Calanais Survey and Excavation, 1979-88 (part 3) by Historic Environment Scotland

Calanais Survey and Excavation, 1979-88

20. The Vegetation Survey 20.1 Introduction Alan Fairweather and Marion Wade-Evans carried out the survey. The text here is largely that produced by them (Technical Note 20.1). A brief overall description of the vegetation was produced in 1980. The main survey was undertaken in 1981. A subjective overall view of the site was taken. In all about 20 areas with different communities from their neighbours were discernible (Illus 20.1). Species occurrences and their relative frequencies were recorded. In the time available it was not possible to write a detailed comparative description of each of the 20 areas discerned but a general account of broad areas are given. 50 species of flowering plants were recorded as growing on the site (Table 19.1). Some species in addition to those identified were not seen at full expression complete with inflorescence so specific attribution could not be made. Gramineae identifications were aided by Hubbard, (1954) and flora by Keble Martin (1965); the most important aid to plant community identifications was McVean & Ratcliffe (1962). 20.2 Factors affecting the vegetation Areas tended to merge from one community to the other and the differences were rarely clear cut. They owed more to varying abundance of dominant species than to the existence of totally different communities. Most species were present in most areas and those that were infrequent or rare within the survey area as a whole did not form a very significant part of any one community. (Exceptions to species absence versus dominance were Calluna vulgaris, Ciprus Caespitosus, Juncus squarrosis, and Juncus effusis). 20.2.1 Previous agricultural use. The mixed nature of the vegetation covering the area investigated and the comparatively rich nutrient status is in large part attributable to agriculture. The making and cultivating of cultivation

beds probably involved the incorporation of seaweed, dung and shell sand (most probably at the same time) and not only improved nutrient status but changed local topography. 20.2.2 Peat stripping Before cultivation (the visible remains of which probably belong in the 18th and early 19th centuries, peat was stripped from the site and the activities involved in this operation, including soil disturbance, trampling, soil peat mixing and exposure have contributed to the current vegetation mosaic. 20.2.3 Grazing The grazing regime has had a considerable affect in that most of the site shows evidence of having been closely grazed by sheep until shortly before 1980 with a very high and ubiquitous incidence of Plantago lanceolata, Hypochaeris radicata, Bellis perennis and an almost complete absence of Polygala spp.(Milkwort). It is not detrimental to the maintenance of a close sward to allow grazing to continue; indeed now that sheep have been excluded the site could soon revert to a coarser rank growth of tussocks requiring mowing and reseeding etc. 20.2.4 Victorian and later repairs Conservation of the archaeology involved the introduction of turf composed of Festuca species, Poa Plantago maritima and other species associated with a turf from a seaside or machair. This was especially noticeable on the cairn itself. 20.2.5 Disturbance by visitors Visitor presence and pressure has caused a good deal of soil compaction and wear on the site round the stones and the cairn and on both formal (Lady Mathesonâ&#x20AC;&#x2122;s Path) and informal paths between entrances and points of interest. This was shown by the dominance of Plantago lanceolata and Poa species, with other trampling-resistant species present. The vegetation survey \ 815

Calanais Survey and Excavation, 1979-88

In addition visitor pressure has led to some soil disturbance (puddling) and impeded surface drainage with localised peripheral effects. Close to paths an expected enrichment of the vegetation was observed, probably because of nutrient import on footwear and more rapid freeing of nutrients in solution by surface disturbance. This effect is most noticeable on vegetation growing on a base-poor soil as that at Calanais. It is not so easily seen on richer swards. 20.2.6 Archaeological investigations These have caused very significant effects. The most obvious is deturfing (1980 excavation and current) and heavy trampling and sward cover by spoil tips etc. The long term effects of this disturbance will probably favour a better growth and richer vegetation mat as soil and subsoil and peat are mixed in back fill, and the replacement of turf will physically change the natural succession of colonisation of returfed areas favouring finer and lower growing species. Adequate drainage and restriction of trampling will help in this respect. 20.3 The vegetation The vegetation areas are shown on Illus 20.1. They are described from south to north. 20.3.1 Area 16A The south end of the site, bounded by a dry stone wall, was found to have a flora reflecting a rich nutrient status as shown by the presence of Dactylus glomerata, Lolium perenne, and Festuca pratensis, and a lush growth of other species and a notable absence of Scirpus and Calluna. This was probably because of the enrichment of the soil by leached material from the wall surface and former use of the wall as shelter by sheep and consequent enrichment from faeces and urine. 20.3.2 Area 16 On the west side of the site above test pit G1 and G2 an area of early modern cultivation beds

showed a good cover of species indicating a relatively high nutrient status (but not as high as 16A) The dominant grass was Anthoxanthum odoratum with a scatter of Angelica sylvestris and Rumex acetosa whilst Agrostis app were well grown. Area 17 and 17A to 17D The early modern cultivation beds showed a comparatively high nutrient status reflected by a good growth of Anthoxanthum and Agrostis with Molinia present but not in tussock form. The richness of growth became more obvious as the beds went southwards, (for instance 17C was richer than 17A). 17a was more acid as shown by a relict patch of Calluna vulgaris and by the dry poor appearance of species growing in shallow soil. 20.3.3 Area 18 The path running northwards from the S end was readily distinguishable and showed an expected dominance of Plantago lanceolata and P maritima, both species resistant to trampling. The latter species may reflect turfs imported from a seaside or maritime environment for repairs. 20.3.4 Area 4 This area to the southwest of the Ring was in marked contrast to early modern cultivation beds, paths and the Ring. It was an example of a relict moorland community under heavy trampling and grazing pressure (surviving in area with increasing fertility?). It had typical heath mosaic associations with Calluna vulgaris (very poorly grown and worn down) Scirpus caespitosis and Juncus squarrosis tussocks with detritus mats of undecayed vegetation between them. Few other species were present, and those like Potentilla erecta confirmed an acidophilous community. 20.3.5 Area 14 This area to the northwest of the Ring was very similar to 4 but Calluna faded out and there were more Scirpus tussocks because these are more resistant to heavy trampling than Calluna. The vegetation survey \ 816

Calanais Survey and Excavation, 1979-88

20.3.6 Area 19

20.3.11 Area 12

The area of the stone circle was heavily trampled. By each stone leached material from their faces had enriched the ground at their bases where a richer growth and less acidophilous species were found e.g. Poa, Festuca etc. Within the circle the cairn had been covered in turf from machair or near the sea. This was much worn and trampled.

Area 12, in the middle of the Avenue to the west of Area 11 was dominated by Molinia caerulea (ubiquitous throughout the site but dominant here although not in tussocks) and Nardus stricta. There were fewer tussocks of Scirpus.

