture, vol. 5: Lincolnshire. Oxford. Fanning, T. 1976 Early Christian monastic sites in Tipperary. In W. Hayes (ed.), Tipperary remembers, 34–41. Freshford. Fanning, T. 1981 Excavation of an Early Christian cemetery and settlement at Reask, Co. Kerry. PRIA 81C, No. 3, 67–172.

Fasola, U. 1975 Le catacombe di S Gennaro a Capodimonte. Rome. Fenton, A. 1978 (reprinted 1997) The Northern Isles: Orkney and Shetland. Edinburgh. Fisher, I. 2001 Early medieval sculpture in the West Highlands and Islands. Edinburgh. Fitzpatrick, E. 2007 Lecture on excavations carried out in the Cabhail Tighe Breac settlement, Cahermacnaghten, Burren, Co. Clare. National Museum of Ireland. Fleming, J. 1966 The “Dream of the Rood” and AngloSaxon monasticism. Traditio 22, 43–72. Forbes, R.J. 1956 Metallurgy. In C. Singer (ed.), History of technology 2, 47. London. Forsyth, K. 1995 The ogham-inscribed spindle-whorl from Buckquoy: evidence for the Irish language in pre-Viking Orkney? PSAS 125, 677–96. Fredengren, C., McClatchie, M. and Stuijts, I. 2004 Reconsidering crannógs in early medieval Ireland: alternative approaches in the investigation of social and agricultural systems. Environmental Archaeology 9(2), 161–6. Fried, J. 2003 The apocalyptic year 1000: religious expectation and social change, 950-1050. In R. Landes, A. Gow and D. Van Meter (eds), The apocalyptic year 1000. Oxford. Fry, S.L. 1999 Burial in medieval Ireland, 900-1500. A review of the written sources. Dublin. Gale, R. 1996 The pieces of charcoal. In P. Ashbee, Halangy Down, St. Mary’s, Isles of Scilly. Cornish Archaeology 35, 5–201. Gale, R. 1999 Knockans, Rathlin Island, SKR 94-96. Unpublished report prepared for Ulster Museum. Gale, R. 2003 Wood-based industrial fuels and their environmental impact in lowland Britain. In P. Murphy and P. Wiltshire, The Environmental Archaeology of Industry. Symposia of the Association for Environmental Archaeology 20, 30–47. Oxford. Gale, R. 2012 Charcoal. In C. Cotter, The Western Stone Forts Project: excavations at Dún Aonghasa and Dún Eoghanachta 2, 201–10. Dublin.

338 high island: excavation of an early medieval monastery

Gale, R. and Cutler, D. 2000 Plants in archaeology. London.

Hamilton-Dyer, S. 1994 Fish and marine invertebrates from Omey Island, Co. Galway. Unpublished report.

Gannon, A. 2003 The iconography of early Anglo-Saxon coinage. Oxford.

Hamilton-Dyer, S. 2000 Fish bones. In E.V. Murray and F. McCormick, Excavations of two Early Christian sandhill sites at Doonloughan, Slyne Head, Co. Galway. Unpublished report.

Gerard, J. and Johnson, T. 1633 The herball or generall historie of plantes: very much enlarged and amended by Thomas Johnson. London. Gibbons, M. and Gibbons, M. 2006 Soapstone as a cultural indicator on the Atlantic Seaboard – Scandinavian no more. Cathair na Mart, JWHS 25, 25–32. Gibbons, E.K. and Kelly, E.P. 2003 A Viking Age farmstead in Connemara. Archaeology Ireland 17 (1), 28–32. Gilchrist, R. 2008 Magic for the dead? The archaeology of magic in later medieval burials. Medieval Archaeology 52, 119–59. Gosling, P. 1993 Archaeological inventory of County Galway Volume 1: West Galway (including Connemara and the Aran Islands). Dublin. Gougaud, L. 1911 Étude sur les loricae Celtiques et sur les prières qui s’en rapprochent (pt. 1). Bulletin d’Ancienne Litérature et d’Archéologie Chrétienne 1, 265–81. Grant, A. 1975 The animal bones and Appendix B: the use of tooth wear as a guide to the age of domestic animals. In B. Cunliffe (ed.), Excavations at Portchester Castle, Vol. 1: Roman, 378–408; 437–50. London. Greene, S.A. 2009 Settlement, identity and change on the Atlantic islands of northwest Co. Mayo, c. AD 400-1100. Unpublished PhD thesis, UCD School of Archaeology. Grigson, G. 1958 The Englishman’s flora. London. Gwynn, A. and Hadcock, R.N. 1970 Medieval religious houses: Ireland. London. Haak, W., Brandt, G., de Jong, H.N., Meyer, C., Gansimeier, R., Heyd, V., Hawkesworth, C., Pike, A.W.G., Meller, H. and Alt, K.W. 2008 Ancient DNA, strontium isotopes and osteological analyses shed light on social and kinship organization of the later stone age. Proceedings of the National Academy of Sciences of the United States of America 105, 18226–31. Hall, V.A. 1988 The role of harvesting techniques in the dispersal of pollen grains of Cerealis. Pollen et spores 30, 265–70.

Hamilton-Dyer, S. 2002 Appendix 2: Fish and fowl bones. In M. Comber, M.V. Duignan’s excavations at the ringfort at Rathgurreen, Co. Galway, 1948–9. PRIA 102C, No. 5, 192–3. Hamilton-Dyer, S. 2005 Fish bone. In J. White Marshall and C. Walsh, Illaunloughan Island: an early medieval monastery in County Kerry, Appendix 4. Bray. Hamilton-Dyer, S. 2011 Bird and fish bones. In E. Bourke, A.R. Hayden and A. Lynch, Skellig Michael, Co. Kerry: the monastery and South Peak. Archaeological stratigraphic report: excavations 1986-2010, 437–49. [http://www.worldheritageireland.ie/skellig-michael/ publications]. Hammersley, G. 1973 The charcoal iron industry and its fuel 1540–1750. The Economic History Review, second series, 26 (4), 593–613. Hancock, W.N., O’Donovan, J., O’Mahony, T., Richey, A.G., Atkinson, R. (eds) 2000 Ancient laws and institutes of Ireland (6 vols). New York. Harbison, P. 1991 Pilgrimage in Ireland. The monuments and the people. London. Harbison, P. 1992 The High Crosses of Ireland. 2 vols. Bonn. Harbison, P. 1994 Irish High Crosses with the figure sculptures explained. Drogheda. Harbison, P. 1999 The Golden Age of Irish Art: the medieval achievement 600–1200. New York. Harbus, A. 2002 Helena of Britain in medieval legend. Woodbridge. Hardiman, J. (ed.) 1846 A chorographical description of west or H-Iar Connaught, written A.D. 1694 by Roderic O’Flaherty. Dublin. Harrison, S. 2001 Viking graves and grave-goods in Ireland. In A. Larsen (ed.), The Vikings in Ireland, 61–76. Denmark. Hayden, A.R. 2013 Early medieval shrines: new perspectives from Church Island, near Valentia, Co. Kerry. PRIA 113C, 67–138. Healy, J. 1890 An island shrine in the west. The Irish Ecclesiastical Record, third series, 11, 673–86. Hencken, H. 1936 Ballinderry crannog no. 1. PRIA 43C, 103–239.

bibliography 339

Hencken, H. 1938 Cahercommaun, a stone fort in Co. Clare. JRSAI 38 (extra volume), 1–82. Hencken, H. 1942 Ballinderry Crannog No. 2. PRIA 47C, 1–76. Henderson, G. 1980 Bede and the Visual Arts. Jarrow. Henderson, I. and Okaska, E. 1992 The Early Christian and inscribed stones of Tullylease, Co. Cork. Cambridge Medieval Celtic Studies 24, 1–36. Henry, F. 1933 La Sculpture Irlandaise pendant les douze premiers Siècles de l’ere chrétienne. Paris. Henry, F. 1937 Early Christian slabs and pillar-stones in the west of Ireland. JRSAI 67, 265–79. Henry, F. 1945 Remains of the Early Christian period on Inishkea North, Co. Mayo. JRSAI 75, 127–55. Henry, F. 1947 The Antiquities of Caher Island, Co. Mayo. JRSAI 77, 23–38. Henry, F. 1951 Miscellanea: habitation sites on Inishkea North, Co. Mayo. JRSAI 81, 75–6. Henry, F. 1952 A wooden hut on Inishkea North, Co. Mayo (Site 3, House A). JRSAI 82(2), 163–78.

Herity, M. 1995b Two island hermitages in the Atlantic: Rathlin O’Birne, Donegal and Caher Island, Mayo. JRSAI 125, 85–128. Herren, M.W. (ed.) 1974 The Hisperica famina: a new critical edition with English translation and philological commentary 1. Toronto. Herren, M. 1987 The Hisperica Famina II. Related Poems. Toronto. Higgins, J. 1987 Early Christian cross-slabs, pillar stones and related monuments in Co. Galway, Ireland. British Archaeological Reports: International Series 375, 2 vols. Oxford. Hill, I. 1992 The fish of Ireland. Belfast. Hill, P. (ed.) 1997 Whithorn and St. Ninian: the excavation of a monastic town, 1984-91. Stroud. Hillman, G.C. 1981 Reconstructing crop husbandry practices from charred remains of crops. In R. Mercer (ed.), Farming practice in British prehistory, 123–62. Edinburgh.

Herity, M. 1977 The High Island hermitage. Irish University Review 7, Spring, 52–69.

Hillman, G.C. 1984 Interpretation of archaeological plant remains: the application of ethnographic models from Turkey. In W. van Zeist and W.A. Casparie (eds), Plants and ancient man: studies in palaeoethnobotany, 1–41. Rotterdam.

Herity, M. 1987 The ornamented tomb of the saint at Ardoileán, Co. Galway. In M. Ryan (ed.), Ireland and insular art, A.D. 500-1200, 141–2. Dublin.

Hooton, E.A. and Dupertuis, C.A. 1955 The physical anthropology of Ireland. Papers of the Peabody Museum of Archaeology and Ethnology 30(1, 2).

Herbert, M. and McNamara, M. (eds) 1989 Irish biblical apocrypha: selected texts in translation. Edinburgh.

Herity, M. 1989 Cathair na Naomh and its cross-slabs. Cathair na Mart, JWHS 9 (1), 91–100. Herity, M. 1990a The hermitage on Ardoileán, County Galway. JRSAI 120, 65–101. Herity, M. 1990b, Carpet pages and chi-rhos: some depictions in Irish Early Christian manuscripts and stone carvings. Celtica 21, 208–22. Herity, M. 1993a Gleanncholmcille: A guide to 5,000 years of history in stone. Dublin. Herity, M. 1993b The forms of the tomb-shrine of the founder saint in Ireland. In R.M. Spearman and J. Higgitt (eds), The age of migrating ideas: early medieval art in Northern Britain and Ireland. Proceedings of the Second International Conference on Insular Art held in the National Museums of Scotland in Edinburgh, 3-6 January 1991, 185–95. Edinburgh. Herity, M. 1995a Studies in the layout, buildings and art in stones of early Irish monasteries. London.

Horn, W., Marshall, J.W. and Rourke, G. 1990 The forgotten hermitage of Skellig Michael. Los Angeles. Horne, L. 1972 Fuel for the metal worker. Expedition 25, Fall, 6–13. University of Pennsylvania Museum. Hornell, J. 1937 The curraghs of Ireland. Mariner’s Mirror 23, 1, 74–83. Hornell, J. 1938 The curraghs of Ireland. Mariner’s Mirror 24, 1, 5–39. Howlett, D. 2006 Dictionary of medieval Latin from British sources. Fascicule X. Oxford. Hughes, K. and Hamlin, A. 1977 (second edn 1997) The modern traveller to the Early Irish Church. Dublin. Iacumin, P., Bocherens, H., Maariotti, A. and Longinelli, A. 1996 Oxygen isotope analyses of co-existing carbonate and phosphate in biogenic apatite: a way to monitor diagenetic alteration of bone and phosphate? Earth and Planetary Science Letters 142, 1–6.