20.3.7 Area 13 The returfed 1980 Area F trenches showed poor recovery because of surface puddling and heavy traffic. They were poorly vegetated. 20.3.8 Area 15 The slope from the plateau to the western bluff had a flora typical of shallow soil acid moorland with Calluna, Eriophorum angustifolium and Festuca vivipira (the presence of E angustifolium is somewhat anomalous as although it is very acid-tolerant it is usually found in very wet situations in peat cuttings). The area showed a sharp and dramatic transition to deeper soil and a greener community including Agrostis and Anthoxanthum. 20.3.9 Area 20 This area lay to the north of Area 15. It was heavily trampled with frequent patches of Juncus squarrosus and Scirpus, with Juncus effusis in evidence. It was an area of impeded drainage with an acidophilous flora. 20.3.10 Area 11 This area lay on early modern cultivation beds between the east entrance to the site and the Avenue and Ring. It was much affected by trampling and broken up by paths. It was dominated by Juncus squarrosus patches and Scirpus tussocks increased in frequency. There does not seem to have been much distinction between the parts of this area identified topographically as probable cultivation beds and the area immediately to their south.

Area 3 This area surrounded the test pits G1 and G2 and was approximately 10m in radius. It was extremely diverse within a comparatively small area and for this reason was divided into several facies (zones or communities). 20.3.12 Area 3A This area lay just inside the fence. It bore a good lush growth of bent grass (Agrostis tenuis) with Yorkshire fog and creeping soft grass (Holcus lanatus and Homollis) with common rush ( Juncus conglomeratus). Beside this dominant area were frequent tussocks of Molinia caerulea in the damp area. This abundance, with the moss Thuidium tameriscinum, indicated a damp acid area with a moving but high water table rarely drying out. Other species such as Viola riviniana (violet), Hieracium species and Leontodon and Centaurea nigra showed a reasonable high nutrient status (confirmed by the presence of Holcus species). This area with 3B may be described as a â&#x20AC;&#x153;flushâ&#x20AC;?, in that nutrients draining from higher areas became relatively concentrated or were regularly replenished by leached material. 20.3.13 Area 3B The boggy depressions in the area supported a flora of impeded drainage and even stagnation (c.f. mosses of impeded drainage in surrounding hill land). Prime among these were sphagnum hummocks (S papillosum?) which in time may dominate and raise vegetation level above nutrient supplying ground water making a base poor ombrogenous bog. Within this area however, Viola, Festuca and Juncus conglomerata indicated a reasonable nutrient supply. The vegetation survey \ 817

Calanais Survey and Excavation, 1979-88

Illus 20.1Â The vegetation areas The vegetation survey \ 818

Calanais Survey and Excavation, 1979-88

20.3.14 Area 3C The edge of the cultivation bed within the fence was similar to Area 3A but more dominated by Molinia with Scirpus tussocks. Other species represented were Agrostis and Hieracia sp. The impression is of a Molinia facies representing a high but moving water table with an inflow of nutrients. Scirpus tussocks were an apparent relict of the former acid peat community generally represented throughout the site. 20.3.15 Area 3D This area was a flat cultivation bed surface near test pit G2 dominated by Molinia with abundant Agrostis tenuis and a high incidence of Plantago lanceolata with a mix of occasional Senecio jacobea, Angelica sylvestris, Succisa pratensis and Hieracea species, indicating improved nutrient status. An abundance of Thuidium tamariscinum, and other mosses showed a high water table and relatively acid conditions but the other species mentioned above showed relative nutrient availability. 20.3.16 Area 3E The â&#x20AC;&#x2DC;ditchâ&#x20AC;&#x2122; at the edge of the cultivation bed showed an expected flora indicating a high water table and poor base status but with some influx of nutrients. It included frequent Thuidium tamariscinum moss and other moss species encountered in acid moorland conditions. Sphagnum species hummocks wee initiating in the ditch with Molinia and Juncus conglomerata, indicating very restricted water flow and a high water table. The slopes of the ditch had small plants of Calluna vulgaris with Festuca rubra and Deschampsia flexuosa all in all demonstrating an acidophilous association. 20.3.17 Area 3F Numerous species were represented in this area (see Table 19.1). The variety reflected the topography of the area and thus the variable edaphic conditions mainly because of water table variations and nutrient flow concentrations and/or imped-

ance. The overall picture was of acid ground water flow and more acid free-draining high points. The slow or impeded drainage flow was in the lower areas and allowed a build up of nutrients with a greater number of species existing in close proximity, although not really forming a community per se, more a mosaic of communities. 20.3.18 Area 3G This area was much more trampled and bore a mat of undecayed vegetation, indicating low biological activity. The mass of the mat was made of Agrostis tenuis through which Holcus mollis and H lanatus occurred, indicating previous close grazing and a reasonably high nutrient status. In addition Plantago lanceolata was present confirming a sward kept very short by grazing or trampling. 20.3.19 Area 3H This area lay outside the fence on a raised early modern cultivation bed. It had a closely grazed sward with distinct composition including much more Holcus mollis, Anthoxonthum odoratum, Plantago lanceolata and Bellis perennis. The species present indicated a drier richer area but still acid and obviously very closely grazed. 20.3.20 Area 3I This area lay on the edge of a cultivation bed. It was characterised by Polytrichum (moss) species with some dark bare ground sparsely colonised by Hypochoeris radicata (Catsear) and P lanceolata. This was a very acid surface unmodified by vegetational colonisation. 20.3.21 Area 3J The flora of this area was dominated by Molinia tussocks raised above the high water level and supporting species of drier habitat within and on top of the tussocks, for example Festuca rubra and Potentilla erecta. Again this was an acid area but reasonably rich with some presence of Anthoxanthum odoratum.

The vegetation survey \ 819

Calanais Survey and Excavation, 1979-88

20.3.22 Area 3K The association in this area was dominated by Juncus effusis tussocks and there were fewer grass species and much more moss indicative of a high water table. The area below the rocky bluff above the test pit was very rich in comparison to the other areas. There was a lush growth of Rumex acetosa, Festuca and Agrostis with Viola species growing in rock clefts. This indicated a nutrient flush zone. This area was often used by male diggers in 1980 and 1981 to urinate and the growth observed was in large part because of this. 20.3.23 Area 2 This area was a raised green area by Stone 19 at the north end of the Avenue. It had a freer draining soil with higher nutrient status, indicated by the greater presence of Trifolium repens and its effect (nitrogen fixation) on other plants in this area. The south-facing base of the stone was dry in appearance compared with the north side. This was an example of a phenomenon typical of all stone base areas not trampled or disturbed by excavation. The stone face intercepts more rain and the consequent run-off leaches small but significant amounts of nutrients down into the soil at the stones base. This enrichment effect is most

apparent in soils that are nutrient deficient. The effect is speeded up by the absence of lichens (a similar and more striking effect is noticed at the base of the north-facing wall at the south end of the monument), slow water flow on the stone surfaces and thus slow washing effecting the abundance and dominance of rosette-forming and low growing species. Bellis perennis, Hieracia, clover and P lanceolata indicated the pressures keeping the sward short i.e. grazing and/or trampling. 20.4 Discussion The whole site could be described as a Molinietum, i.e. a community consistently containing an abundance of Molinia caerulea although not always forming tussocks. This was probably because of frequent disturbance and heavy grazing and trampling pressure. The communities are best described as a superimposition on the Molietum. Molinia itself is indicative in general of a high but moving water table on acid land. Areas 4, 14 and acid 15 (described above) with Calluna, Molinia, Scirpus, and J squarrosus forming a mosaic gave the nearest approximation to the probable site vegetation before peat cutting cleared it (compare Calluna etc. land in surrounding district).