340 high island: excavation of an early medieval monastery

Ivens, R.J. 1989 Dunmisk fort, Carrickmore, Co. Tyrone, excavations 1984–1986. UJA 52, 17–110.

Lacy, B. 1983 Archaeological Survey of Co. Donegal. Lifford.

Jackson, R. 2006 Excavations at St James’s Priory, Bristol. Oxford.

Lancaster, S. 2005 Preliminary assessment of soil thin sections from High Island. Unpublished report.

James, M. 1924 The Apocryphal New Testament. Oxford.

Lancaster, S. 2006 Analysis of soil samples from High Island. Unpublished report.

Joyce, P.H. 1906 Letter dated 17 November 1906 from P.H. Joyce, Clerk of Oughterard Union and Rural District Council, to the Geological Survey of Ireland. Unpublished, Geological Survey of Ireland Office, Dublin.

Ladner, G. 1995 God, cosmos and humankind: the world of Early Christian symbolism. Berkeley.

Landes, R. 2000 The fear of an apocalyptic year 1000: Augustinian historiography, medieval and modern. Speculum 75, 97–114. Lane, A. and Campbell, E. 2000 Dunadd, an early Dalriadic capital. Oxford.

Katz, N.J., Katz, S.V. and Kipiani, M.G. 1965 Atlas and keys of fruits and seeds occurring in the quaternary deposits of the USSR. Moscow.

Lang, J. 2001 Corpus of Anglo-Saxon stone sculpture VI: Northern Yorkshire. Oxford.

Keepax, C.A. and Morgan, G. 1978 Charcoal. In S.A. Butcher, Excavations at Nornour, Isles of Scilly, 1969–73: the pre-Roman settlement. Cornish Archaeology 17, 29–112.

Lewis, S. 1837 A topographical history of Ireland. London.

Kelleher, J.V. 1971 Unpublished letter from J.V. Kelleher to R. Murphy. Kelly, D. 1988 Cross-carved slabs from Latteragh, Co. Tipperary. JRSAI 118, 92–100. Kelly, F. 1998 Early Irish Farming. Dublin. Kenney, J.F. 1929 The sources for the early history of Ireland: 1. Ecclesiastical (rev. edn 1966). New York. Keogh, J. 1735 Botanalogia universalis Hibernica or a general Irish herbal. Cork. Kermode, P. 1907 Manx crosses (1994 reprint). Balgavies. Kinahan, G.H. 1868–9 Report on the state of the ancient ruins on the islands off the western coast of Iar-Connaught. JRSAI 1 (third series), 348. Kinahan, G.H. 1869 The ruins on Ardillaun, Co. Galway. PRIA 10C 551–5. Kinahan, G.H., Nolan, J., Leonard, H. and Cruise, R.J. 1878 Explanatory memoir to accompany sheets 93 and 94, with the adjoining portions of sheets 83, 84 and 103, of the maps of the Geological Survey of Ireland illustrating the district around Clifden, Connemara. Geological Survey of Ireland, Dublin.

Leask, H.G. 1955 Irish churches and monastic buildings; Vol. 1 Early phases and the Romanesque. Dundalk. Lewis, S. 1980 Sacred calligraphy: the Chi-Rho page in the Book of Kells. Traditio 36, 139–59. Lionard, P. 1961 Early Irish grave-slabs. PRIA 61(C), 95–169. Little, L. 1993 Benedictine maledictions. Liturgical cursing in Romanesque France. Berkeley. Longinelli, A. 1984 Oxygen isotopes in mammal bone phosphate: a new tool for palaeohydrological and palaeoclimatological research? Geochimica et Cosmochimica Acta 48, 385–90. Lowden, J. 1997 Early Christian and Byzantine art. London. Lowe, C. 2008 Inchmarnock: an early historic island monastery and its archaeological landscape. Edinburgh. Lynn, C.J. 1976 Excavations at Bank of Ireland, Scotch St. Armagh. In I. Bennett (ed.), Excavations 1976, 24–5. Bray. Lynn, C. 1978 Early Christian period domestic structures: a change from round to rectangular plans? Irish Archaeological Research Forum 5, 29–45. Lynn, C. 1984 Some fragments of exotic porphyry found in Ireland. The Journal of Irish Archaeology 11, 19–32.

King, H. 1992 New graveyard at Clonmacnoise. In I. Bennet (ed.), Excavations 1992, 53–4. Bray.

Lyons, M.C. 1989 Weather, famine, pestilence and plague in Ireland 900–1500. In E.M. Crawford (ed.), Famine: the Irish experience 900–1900, 31–74. Edinburgh.

Kock, P., Tuross, N. and Fogel, M.L. 1997 The eﬀects of sample treatment and diagenesis on the isotopic integrity of carbonate in biogenic hydroxylapatite. Journal of Archaeological Science 24, 417–29.

Lysaght, P. 2000 Food preservation strategies on the Great Blasket Island: sea bird fowling. In P. Lysaght (ed.), Food from nature: attitudes, strategies and culinary practices, 33–63. Uppsala.

bibliography 341

Mac Airt, S. and Mac Niocaill, G. (eds) 1983 The annals of Ulster. Dublin.

McCarthy, M. In prep. Faunal remains from Caherlehillan, Co. Kerry.

Macalister, R.A.S. 1896 The antiquities of Ardilaun, County Galway. JRSAI 26, 197–210.

McCarthy, M. Unpublished. Faunal remains from Drumcliffe. Report lodged in National Monuments Service Archive, Department of Arts, Heritage and the Gaeltacht.

Macalister, R.A.S. 1909 The memorial slabs of Clonmacnoise, King’s Co. Dublin. Macalister, R.A.S. 1916 The history and antiquities of Inis Cealtra. PRIA 33, 93–174. Macalister, R.A.S. 1945–1949 Corpus inscriptionum insularum celticarum. 2 vols. Dublin. MacDonald, A. 1997 Adomnan’s monastery of Iona; a discussion of the monastery of Iona as it is reflected in Adomnan’s life of Columba (VC). In C. Bourke (ed.), Studies in the cult of Saint Columba, 24–44. Dublin. MacEoin, G. 1989 Orality and literacy in some Middle Irish king-tales. In H.L.C. Tristram, S.N. Tranter (eds), Early Irish literature: media and communication/Mündlichkeit und Schriftlichkeit in der frühen irischen Literatur, 149–83. Tübingen. MacGiolla Choille, B. and Simington, R.C. (eds) 1962 Books of survey and distribution Vol. III County Galway. Dublin. Maddern, C. 2013 Raising the dead: early medieval name stones in Northumbria. Turnhout. Mallory, J.P., Woodman, P., Voss, P., Van WijnagaardenBakker, L., McCormick, F. and Ross, H. 1984 Oughtymore: an Early Christian shell midden. UJA 47, third series, 51–62. Manning, C. 1988 Killelton Oratory. In I. Bennet (ed.) Excavations 1988, 21. Bray. Manning, C. 1995 Early Irish monasteries. Dublin. Manning, C. 1998a Clonmacnoise Cathedral. In H.A. King (ed.), Clonmacnoise Studies 1, Seminar Papers 1994, 57–86. Dublin. Manning, C. 2009 A suggested typology for the pre-Romanesque stone churches in Ireland. In N. Edwards (ed.), The archaeology of the early medieval Celtic churches, 265–79. Leeds. Marshall, J.W. and Walsh, C. 2005 Illaunloughan Island, an early medieval monastery in County Kerry. Bray. Martene, E. 1788 De Antiquis Ecclesiae Ritibus. 2 vols. Antwerp. McCarthy, M. 2012 Fish bone and sea mammal bone. In C. Cotter, The Western Stone Forts Project: excavations at Dún Aonghasa and Dún Eoghanachta, Vol. 2, 167–74. Dublin.

McClatchie, M. 2003 The plant remains. In R.M. Cleary and M.F. Hurley (eds), Cork city excavations 1984-2000, 391–413. Cork. McCormick, F. 1991 The effects of the Anglo-Norman settlement on Ireland’s wild and domesticated fauna. In P. Crabtree and K. Ryan (eds), Animal Use and Culture Change, Masca Research Papers in Science and Archaeology 8, 40–52. McCormick, F. 1997 Iona: the archaeology of the early monastery. In C. Bourke (ed.), Studies in the cult of Saint Columba, 45–68. Dublin. McCormick, F. and Murphy, E. 2012 Mammal bones. In C. Cotter, The Western Stone Forts Project: excavations at Dún Aonghasa and Dún Eoghanachta, Vol. 2, 153–66. Dublin. McCutcheon, S. 1997 The stone artefacts. In M. Hurley and O. Scully (eds), Late Viking Age and medieval Waterford, excavations 1986-1992, 404–32. Waterford. McErlean, T. and Crothers, N. 2007 Harnessing the tides: the early medieval tide mills at Nendrum monastery, Strangford Lough. Belfast. McKenzie, C. 2010 A palaeopathological study of the adult individuals from the Ballyhanna cemetery site, Co. Donegal. Belfast. McLaughlin, E. 1990 Did cursing stones really sink the HMS Wasp? Donegal Annual 1990 42, 17–24. Meehan, B. 1996 The Book of Durrow: a medieval masterpiece at Trinity College Dublin. Dublin. Mitchell, F. 1986 (1990 edn) Shell Guide to reading the Irish landscape. Dublin. Mitchell, F. 1987 Archaeology and environment in early Dublin. Dublin. Mitchell, F. 1995 The dynamics of Irish post-glacial forests. In J.R. Pilcher and S.S. Mac an tSaoir (eds), Woods, trees and forests of Ireland, 13–22. Dublin. Molloy, K., Fuller, J.L. and Conaghan, J. 2000 Vegetation and land-use on High Island: the results of preliminary investigations. In J.W. Marshall and G.D. Rourke, High Island: an Irish monastery in the Atlantic, Appendix 2, 233–43. Dublin. Moloney, M.F. 1919 Irish ethno-botany and the evolution of medicine in Ireland. Dublin.

342 high island: excavation of an early medieval monastery

Monk, M.A. 1987a Appendix IV: charred plant remains. In M.G. Doody, Moated site, Ballyveelish 1, Co. Tipperary. In R.M. Cleary, M.F. Hurley and E.A. Twohig (eds) 1987 Archaeological excavations on the Cork–Dublin Gas Pipeline, 74–87. Cork. Monk, M.A. 1987b Appendix II: charred plant remains. In M.F. Hurley, Kilferagh, Co. Kilkenny. In R.M. Cleary, M.F. Hurley and E.A. Twohig (eds) 1987 Archaeological excavations on the Cork–Dublin Gas Pipeline, 88–100. Cork. Monk, M.A. 1987c Appendix III: charred plant remains. In R.M. Cleary, Drumlummin, Co. Tipperary. In R.M. Cleary, M.F. Hurley and E.A. Twohig (eds) 1987 Archaeological excavations on the Cork–Dublin Gas Pipeline, 116–45. Cork. Monk, M.A. 1991 The archaeobotanical evidence for field crop plants in early historic Ireland. In J. Renfrew (ed.), New light on early farming: recent developments in palaeoethnobotany, 315–28. Edinburgh. Monk, M.A. and Kelleher, E. 2005 An assessment of the archaeological evidence for Irish corn-drying kilns in the light of results of archaeological experiments and archaeobotanical studies. Journal of Irish Archaeology 14, 77–114. Monk, M.A., Tierney, J. and Hannon, M. 1998 Archaeobotanical studies and early medieval Munster. In M.A. Monk and J. Sheehan (eds), Early medieval Munster: archaeology, history and society, 65–75. Cork. Morgan, R.A. 1982 Tree-ring studies in the Somerset Levels; the examination of modern hazel growth in Bradfield Woods, Suffolk, and its implications for the prehistoric data. Ancient Monuments Laboratory Report 3839. Murphy, C.P. 1986 Thin section preparation of soils and sediments. Berkhamsted. Murphy, M. and Potterton, M. 2005 Investigating living standards in medieval Dublin and its region. In S. Duffy (ed.), Medieval Dublin VI. Proceedings of the Friends of Medieval Dublin Symposium 2004, 224–56. Dublin. Murray, E.V. 2002 The mammal bones from Rathgureen ringfort. In M. Comber, M.V. Duignan’s excavations at the ringfort of Rathgurreen, Co. Galway, 1948–9. PRIA 102C, 137–97. Murray, E. 2011 The mammal bones. In E. Bourke, A.R. Hayden and A. Lynch, Skellig Michael, Co. Kerry: the monastery and South Peak, Archaeological stratigraphic report: excavations 1986-2010. (http://www.worldheritageireland.ie/skellig-michael/publications).