The vegetation survey \ 820

Calanais Survey and Excavation, 1979-88

20.5 Calanais species list Latin name

Common name

Latin name

Common name

Agrostis canina

Bent grass

Luzula campestris

Woodrushes

Agrostis tenuis

Angelica sylvestris Anthoxanthum odoratum

Bent grass Angelica

Sweet vernal grass

Bellis perennis

Daisy

Carex species

Sedges

Calluna vulgaris Centaurea nigra

Cynosurus cristatus Dactylis glomerata Eriophorum angustifolium

Cardomine pratensis

Heather Knapweed, hard heads

Crested dogâ&#x20AC;&#x2122;s tail grass Cocks foot grass Bog cotton Cuckoo flower

Dactylorhiza maculata Heath orchid

Deschampsia flexuosa Wavy hair grass Festuca rubra

Red fescue

Festuca pratensis

Meadow fescue

Festuca vivipira

Hieracium species Holeus mollis

Holcus lanatus Juncus effuses

Viviparous fescue Hawk bits

Creeping soft grass Yorkshire fog grass Soft rush

Luzula multiflora

Woodrushes

Lolium perenne

Perennial rye grass

Molinia caerulea

Purple moor grass

Nardus stricta

Moor mat grass

Pedicularis sylvatica

Louse wart

Plantago lanceolata

Ribwort plantain

Plantago maritima

See plantain

Poa annua

Annual meadow grass

Poa pratensis

Meadow grass

Poa trivialis

Meadow grass

Potentilla erecta

Tormentil

Potentilla anserina

Pig weed, silver weed

Polygala vulgaris

Milkwort

Ranunculus acris

Buttercup

Ranunculus repens

Buttercup

Senecio jacobea

Ragwort

Succisa pratensis

Revels bit scabious

Viola riviniana

Violet

Scirpus caespitosus (Trichophorum cespitosum)

Deer grass

Trifolium repens

White clover

Juncus conglomeratus Common rush

Thudium tamariscinum

Leontodon autumnalis Autumn hawk bit

Polytrichum species

Moss

Juncus squarrosis

Health rush

Sphagnum species

Sphagnum moss

Luzula pilulifera

Woodrushes

Athyrium felix-femina Lady fern

Moss

The vegetation survey \ 821

Calanais Survey and Excavation, 1979-88

21. Palaeoenvironment S J P Bohncke, P J Ashmore and R Tipping

Illus 21.1 The main stone setting and the peninsula of Callanish Leobag 21.1 Introduction 21.1.1 Introductory note by P J Ashmore A full draft report on the Calanais palaeoenvironment was prepared by Dr S Bohncke in 1982. My long delay in producing the Calanais excavation report meant that it became impossible for him to revise his draft. Dr Tipping has added some discussion in collaboration with Dr Bohncke. I have added dating information, illustrations showing where spot samples were taken, archaeological context and discussion. 21.1.2 Research into the prehistoric environment of the Calanais area Much exploration of the past environment around Calanais has been undertaken over the past few decades (see Chapter 21.3.14). In the Callanish Leobag project three pollen columns were obtained within a few metres of each other from blanket peat (Bohncke 1988, fig. 2). The pollen source areas, the area from which most pollen comes, is small, probably only tens to a few hundreds of metres around. CN1 and CN3 share the same pollen source area. Columns CN1 and CN3 were radiocarbon-dated in detail (Table 21.1;

Bohncke 1988). The pollen contents of CN1 and CN3 at time intervals relevant to this project were re-analysed by Dr Bohncke for this report (Section 21.3). 21.1.3 On-site spot samples from Calanais 57 spot samples taken during the 1980 and 1981 excavations at Calanais were analysed to help create a refined bio-stratigraphic dating of the features discovered and to resolve correlation problems, to assess ‘turf lines’ and plough-turned layers and to throw light on the origin of the material in some layers. Some of the turf lines discovered then were sampled in Kubiena boxes taken in 1981 (Areas E and H) and in partially re-opened excavation trenches in 1982 (Area D). 21.1.4 Treatment of the pollen samples The sub-samples taken from the peat columns CN3 and CN1 at Callanish Leobag were treated with KOH and subsequently acetolysed according to Faegri and Iversen (1975). The samples taken from the Kubiena boxes and the spot samples from the excavation site were treated in the same way; but for the removal of sand and clay particles a separation in heavy liquid (bromoform-alcohol Palaeoenvironment \ 822

Calanais Survey and Excavation, 1979-88

Illus 21.2Â The peninsula of Callanish Leobag from the main stone setting at Calanais [Cal 1981 Large Format 8] mixture, specific gravity 2) followed by a night in 40% HF was added to the preparation method. The residue was mounted on a slide embedded in glycerine jelly and sealed with paraffin wax. The spot samples from the 1980 and 1981 excavations at Calanais had suffered from desiccation. This may have affected the condition of the pollen. The spot samples and samples from Kubiena boxes sometimes gave problems in this respect due to the poorer preservation of the pollen. Some pollen appeared to have shrunk, which made the identification of pollen of the Cerealia type difficult. 21.1.5 The pollen diagrams and tables The pollen sum used for the construction of pollen-diagrams and tables is of total land pollen (%tlp), excluding pollen of aquatic plants ands spores of mosses and ferns. A pollen sum of at least 500 was sought, to allow confidence that species occurring in low frequencies could still be recorded.