Murray, E. and McCormick F. 2005 Environmental analysis and the food supply. In J. White Marshall and C. Walsh, Illaunloughan Island: an early medieval monastery in County Kerry. Bray. Murray, E. and McCormick, F. 2012 Doonloughan: a seasonal settlement on the Connemara coast. PRIA 112C, 1–52. Murray, E., McCormick, F. and Plunkett, G. 2004 The food economies of Atlantic island monasteries: the documentary and archaeo-environmental evidence. Environmental Archaeology 9(2), 179–88. Murtagh, B. 1997 The architecture of St. Peter’s Church. In M.F. Hurley and O.M.B. Scully with S.W.J. McCutcheon, Late Viking Age and medieval Waterford Excavations 1986-1992, 228–43. Waterford. Mynors, R.A.B. 1956 The Stonyhurst Gospel. In C.F. Battiscombe (ed.), The relics of St Cuthbert, 356–74. Oxford. Nash-Williams, V. 1950 The Early Christian monuments of Wales. Cardiff. Nees, L. 1978 A fifth-century book cover and the origin of the four Evangelist symbols page in the Book of Durrow. Gesta 17, 3–8. Nees, L. 2002 Early medieval art. Oxford. Ní Shéaghdha, N. 1967 Catalogue of Irish manuscripts in the National Library of Ireland, vol. 1. Dublin. Norton, J. 2004 Clay pipes. In E. Fitzpatrick, M. O’Brien and P. Walsh (eds), Archaeological investigations in Galway City, 1987–1998, 427–47. Bray. O’Brien, C. 1994 New finds from Co. Offaly. Archaeology Ireland 8(1), 16–17. O’Brien, C. and Sweetman, P.D. 1997 Archaeological Inventory of Co. Offaly. Dublin. Ó Carrag�� á�� in, T. 2003a A landscape converted: archaeology and early church organisation on Iveragh and Dingle, Ireland. In M. Carver (ed.), The cross goes North: processes of conversion in Northern Europe AD 300–1300, 127–52. Suffolk. Ó Carragáin, T. 2003b The architectural setting of the cult of relics in early medieval Ireland. JRSAI 133, 130–76. Ó Carragáin, T. 2005 Habitual masonry styles and the local organization of church building in early medieval Ireland. PRIA 105C, 99–149.

bibliography 343

Ó Carragáin, T. 2010 Churches in early medieval Ireland. New Haven and London. Ó Carragáin, T., O’Sullivan, J. and Ó Caoimh, T. 2005 A flying visit to Bishop’s Island, Co. Clare. Archaeology Ireland 19 (1), 34–7. O’Connell, M. 1994 Connemara vegetation and land use since the last Ice Age. Dublin. Ó Corráin, D. 2009 The Vikings on Iveragh. In J. Crowley and J. Sheehan (eds), The Iveragh Peninsula: a cultural atlas of the Ring of Kerry, 141–7. Cork. Ó Corráin, D. and Maguire, F. 1981 Gaelic personal names. Dublin. Ó Cróinín, D. 1995 Early medieval Ireland 400–1200. Harlow. O’Donnabhain, B. 2010 The human remains. In A. Lynch, Tintern Abbey, Co Wexford: Cistercians and Colcloughs. Excavations 1982–2007. Archaeological Monograph Series 5, 105–25. Dublin. O’Donnabhain, B. (forthcoming) The human remains from the Wood Quay excavations, Dublin 1974–1981. O’Donnabhain, B, and Cosgrave, U. 1994 The bioarchaeology of the cemetery at Temple Lane, Dublin. Unpublished manuscript submitted to Margaret Gowen and Co. Ltd, April 1994. O’Donnchadha, B. 2007 The oldest church in Ireland’s ‘oldest town’. Archaeology Ireland 21 (1), 8–10. O’Donnell, L. 2011 Charcoal and wood. In E. Bourke, A.R. Hayden and A. Lynch, Skellig Michael, Co. Kerry: the monastery and South Peak. Archaeological stratigraphic report: excavations 1986–2010. (http:// www.worldheritageireland.ie/skellig-michael/publications). O’Flaherty, T. 1935 Cliffmen of the west. London.

O’Keeffe, T. 1992a Interim report on archaeological excavations on Omey Island, 92E0053, 1–5. Unpublished report lodged in National Monuments Service Archive, Department of Arts, Heritage and the Gaeltacht. O’Keeffe, T. 1992b Early Christian hermitage discovered on Omey Island. Archaeology Ireland 6 (3), 5. O’Keeffe, T. 1993 Omey Island – Gooreen and Sturakeen. In I. Bennett (ed.), Excavations 1993, 39. Bray. O’Keeffe, T. 1994a Lismore and Cashel: Reflections on the beginnings of Romanesque architecture in Munster. JRSAI 124, 118–51. O’Keeffe, T. 1994b Omey and the sands of time. Archaeology Ireland 8 (2), 14–17. O’Kelly, M.J. 1956 An island settlement at Beginish, Co. Kerry. PRIA 57C, 159–94. O’Kelly, M. 1958 Church Island, near Valencia Co. Kerry. PRIA 59C, 2, 57–136. O’Kelly, M.J. 1989 (1997 edn) Early Ireland. Cambridge. O’Loughlin, T. 2000 The diffusion of Adomnán’s De Locis Sanctis in the medieval period. Ériu 51, 93–106. Ó hÉailidhe, P. 1967 The crosses and slabs at St Berrihert’s Kyle in the Glen of Aherlow. In E. Rynne (ed.), Munster Studies: essays in commemoration of Monsignor Michael Maloney, 102–26. Limerick. O’Reilly, J. 1998 Patristic and insular traditions of the Evangelists: exegesis and iconography of the foursymbols page. In A.M. Luiselli Fadda and É. Ó Carragáin (eds), Le Isole Britanniche e Roma in Éta Romanobarbarica, 49–94. Rome. Ó Riain, P. 2003 Four Irish martyrologies: Drummond, Turin, Cashel, York. Henry Bradshaw Society 115. Woodbridge.

O’Flanagan, M. Rev. 1927 Letters containing information relative to the antiquities of the County of Galway collected during the progress of the Ordnance Survey in 1838. Typescript in 3 vols. Bray.

Ó Riain, P. 2011 A dictionary of Irish saints. Dublin.

Ó Floinn, R. 1995 Clonmacnoise: art and patronage in the early medieval period. In C. Bourke (ed.), From the isles of the North: early medieval art in Ireland and Britain, 251–60. Belfast.

Ó’Ríordáin, S.P. 1949 Lough Gur excavations: Carraig Aille and the Spectacles. PRIA 52C, 6–111.

Okasha, E. and Forsyth, K. 2001 Early Christian inscriptions of Munster: a corpus of the inscribed stones. Cork. O’Keeffe, T. 1990 Omey �� Island –�� Gooreen and Sturrakeen. In I. Bennett (ed.), Excavations 1990, 33. Bray. O’Keeffe, T. 1992 Omey �� Island – Gooreen and Sturrakeen. In I. Bennett (ed.), Excavations 1992, 30–31. Bray.

Ó Riain-Raedel, D. 2001 Reform and renewal – Ireland and Europe in the early twelfth century. Archaeology Ireland 15 (3), 25–7.

O’Rorke, T. 1878 History, antiquities and present state of the parishes of Ballysadare and Kilvarnet in the county of Sligo. Dublin. O’Sullivan, A., McCormick, F., Kerr, T. and Harney, L. 2008 Early medieval Ireland: archaeological excavations 1930–2004. Early Medieval Archaeology Project (EMAP) Report 2.1. (http://www.ucd.ie/t4cms/ emap_report_2.1_complete.pdf ).

344 high island: excavation of an early medieval monastery

O’Sullivan, A., McCormick, F., Harney, L., Kinsella, J. and Kerr, T. 2010 Early medieval dwellings and settlements in Ireland, AD400–1100. Vol. 1: Text. Dublin. O’Sullivan, A. and Sheehan, J. 1996 The Iveragh Peninsula: an archaeological survey of South Kerry. Cork. O’Sullivan, C.M. 2004 Hospitality in medieval Ireland, 900–1500. Dublin. O’Sullivan, J. 1994 Excavations of an early church and a women’s cemetery at St Ronan’s medieval parish church, Iona. PSAS 124, 327–65. O’Sullivan, J. and Ó Carragáin, T. 2008 Inishmurray: monks and pilgrims in an Atlantic landscape. Vol. 1: Archaeological Survey and Excavations 1997–2000. Cork. O’Sullivan, T. 2008 Bird and animal bones. In J. O’Sullivan and T. Ó Carragáin, Inishmurray: monks and pilgrims in an Atlantic landscape, 285–6. Cork. O’Sullivan, W. 1949 The earliest Irish coinage. JRSAI 79, 33–4. O’Sullivan, W. 1964 The earliest Anglo-Irish coinage. Dublin. Ó Tuathail, E. 1948–52 Léim Lára. Eigse 6, 155–6. Parsons, H.A. 1921/2 An Irish eleventh-century coin of the Southern O’Neill. British Numismatic Journal 16, 59–71. Pavía, S. and Bolton, J. 2001 Stone monuments decay study 2000: assessment of the degree of erosion and degradation of a sample of stone monuments in the Republic of Ireland. Kilkenny. Petrie, G. 1845 The ecclesiastical architecture of Ireland: an essay on the origin of Round Towers in Ireland. 2nd edn. Dublin.

Prummel, W. and Frisch, H.T. 1986 A guide for the distinction of species, sex and body side. Journal of Archaeological Science 13, 557–77. Rackham, O. 1976 (rev. 1990) Trees and woodland in the British landscape. London. Rackham, O. 1995 Looking for ancient woodland in Ireland. In J.R. Pilcher and S.M. Tsaori, Woods, trees and forests in Ireland, 1–12. Dublin. Radford, C. 1956 The portable altar of St Cuthbert. In C. Battiscombe (ed.), The relics of St Cuthbert, 326–35. Oxford. Raftery, J. 1960 A Viking burial in Co. Galway. JGAHS 29, 1, 3–6. Rahtz, P., Hirst, S. and Wright S.M. 2000 Cannington cemetery. Britannia Monograph Series 17. London. Royal Commission on the Ancient and Historical Monuments of Scotland 1982 4 Iona. Edinburgh. Royal Commission on the Ancient and Historical Monuments of Scotland 1984 5: Islay, Jura, Colonsay and Oronsay. Edinburgh. Redknap, M. and Lewis, J. 2007 Corpus of early medieval inscribed stones and stone sculpture in Wales. Vol.1: South-East Wales and the English Border. Cardiff. Resi, H.G. 1987 Reflections on Viking Age local trade in stone products. In J.E. Knirk (ed.), Proceedings of the Tenth Viking Congress, Larkollen, Norway, 1985. Oslo. Roberts, C. and Manchester, K. 2005. The archaeology of disease. New York. Robinson, T. 1990 Connemara, Part 1: Introduction and Gazetteer. Roundstone.

Plummer, C. (ed.) 1910 Vitae sanctorum Hiberniae. 2 vols. Oxford.

Robinson, T. 2008 Connemara, the last pool of darkness. Dublin.

Plummer, C. 1922 Bethada Náem nÉrenn=Lives of the Irish Saints. 2 vols. Oxford.

Roche, G. 1958 Appendix No. II: The zoological evidence. In M.J. O’Kelly, Church Island near Valencia, Co. Kerry. PRIA 59C, 1334–6.