21.2 Dating 21.2.1 Introduction C14 assays are from 1.0cm or 2.0cm thick peat slices except GU-1234, from a 6-7cm thick chunk. Mid-depths are listed in Table 21.1. Bohncke (1988) reported three uncalibrated C14 assays from peat for the 64cm thick column of CN1 (samples GU-1170, -1171 and 1095). Two additional assays are reported in Table 21.1. Bohncke (1988) lists nine assays from peat at CN3 and two from Betula (birch) wood. In Table 21.1 these assays (GU-1150 and -1151) are regarded as separate, from a small distance away, dating the late stage of a Betula carr. In Table 21.1 the errors quoted by the laboratory for the first eight assays and the last have been corrected by multiplying by 1.4 and then increasing the result to +/-110 if less (Ashmore et al 2000). The errors attached to ages in the main sequence from pollen profile CN3 did not Palaeoenvironment \ 823

Calanais Survey and Excavation, 1979-88

21.3.2 The radiocarbon ages Table 21.1Â The radiocarbon ages Description Betula wood from a late stage of natural carr formation Betula wood from a late stage of natural carr formation CN1: peat from 18.0cm below the top peat from 19cm below the top peat from 48.0cm below the top peat from 49.0cm below the top peat from 64.0cm below the top, at the base CN2: peat from 30cm below the top CN3: peat from 10cm below the top peat from 15cm below the top peat from 19cm below the top peat from 23cm below the top peat from 27cm below the top peat from 30cm below the top peat from 33cm below the top peat from 37cm below the top peat from 53.5cm below the top and 10cm above the base

Lab Code

Lab Age

Lab Error

Corr Error

5110

126

5180

126

2440

112

2355

110

2960

133

3220

110

4810

110

255

110

3010

3430

3890

4005

4225

4525

4860

GU1993

5035

GU1234

7270

100

GU1150

GU1151

GU1289

GU1170

GU1290

GU1171

GU1095

GU1235

GU1986

GU1987

GU1988

GU1989

GU1990

GU1991

GU1992

140

Palaeoenvironment \ 824

Calanais Survey and Excavation, 1979-88

require correction. Calibrated dates are listed in Table 21.2. 21.2.3 Calibration and dating of the pollen diagrams from CN1 and CN3 The dates were calibrated using OxCal 3.10 (Bronk Ramsay 1995, 2005) and the 2004 calibration curve. Although the latter has been superseded the changes are fairly insignificant in the periods discussed here, particularly given the size of the errors attached to the dates. Pollen zone and sub-zone boundaries are defined in detail in Chapter 21.3. In Table 21.2 the combined date for the boundary between subzones CaN3b and 3c, seen in both the CN1 and CN3 cores, was calculated using OxCal 3.10. The dates initially showed poor agreement (A=55.6%; Ac = 60%). After further calculation marginal agreement was reached (A=66.1%; Ac = 60%) for the calculated date shown in Table 21.2. With one exception interpolated dates are based on adjacent dates in CN1 or in CN3 rather than the longer trend lines of age against peat depth. Thus each interpolation depends heavily on a single pair of dates which suffer all the imprecision inherent in single radiocarbon dates. But using more than the closest dates was rejected because it is likely that peat thickness per century varied within each column and therefore using longer trend lines would introduce new errors. The exception is the transition between subzone CaN-3ai and CaN-3aii for which two methods were used. In Method 1 adjacent dates were used but that made the result very dependent on GU1989; in Method 2 the calculation was based on GU-1988 and the interpolated date of the boundary between CN3 subzone 2d and CN3 3a; which in turn was based on GU-1989 and GU-1990. The peat in the earlier part of column CN3 seems to have grown more slowly than that above about 37cm (Illus 21.3). Given the lack of check dates between 37cm and 53.5cm the interpolated dates and those projected back to the start of the column, based as they are on an assumption of even growth between the 37cm and 60cm levels (CN3 subzones 2a to 2b and 1 to 2a and the start

of CN3 zone 1), must be regarded as unreliable. If that problem is ignored they imply that peat started to grow between two and three millennia after the end of the Younger Dryas (taken here as 9600 cal BC). Table 21.3 Dates with severe interpretational problems, omitted from Table 21.2a

Description

early late 2 σ 2 σ cal BC cal BC

Peat CN/IL 3950 from base of 3b GU- 3350 cal cal peat profile CN1 1095 BC BC 65cm from top CN3/119

1470 GU- 1960 cal cal 1235 AD AD

Two assays have been left out of the analysis to date. Assay GU-1095 from the base of the CN1 column was much older than expected given the other dating information and the depth from which it was obtained (Illus 21.3). The sample may have been derived from soil that contained residual organic carbon, or it may represent early shallow local peat formation in an area of impeded drainage with, in effect, a discontinuity before the start of continuous peat growth. It might be suggested as an alternative that the basal peat grew very slowly but there is no suggestion of that at the nearby CN3 column in this period: assay GU1095 has been ignored in what follows. Assay GU-1235 (225+/65 BP) probably reflects post-medieval peat growth after cutting. This assay has also been ignored. Illus 21.3 shows the preferred solution to matching the dating information from the CN1 and CN3 peat columns. It comprises two elements. 1. The coloured rectangles represent the 1 and 2 sigma (σ) C14 age errors (vertical) and date errors (horizontal); they are shown on the calibration curve and the true chronological scale along the bottom applies to them.The green rectangles are CN1 column dates and the orange ones are CN3 column dates. Palaeoenvironment \ 825

Calanais Survey and Excavation, 1979-88

Table 21.2 The one and two sigma dates for the radiocarbon ages and (in red) those calculated for zone and sub-zone boundaries

Description

Pollen CN1: peat CN3: peat Code sub-zones depth depth

early late 2 σ cal late 1 σ cal 1 σ cal BC BC BC

Projected

3e-4

-86

Peat

Interpolated

early 2σ cal BC

516

564

GU-1289 200

400

760

850

GU-1170 150

200

750

800

3d-3e

262

340

859

921

Interpolated

3c-3d

424

519

993

1072

Peat

GU-1986 1110

1130

1380

1410

Peat

GU-1290 800

1010

1380

1500

Peat

3b-c

GU-1171 1200

1380

1640

1800

Calculated

3b-3c

1520

1750

1880

Peat

3b-c

GU-1987 1600

1660

1880

1900

Peat

3a-b

GU-1988 2200

2290

2470

2560

Interpolated Method 1

3ai-3aii

22.5

2290

2445

2605

2865

Interpolated Method 2

3ai-3aii

22.5

2355

2455

2655

2770

Peat

GU-1989 2300

2460

2620

2900

Interpolated

2d-3a

2510

2620

2840

2985

Peat

GU-1990 2580

2670

2910

3010

Peat

2c-d

GU-1991 3020

3100

3360

3490

Peat

GU-1992 3510

3530

3710

3780

Interpolated

2c cereal

3605

3650

3830

3875

Peat

GU-1993 3700

3770

3950

3970

Interpolated

2b-c

3830

3905

4095

4120

Interpolated

2a-b

5525

5660

5980

6075

Interpolated

sharp Betula fall

5655

5795

6215

6225

Peat

53.5

GU-1234 5850

6000

6340

6450

Projected

1-2a

6045

6205

6555

6675

Projected

charcoal appears

6305

6475

6845

6975

Projected

start

6695

6880

7280

7425

Palaeoenvironment \ 826

Calanais Survey and Excavation, 1979-88

2. Thin blue vertical bars for the 2 Ď&#x192; ages of peat samples in CN1, arranged horizontally by peat depths (top green scale) and thin red vertical bars for CN3 ages (top pink scale). The age-bars are connected by straight lines and the areas between them coloured grey-green for CN1 and magenta for CN3. In constructing the diagram the horizontal positions of the blue and red age-bars were defined by the peat depth scales and the vertical positions fixed by the age scale (horizontal grid lines). The main constraint was that the age-bars had to fit within the corresponding coloured rectangles, which were in fixed date and age positions defined by the grid lines. It is clear that the peat in CN1 was on average about three times as thick per century as the peat in CN3 above 37cm, perhaps because it had grown three times as fast or possibly because it was less compressed, or both. Between c. 6000 BC and