Pochin Mould, D.D.C. 1955 Irish pilgrimage. Dublin. Pollard, A.M., Ditchfield, P., Piva, E., Wallis, S., Falys, C. and Ford, S. 2012 Sprouting like cockle amongst the wheat: the St Brice’s day massacre and the isotopic analysis of human bones from St John’s College, Oxford. Oxford Journal of Archaeology 31:1, 83–102.

Roe, H. 1960 A stone cross at Clogher, Co. Tyrone. JRSAI 90/2, 191–206.

Power, C. 1995 The burials. In M.F. Hurley and C.M. Sheehan (eds), Excavations at the Dominican Priory, St Mary’s of the Isle, Crosse’s Green, Cork, 56–83. Cork.

Rourke, G. 2009 Skellig Michael: monastic island retreat in the Atlantic. In J. Crowley and J. Sheehan (eds), The Iveragh Peninsula: a cultural atlas of the Ring of Kerry, 129–35. Cork.

Rosenthal, E. 1954 Pottery and ceramics. London. Ross, A. 2000 Folklore of the Scottish Highlands. Glouchestershire.

bibliography 345

Rowell, D.L. 1994 Soil science: methods and applications. London.

Sharpe, R. 1991 Medieval Irish saints’ lives: an introduction to Vitae sanctorum Hiberniae. Oxford.

Ryan, J. 1940 The abbatial succession at Clonmacnoise. In John Ryan (ed.), Féilsgríbhinn Eóin Mhic Néill, 490–507. Dublin. Repr. Dublin, 1995.

Sharpe, R. (ed.) 1995 Adomnán of Iona: Life of Columba. London.

Ryan, M. 2000 Furrows and browse: some archaeological thoughts on agriculture and population in early medieval Ireland. In A.P. Smyth (ed.), Seanchas: studies in early and medieval Irish archaeology, history and literature in honour of Francis J. Byrne, 30–36. Dublin. Rynne, C. 1998 The craft of the millwright in early medieval Munster. In M.A. Monk and J. Sheehan (eds), Early medieval Munster: archaeology, history and society, 87–101. Cork.

Sheehan, J. 1987–88 A reassessment of the Viking burial from Eyrephort, Co. Galway. JGHAS 41, 60–72. Sheehan, J. 1999 Caherlehillan, Co. Kerry. In I. Bennett (ed.), Excavations 1999, 112. Sheehan, J. 2009 A peacock’s tale: excavation at Caherlehillan, Iveragh, Ireland. In N. Edwards (ed.), The archaeology of the early medieval Celtic Churches. Society for Medieval Archaeology Monograph 29 / Society for Church Archaeology Monograph 1, 191–206. Leeds.

Rynne, C. 2000 The early medieval monastic watermill. In J.W. Marshall and G. Rourke, High Island, an Irish monastery in the Atlantic, 185–213. Dublin.

Sheehan, J., Stummann Hansen, S. and Ó Corráin, D. 2001 A Viking Age maritime haven: a reassessment of the island settlement at Beginish, Co. Kerry. Journal of Irish Archaeology 10, 93–119.

Salmon, W. 1710 Botanologia: the English herbal. London.

Silver, I.A. 1971 The ageing of domestic animals. In D.R. Brothwell and E. Higgs (eds), Science in Archaeology, a survey of progress and research. Bristol.

Savage, G. and Harold Newman, H. 1985 An illustrated dictionary of ceramics. London. Reprinted 2000. Scally, G. 1995–2002 Excavations at the monastic site, High Island, Co. Galway. Unpublished reports submitted to the National Monuments Service Archive, Department of Arts, Heritage and the Gaeltacht. Scannell, M.J.P. 1958 Appendix No. III: report on the botanical material. In M.J. O’Kelly, Church Island, near Valencia, Co. Kerry. PRIA 59C, 57–136. Schapiro, M. 1980 Late Antique, Early Christian and medieval art: selected papers. London. Schiller, G. 1972 Iconography of Christian art, Vol. 2: The Passion of Jesus Christ. London Scott, R.E. 2006 Social identity in early medieval Ireland: a bioarchaeology of the Early Christian cemetery on Omey Island, Co. Galway. A dissertation in Anthropology. Presented to the Faculties of the University of Pennsylvania in partial fulfilment of the requirements for the degree of Doctor of Philosophy. UMI No. 3246234. Seaby, P. 1970 Coins and tokens of Ireland. Part 3. London. Seaby, W.A. 1972 The present location of HibernoNorse coins. Seabys Coin and Medal Bulletin, June, 271. Sexton, R. 1998 Porridges, gruels and breads: the cereal foodstuffs of early medieval Ireland. In M.A. Monk and J. Sheehan (eds), Early medieval Munster: archaeology, history and society, 76–86. Cork.

Simms, M. 2001 Exploring the limestone landscapes of the Burren and the Gort lowlands: a guide for walkers, cyclists and motorists. Belfast. Smith, J.M.H. 1990 Oral and written: saints, miracles and relics in Brittany, circa 850–1250. Speculum 65, 309–43. Smith, P.J. 2004 Muirchertach (d. 1014/1016). In Oxford dictionary of national biography. Oxford. [http://www. oxforddnb.com/view/article/17681�� , accessed 2 February 2008] Smith, C.S. and Gnudi, M.T. 1990 The Pirotechnia of Vannoccio Biringuccio. New York. Stace, C. 1991 New flora of the British Isles. Cambridge. Stalmans, N. and Charles-Edwards, T.M. 2004 Meath, saints of (act. c.400–c.900). In Oxford dictionary of national biography. Oxford.[http://www.oxforddnb. com/view/article/51010, accessed 2 February 2008] Stevenson, J. 1989 The beginnings of literacy in Ireland. PRIA 89 (C), 127–65. Stevenson, R. 1956 The chronology and relationships of some Irish and Scottish crosses. JRSAI 86, 1, 84–96. Stokes, M. 1878 Early Christian architecture in Ireland, 2 vols. London. Stokes, P. and Rowley-Conway, P. 2002 Iron Age Cultigen? Experimental return rates for fat hen (Chenopodium album L.). Environmental Archaeology 7, 95–9.