4000 BC the peat in CN3 grew on average more slowly than in the next 3500 years; as discussed below there may have been another variation in peat growth near the top of CN3. With the (at best approximately right) assumption of even peat growth in most of CN1 and CN3 several subtly different matches were possible between the red and blue age bars and the corresponding sets of coloured rectangles representing C14 ages. The matches must be regarded as indicative rather than correct. Nevertheless, there are some severe constraints in making them. Illus 21.3a shows the preferred compromise from a C14 dating point of view between dating and zoning information; but it does conceal a problem in matching CN1 and CN3. Correlation on palynological grounds indicates that the 15cm level in CN3 should correspond to the 50cm level in CN1. Assuming slower peat growth in CN1 and in the upper 20cm of CN3 could remove that problem (Illus 21.3b). But if correct the C14 date GU-

llus 21.3aÂ The dates from CN1 and CN3 Palaeoenvironment \ 827

Calanais Survey and Excavation, 1979-88

1290 was too young by over 2 sigma and GU-1289 and GU-1170 were both rather ‘young’, although both just within the quoted 2 sigma limits. Illus 21.4 shows the two pollen columns with CN1 and the part of CN3 between about 10cm and 37cm at approximately the same timescale as each other (assuming even peat growth in each column during the period covered). The two red lines, at 15cm on CN3 and at 50cm on CN1 represent the boundary between CN3 subzones 3b and 3c, used to correlate the two columns in the pollen analysis. This match is different from those shown in Illus 21.3a and 21.3b because it is based purely on a palynological assessment of human interference in the landscape. 21.2.4 Dating of the Betula carr Two dates were obtained for Betula carr (Tables 2.1, 21.2). They suggest that the trees were grow-

ing at some date between 4350 and 3700 cal BC, towards the end of CaN- subzone 2b, when there was birch regeneration (below). 21.2.5 Discussion A richer set of radiocarbon dates would not have come amiss, to help capture variations in peat thickness per century. Interpolation of dates was based on an assumption that peat depth in each column did not vary between each pair of dated samples. The alternative of using peat growth rates averaged over longer periods would have been more risky (see Illus 21.3 and Table 21.2). The dating information is unsatisfactory in another way. The measurement of the CN1 samples took place at a time when dating equipment required considerable further development and even those for the CN3 samples were not of the precision which could be expected for dates measured

Illus 21.3b The top left part of Illus 21.3 with reduced peat growth rates for all of CN1 and the upper part of CN3. Palaeoenvironment \ 828

Calanais Survey and Excavation, 1979-88

today. That may have contributed to the fact that the dates for the CN3 subzone boundary 3b-3c, one from the CN1 column and the other from the CN3 column, were nearly significantly different from one another; there is only a bit better than 1 in 20 chance that the ages could have come from contemporaneous samples (Table 21.2). The apparent conflict between dating and pollen-analytical information around the CN3 subzones 3b-3c boundary suggests that further intensive work at Callanish Leobag is desirable. The conflict does not affect the earlier period when most of the structures at Calanais were built but abundant modern dates would refine the chronology.

21.3 The Callanish-3 (CN3) and Callanish-1 (CN1) pollen columns The 60cm deep CN3 peat column presented by Bohncke (1988) was re-analysed in 1cm intervals between 10 and 55cm. Likewise the 65cm thick pollen record at CN1 was re-analysed between 15 and 25cm. This refinement leads to a more detailed zonation of the successive stages of vegetation change in comparison with the originals of Bohncke (1988). The pollen diagrams are plotted against depth separately in Illus 21.6 and 21.7, but following chrono-stratographic and bio-stratigraphic correlation, the two share a common zonation scheme from CaN-1 to CaN-4: several

Illus 21.4Â The CN1 and CN3 columns at approximately the same vertical timescale as each other Palaeoenvironment \ 829

Calanais Survey and Excavation, 1979-88

zones are sub-divided. Note that curves for pollen representing the heath communities, Calluna, other Ericaceae, Empetrum and Arctostaphylos were left out in error from Illus 21.6 and 21.7: they can be found in Bohncke (1988: figures 5a, band c). Their proportions at CN3 and CN1 are described in the text below. Glossary of common names for pollen types (taxa) reading left to right in Figures 21.6 and 21.7

Chenopodiaceae = goosefoots Spergula arv = corn spurrey Convulvulus = field bindweed Polygonum pers = redshank Plantago mar = sea plantain Triglochin = arrow-grasses Pteridium = bracken Sphagnum = bog-moss

Betula = birch Quercus = oak Corylus = hazel Salix = willow Populus = poplar / aspen Alnus = alder Pinus = Scotâ&#x20AC;&#x2122;s pine Ulmus = elm Poaceae = undifferentiated grasses Cyperaceae = sedges / rushes Lonicera = honeysuckle Melampyrum = cow-wheat Potentilla-type = tormentils Narthecium = bog asphodel Pedicularis = lousewort Polygala = milkwort Succisa = devilsbit scabious Ranunculaceae = buttercups Brassicaceae = part of the cabbage family Plantago lanc = ribwort plantain Trifolium rep = clover Urtica = nettles Rumex acet = docks, sorrels Cerealia-type = cultivated and some wild grasses Anthoceras = hornwort Asteraceae tub = part of the daisy family Asteraceae lig = part of the daisy family

21.3.1 Zone CaN-1 (60-55cm at CN3; not present at CN1) Zone CaN-1 is not directly dated. The abundance of Betula pollen and the very low representation of other tree pollen types with region-scale synchroneity (Corylus, Quercus, Ulmus: Birks 1989) restrict accurate age estimates. Projection backward of peat growth rates suggests that the record commenced in the last half of the 8th or the first third of the 7th millennium BC. Although there are uncertainties about peat growth rates its upper boundary is fairly close to the level dated by GU1234 at 53.5cm (Table 21.1) and the projected date for its end at between 6675 and 6045 cal BC should be fairly accurate (Table 21.2). The zone is characterised by high arboreal pollen (AP) percentages, almost entirely of Betula (birch). There is no requirement to regard Coryloid pollen (cf. Corylus), Sorbus, Ulmus, Pinus or Quercus as having grown locally, athough Bohncke (1988) argues this for Sorbus and Corylus. With this lack of competition, Betula will have colonised all substrates, on mineral soils and on Sphagnum-rich peat as at CN3 itself. Salix (willow) is under-represented as a pollen type, so there were actually more willow plants than the pollen

Illus 21.5Â The C14 assays for a late stage in formation of the Betula (birch) carr Palaeoenvironment \ 830

Calanais Survey and Excavation, 1979-88

Illus 21.6Â The CN3 pollen diagram from Callanish Leobag Palaeoenvironment \ 831

Calanais Survey and Excavation, 1979-88

Illus 21.7Â The CN1 pollen diagram from Callanish Leobag Palaeoenvironment \ 832

Calanais Survey and Excavation, 1979-88

count appears to suggest, and it grew alongside Betula. Melampyrum (cf. M. pratense in the Outer Isles) and Lonicera periclymenum are herbs of open woodland and scrub. Areas of open ground were very limited, supporting Gramineae and Ericaceae with some Cyperaceae, Filipendula, Succisa pratensis, Cruciferae, Umbelliferae and some ferns such as Pteridium aquilinum. Empetrum is better represented than Calluna (Bohncke 1988, fig. 5a) which suggests the development of dry heath at this time. Proportions of microscopic charcoal fragments increase. 21.3.2 Zone CaN-2 (55-26cm at CN3; not present at CN1) The projected date for the start of CaN-2 is between 6675 and 6045 cal BC. The zone probably ends at some date between 2985 and 2510 cal BC. Zone CaN-2 zone is subdivided into four subzones, based on fluctuations in the Betula curve. 21.3.3 Sub-zone CaN-2a (55-51cm at CN3; not present at CN1) This sub-zone probably started at some date between 6675 and 6045 cal BC. Betula percentages fall suddenly to 14% at 52cm, probably between 6225 and 5655 cal BC. Microscopic charcoal is relatively abundant above 54cm. There need have been no causal relation between the loss of Betula and burning because charcoal fragments may have become more common in the peat as the Betula canopy was reduced. In the same way, Quercus and Pinus pollen may be more common only because there was less Betula pollen. With Betula loss came the expansion of Salix, to 34% at 53cm. Open ground indicators like Gramineae increase, but perhaps only because the Salix canopy cover was less dense, not because there was more open ground. Consistent with this is the maintenance of the record for Melampyrum, though the abundant grains of Potentilla type, if of P. erecta (tormentil) suggest grassland that was rapidly established. Empetrum pollen is only rarely recorded. Calluna is more common. Local ground conditions around CN3 possibly became markedly drier with the reduction of Sphagnum.

21.3.4 Sub-zone CaN-2b (51-38cm at CN3; not present at CN1) This sub-zone started at some date between 6075 and 5525 cal BC. Betula values fluctuate markedly. In the lower part of the subzone, as in subzone CaN-2a, this is inversely related to the abundance of Salix, possibly cyclic, affecting each generation of a species over the c 500 years between 53 and 47cm. Gramineae values rise slowly, as do those of Calluna and other, undefined Ericaceae. Potentilla and Melampyrum do not rise at all, indicating little increase in open ground. The species of Rosaceae recorded more commonly cannot be defined. The occurrence of Tilletia sphagni reflects the local growth and probable abundance of Sphagnum on the peat surface, peaking at 45cm at the same depth/time as that of the highest values for Cyperaceae. These changes suggest a wetter peat surface then, but the subsequent loss of Tilletia sphagni and Sphagnum above 44cm, between 5375 and 4825 cal BC, suggests a dryer peat surface. Betula was adversely affected. The local Salix population disappeared. Melampyrum probably became more scarce because woodland in which it grew was reduced. Coryloid pollen, probably Corylus, increased modestly but no other tree type expanded significantly to take advantage of the ground surrendered by Betula and Salix: Ulmus and Alnus pollen are recorded contiguously but neither at proportions reflecting local colonisation. Grassland with Potentilla also declined from 45cm to 40cm. It was Ericaceous and Calluna heath that gained most from the shift to dryer ground, expanding greatly, the latter to 46 %tlp at the end of zone CaN-2b (Bohncke 1988). The end of the continuous curve for Ulmus at 40cm probably reflects the reduction of Ulmus pollen from air masses over north west Europe at around 3800 cal BC. The recovery of Betula above 40cm, between 4700 and 4150 cal BC, coincides with an increase in Gramineae, and in Pteridium, a fall in Cyperaceae and the local absence of Sphagnum as the ground became drier. The Ericaceae increase towards the top of the subzone. Numbers of microscopic charcoal fragments fall at the start of the zone but remain comparatively common. Palaeoenvironment \ 833

Calanais Survey and Excavation, 1979-88

21.3.5 Sub-zone CaN-2c (38-30cm at CN3; not present at CN1) This sub-zone starts only 1cm below the peat level dated by GU-1993, some time between 4120 and 3830 cal BC. Betula values fall briefly early in the zone. Salix was re-established as a dominant taxon for a short time before fading to become a low but constant background. Ericaceae also increased. Alnus may have become locally established but it was not a significant tree and could not resist the major recovery of Betula woodland after a time between 4000 and 3450 cal BC. This recovery can be demonstrated at CN3 from the C14 dated Betula wood (Section 21.2.4). To the dry grassland with Potentilla and Pteridium was added species of Cruciferae, Plantago lanceolata, Trifolium and Rumex. These indicate grazed grassland. The rise in Pteridium values may reflect the greater openness of Betula woodland under grazing. Cerealia-type pollen is recorded in contiguous levels from 35cm, at a date probably between 3875 and 3605 cal BC, until the end of the sub-zone. The abundance of Calluna heath declined, perhaps under grazing pressure or regular burning, in essence muirburn, converting heath to a grassy heath. Microscopic charcoal continues to be common, with a peak at the upper zone boundary (not clearly seen in Illus 21.6: see Bohncke 1988). 21.3.6 Sub-zone CaN-2d (30-26cm at CN3; not present at CN1) This sub-zone started between 3490 and 3020 cal BC. Betula once more dominated the pollen source area because, presumably in the close vicinity. A dense canopy probably grew at CN3. This is likely to have been so dense that the representation of all other plant communities markedly suffered. Thus there appears to have been almost no grassland, grazed or not, no cultivation and less evidence for burning, but these may have persisted outside the Betula woodland if this did not spread to all soils. 21.3.7 Zone CaN-3 (26-0cm at CN3; 5010cm at CN1) This zone started between 2980 and 2510 cal BC and lasted until sometime between about 560