346 high island: excavation of an early medieval monastery

Stokes, W. 1868 The life and labours in art and archaeology of George Petrie. London. Stokes, W. (ed.) 1890 Lives of saints from the Book of Lismore. Oxford.

Vogel, C. 1986 Medieval liturgy: an introduction to the sources. Revised and translated by W.G. Storey and N.K. Rasmussen. Washington.

Stokes, G.T. 1892 St Fechin of Fore and his monastery. JRSAI 22, 1–12.

Wakeman, W.F. 1862 Aran – pagan and Christian. Duffy’s Hibernian Magazine 1, 460–71, 567–77.

Stokes, W. (ed.) 1895 Félire hÚi Gormáin. The martyrology of Gorman. Henry Bradshaw Society 9. London.

Wakeman, W.F. 1863 An uninhabited Island. Duffy’s Hibernian Magazine 1V, 213–24.

Stokes, W. and Strachan, J. 1903–1910 Thesaurus palaeohibernicus: a collection of Old Irish glosses, scholia, prose and verse. 2 vols & supplement. Cambridge. Repr. Dublin 1975, 1987.

Wakeman, W.F. 1867 Aird Illawn, or High Island, Connemara. The Dublin Saturday Magazine 2 (83), 367–8.

Stokes, W. (ed.) 1905 Félire Óengusso Céli Dé. The martyrology of Óengus the Culdee. Henry Bradshaw Society 29. London. Stuiver, M. and Pearson, G.W. 1986 High-precision calibration of the radiocarbon timescale, AD 1950500 BC. Radiocarbon 28, 805–38. Stuiver, M. and Reimer, P.J. 1993 Extended 14C database and revised CALIB radiocarbon calibration program. Radiocarbon 35, 215–30. Swift, C. 1997 Ogam stones and the earliest Irish Christians. Maynooth. Synnott, D. 1979 Folk-lore, legend and Irish plants. In E.C. Nelson and A. Brady (eds), Irish gardening and horticulture. Dublin. Taylor, S. (forthcoming) St Vigeans in Pictish times. Taylor, S.R. and McLennan, S.M. 1981 The composition and evolution of the continental crust: rare earth element evidence from sedimentary rocks. Philosophical Transactions of the Royal Society A301, 381–99. Thomas, C. 1967 An Early Christian cemetery and chapel on Ardwall Isle, Kirkcudbright. Medieval Archaeology 11, 127–88. Thomas, C. 1971a Britain and Ireland in Early Christian Times AD 400-800. London. Thomas, C. 1971b The Early Christian archaeology of North Britain. London. Thomas, C. 1994 And Shall These Mute Stones Speak? Post-Roman inscriptions in Roman Britain. Cardiff. Trotter, M. 1970 Estimation of stature from intact long limb bones. In T.D. Stewart (ed.), Personal Identification in Mass Disasters, 71–83. Washington. Tutin, T.G., Heywood, V.H., Burges, N.A., Valentine, D.H., Walters, S.M. and Webb, D.A. 1964–80 Flora Europaea (Volumes 1–6). Cambridge.

Wakeman, W.F. 1885 Inis Muiredaich, now Inishmurray, and its antiquities. JRSAI 7, 175–332. Walsh, P. (ed.) 1918 Genealogiae regum et sanctorum Hiberniae. Dublin. Warren, F.E. 1881 The liturgy and ritual of the Celtic Church (2nd edn, J. Stevenson ed.). Woodbridge. Weber, R. (ed.) 1994 Biblia Sacra, iuxta Vulgatem Versionem. Stuttgart. Wells, C. 1981 Discussion of the skeletal material. In R. Reece (ed.), Excavations at Iona 1964–1974. Occasional Publications 5. London: Institute of Archaeology. Welsh, R. 1902 Rosguil and the old Kingdom of Fanad. JRSAI 32, 231. Werner, M. 1994 Crucifixi, Sepulti, Suscitati: remarks on the decoration of the Book of Kells. In F. O’Mahoney (ed.), The Book of Kells. Proceedings of a Conference at Trinity College, Dublin, September 1992, 609. Aldershot. Westropp, T.J. 1905 The coast and islands of County Galway. In Illustrated Guide to the Northern, Western and Southern islands and coast of Ireland. Royal Society of Antiquaries of Ireland, Antiquarian Handbook Series VI, 45–96. Westropp, T.J. 1911 A folklore survey of Co. Clare. The Transactions of the Folk-lore Society 22 (1), 203–12. Wheeler, A. 1978 Key to the fishes of Northern Europe. London. White, R.E. 1997 Principles and practice of soil science: the soil as a natural resource. Oxford. White Marshall, J. and Rourke, G. 2000 High Island, an Irish monastery in the Atlantic. Dublin. White Marshall, J. and Walsh, C. 2005 Illaunloughan Island: an early medieval monastery in County Kerry. Bray. Williams, B.B. 1978 Excavations at Lough Eskragh, Co. Tyrone. Ulster Journal of Archaeology 41, 37–48.

bibliography 347

Woolf, A. 2007 From Pictland to Alba 789–1070. The New Edinburgh History of Scotland 2. Edinburgh. Wright, D. 1982 The Irish element in the formation of Hiberno-Saxon art: calligraphy and metalwork. In H. Löwe (heraus.), Die Iren und Europa im früheren Mittelalter, teil. 1, 99–100. Stuttgart. Young, T.P. 2005 Evaluation of archaeometallurgical residues from the Heath-Mayfield N7 development (03E0151, 03E0966, 03E0461, 03E0603, 03E0633, 03E0679, 03E0602, 03E0635). GeoArch Report 2005/12. 28pp. Unpublished report submitted to the National Monuments Service Archive, Department of Arts, Heritage and the Gaeltacht. Young, T.P. 2006 Evaluation of archaeometallurgical residues from N30 Moneytucker–Jamestown, sites 1, 4, 5 and 7 (04E0329, 04E0326, 04E0325, 04E0323). GeoArch Report 2005/13. 20pp. Unpublished report submitted to the National Monuments Service Archive, Department of Arts, Heritage and the Gaeltacht.