BC and 90 AD, although its end was not directly dated and the estimate relies on an assumption of even growth rates in the CN1 peat. It lasted for very roughly two and a half millennia and encompassed most of the human activities described in the archaeological report. Based on the trends in the Betula, Corylus, Gramineae, Plantago lanceolata, Plantago maritima, Ranunculaceae, Cruciferae, and Rumex curves, and the absence of Ulmus pollen, an overlap is assumed between the top of the CN3 column and the base of the CN1 column (Illus 21.4); that there was some overlap is demonstrated by the radiocarbon dates (Illus 21.3). For the purpose of the zonation and interpretation data from both columns were used. Based on changes in the Betula curve and herbs five sub-zones are identified. 21.3.8 Sub-zone CaN-3a (26-19cm at CN3; not present at CN1) This sub-zone started sometime between 2985 and 2510 cal BC. Sub-zone CaN-3a begins with a strong decline in Betula values as this woodland declined over, perhaps, 400 years or so. No tree or shrub taxa respond to this decline. Cerealia-type pollen is recorded at the lower zone boundary when Betula values were very high whereas indicators for grazed grassland, Gramineae and Plantago lanceolata increase some 300 years or so later as Betula values fall. With the re-emergence of a farmed landscape, charcoal fragments are sharply reduced in number. This suggests that burning need not be related to agriculture. Sphagnum became locally more abundant. Ericaceae steadily increase to form with Gramineae the majority of the heath and herb plant communities. The small peak in Triglochin may be from coastal marshes. 21.3.9 Sub-zone CaN-3b (19-15cm at CN3; 64-51cm at CN1) The transition from CaN-3a to 3b took place between 2560 and 2200 cal BC. Columns CN1 and CN3 are only a small distance apart and share the same pollen source area. With the exception of one anomalous peak early in the zone, Betula values continue to fall to Palaeoenvironment \ 834

Calanais Survey and Excavation, 1979-88

15% total land pollen. Increases in Coryloid pollen and in Quercus and Alnus may again reflect the easier influx from distant sources. The single peak in Pinus pollen is difficult to interpret. Calluna heath expanded from zone CaN-3a, with Empetrum. The Cyperaceae increase rather than the Gramineae, suggesting wetter ground conditions around CN3, and the slightly stronger representation of species within the Ranunculaceae may also reflect wetter soil, as does Sphagnum in the peat at both CN3 and CN1. Similar slight increases in Plantago lanceolata show that grazing continued, as did cultivation, seen in Cerealia-type pollen and associated weeds from the Cruciferae and Caryophyllaceae. The one record of Chenopodiaceae pollen may also indicate disturbed ground but it might fit with Triglochin and Plantago maritima as representatives of coastal marshes, their pollen blown inland at a time of increased storminess (Walker 1984), their pollen more â&#x20AC;&#x2DC;visibleâ&#x20AC;&#x2122; in this zone as woodland was diminished, or plants adventitiously colonising disturbed ground. Betula recovered to 32% towards the top of this sub-zone. Evidence for burning is constant but comparatively low. 21.3.10 Sub-zone CaN-3c (15-0cm at CN3; 50-30cm at CN1) The transition from CaN-3b to 3c probably took place between about 1880 and 1520 cal BC. Betula percentages declined under 10% for the first time at CN3. Betula trees probably ceased to grow at a date within the age range 2100 and 1750 cal BC. Ulmus pollen disappears from the samples, Pinus has a marked decline and Alnus, Corylus and Quercus declined more slowly. The regional and local landscapes were effectively treeless by mid-zone. Gramineae dominates the assemblage with 30%tlp, with Ericaceae of which Calluna exceeds 53% tlp. Potentilla type pollen is of less significance, possibly because soils continued to be wet, suggested by Cyperaceae, Narthecium ossifragum, Pedicularis, Ranunculaceae and Sphagnum. It is possible that woodland loss led to elevated soil water tables as evapo-transpiration became less effective. Components of the farmed landscape, Plantago lanceolata, Plantago marit-

ima, Ranunculaceae, Rumex, Trifolium repens and Urtica in pasture and Cerealia-type, Cruciferae, Caryophyllaceae all became more abundant. Cerealia-type pollen was continuously present with a maximum at 11cm in the CN3 column: the 10 cm level was dated to between 1410 and 1110 cal BC and at an average peat growth rate calculated as between 230 and 690 years for the growth of the 5cm of peat between the 10 and 15cm levels of CN3, the Cerealia-type maximum probably lay at some time between 1500 and 1200 cal BC. Evidence for burning declines in the middle of the zone but recovers to former values. Towards the end of this sub-zone Betula percentages increase a little together with Alnus and Pinus values, while the Ranunculaceae, Cruciferae, Caryophyllaceae and Plantago lanceolata decline. 21.3.11 Sub-zone CaN-3d (not present at CN3; 30-25cm at CN1) The transition from CaN-3c to 3d probably took place between about 1070 and 420 cal BC. In this comparatively brief period tree pollen percentages increase slightly, especially Betula and Alnus local to CN1, and the Cyperaceae are more strongly represented, with Narthecium ossifragum and Sphagnum common, suggesting that wetter soils became more extensive, although the Ranunculaceae are all but extinguished by the upper zone boundary. Pasture seems to have been unaffected by these changes but Cerealia-type values decline, as do those of the Cruciferae. 21.3.12 Sub-zone CaN-3e (not present at CN3; 25-10cm at CN1) The transition from CaN-3d to 3e took place sometime between 920 and 260 cal BC. The few trees growing locally, of Betula and Alnus, were lost, abruptly. Ericaceae was abundant. Charcoal fragments are not recorded from a time between 920-260 cal BC and 780-120 cal BC. Dry grassland spread at the expense of the Cyperaceae, and Sphagnum was lost from the peat at CN1. Plantago lanceolata values decline and are very low in mid-zone, and this can be seen for Plantago maritima also. Cerealia-type pollen is also increasingly Palaeoenvironment \ 835

Calanais Survey and Excavation, 1979-88

less well represented, particularly above 20cm, a time between 780 and 120 cal BC. Whilst there is some statistical distortion in mid-zone from the abundant pollen of local Potentilla-type plants, Cerealia-type pollen percentages remain low to the upper zone boundary. In mid-zone, however, within the age-range 850-190 cal BC and 660-1 cal BC, several bare or disturbed ground indicators are better represented, such as species in the Cruciferae family, the Compositae, Spergula, Spergularia and Polygonum. The creation of bare ground occurred seemingly without an accompanying crop. This might indicate settlement creating bare ground erosion of soil. These bare ground indicators are then in turn lost as indicators of pasture (Plantago lanceolata, Plantago maritima and Rumex pollen) rise. 21.3.13 Zone CaN-4 (not present at CN3; 101cm at CN1) The transition to CaN4 took place sometime between 560 BC and 90 AD. The pollen of many herb species decline while those of Cyperaceae and Ericaceae increase. Many of these herbs are indicators of agriculture. Cerealia-type itself is not recorded by mid-zone. 21.3.14 The wider context Lewis was not treeless in the Holocene. An open woodland of Betula, Corylus, Populus, Sorbus aucuparia and Salix grew in the early Holocene (Wilkins 1984; Edwards 1996; Fossitt 1996). At Callanish Leobag Betula was dominant but Corylus was common elsewhere (Fossitt 1996; Edwards Mulder, Lomax, Whittington and Hirons 2000). Close to Calanais, land around Loch na Beinne Bige (Edwards 1996) had a complete tree cover by c. 8000 cal BC. Fossitt (1996) reported a sudden woodland decline at Loch Buialaval Beag, north west of Calanais, at c. 6650 cal BC. Areas with Calluna and Cyperaceae increased. Fossit sampled lake sediment which has a larger pollen source area than Callanish Leobag, and from this the loss of Betula trees at Callanish Leobag in subzone CaN-2a, between 6675-6045 and 6075-5525 cal BC, might

have regional significance, but no similar feature is seen at Loch na Beirige Bige (Edwards 1996). Fossitt (1996) argued for a natural, perhaps climatically induced woodland decline of high magnitude and short duration. Bohncke (1988) suggested that the decline was possibly anthropogenic, by mesolithic groups manipulating plant communities to favour wild herbivores like deer, a model originally proposed for landscapes further south (Mellars 1976) and certainly plausible in the Western Isles (Edwards 2000). In 1988 there was no archaeological evidence on the Western Isles for Mesolithic communities. In 2005 Gregory and others reported such evidence, from Northton in Harris. The evidence is not, however, unambiguous. Five burnt Corylus nutshells 14C dated to 7060 to 6090 cal BC, lay within deposits that â&#x20AC;&#x153;are best interpretedâ&#x20AC;? (Gregory et al 2005, 945) as anthropogenic midden deposits otherwise containing material of Neolithic age. It may be that burning was anthropogenic, though natural fires would not have been rare in northern Scotland (Tipping 1996). Work in 2010 however revealed a concentration of Mesolithic artefacts nearby in an old ground surface which also contained ecofacts including charcoal and other plant fossils (Bishop et al 2010, 178; 2011, 185). Burning increased, or was more easily recorded, at Callanish Leobag, sometime between 6975 and 6300 cal BC, becoming abundant at some date between 6675 and 6045 cal BC, but not at Loch Buialaval Beag. Faunal evidence for large wild herbivores, the other factor in explaining anthropogenic manipulation of woodland, is absent on Lewis in the Mesolithic period. Bennett et al (1992) and Edwards (1996) argued that absence of evidence was not evidence for absence in assigning Mesolithic-age vegetation changes on Shetland and the Outer Hebrides to grazing pressure. Woodland did not recover at Loch Buialaval Beag (Fossitt 1996). At Callanish Leobag it did. From this comparison it is possible to argue that subsequent events affecting the Callanish Leobag woodland were small in spatial scale. Betula woodland continued to be disturbed in subzone CaN-2b, perhaps by hunter-gatherers (Bohncke 1988). The short-lived decline of Betula in subzone CaN-2c, at the Mesolithic-Neolithic transiPalaeoenvironment \ 836

Calanais Survey and Excavation, 1979-88

tion was correlated by Bohncke (1988) with a similar though much larger and permanent event in west Orkney, which Keatinge and Dickson (1979) thought to have been climatically induced, more specifically from an increase in storminess. Melton (2008) has stratigraphic evidence for heightened storminess at this time on Shetland. Again there was complete recovery in woodland at Callanish Leobag soon after 3490-3020 cal BC. Anthropogenic effects on plant communities close to Callanish Leobag are probably seen after 3490-3020 cal BC in subzone CaN-2c. Cerealia-type pollen is recorded from 3875-3605 cal BC, synchronous at the resolution of dating controls with the earliest C14 dated cereal pollen in southern Britain (Brown 2007). Herb diversity increased. Cereal growing at some time between 3875 and 3605 cal BC occurred at least from time to time during the next few centuries. However at Loch Buialaval Beag (Fossitt 1996) “Cereal-type” pollen is recorded from c. 6300 cal BC. In coastal locations such pollen can be from wild grasses: Fossitt (1996) argued, perhaps conservatively, that Cerealia were present on the Western Isles only from the late Iron Age. Similarly in coastal plant communities herbs such as Plantago lanceolata grow naturally that would elsewhere characterise grazed grassland (Brayshay et al 2000). Elsewhere in the Western Isles woodland decline, probably anthropogenic because asynchronous, commenced at around 4100 cal BC but at a slow and variable rate (Brayshay and Edwards 1996; Fossitt 1996). Edwards et al (2000) suggested that woodland decline on Lewis may have been somewhat synchronous at around 3950 cal BC. Grazing livestock is more likely than axe or fire to have created this mosaic, although burning might be implicated, directly or indirectly at Callanish Leobag, possibly in heath management for grazing animals domestic or wild (Edwards 1996). This slow and possibly small-scale woodland clearance is in stark contrast to the sediment record at Loch Olabhat on North Uist (Edwards et al 2000; Mills, Armit, Edwards, Grinter and Mulder 2003) where partial woodland loss 4000-3300 cal BC, with probable cereal growing, was followed 3700-3500 cal BC by extensive soil erosion. The interpretation is that erosion was initiated by turf-stripping for con-

struction but soils continued to erode, apparently until the late Iron Age c. 400 cal AD. Subzone CaN-3a started between 2985 and 2510 years cal BC; cultivation beds at Calanais were assigned to the transition between subzone CaN-2d and CaN-3a, when cereals reappeared, or when cereal pollen became more palynologically visible again with tree loss. They were followed by the central monolith, Ring and building of the chambered cairn at Calanais. Much of the subsequent building at Calanais was assigned to CaN-3a or 3b. Evidence for ground working, some with an ard, was found in several areas and cultivation may have continued until peat covered much of the ground around the stone setting at Calanais. During analysis of the excavation samples at Calanais (below: Section 21.4) it was possible to subdivide subzone CaN-3a into phase CaN-3ai in which there were indications of agriculture and phase CaN-3aii in which there were stronger indicators of pastoralism. The change from phase 3ai to 3aii happened sometime between about 2870 and 2290 cal BC. Around Loch Bharabhat (Lomax and Edwards 2000), north west of Callanish Leobag, woodland also declined, from c. 2700 cal BC, and from c. 2000 cal BC there was a major reduction in woodland, in particular of Betula. Calluna heath is much better represented, with other Ericaceae, Cyperaceae and Potentilla-type, characteristic of wet heath from, probably, the expansion of blanket peat. This change is seen at Callanish Leobag in subzone CaN-3b from 2560-2200 cal BC. At Sheshader in the far east of Lewis the expansion of farmed land, arable and pastoral, and a decline of woodland and scrub occurred from c. 2200 cal BC. Heath management by burning may have taken place. But the landscape was also evolving in a different direction, to spreading blanket peat described aptly by Newell (1988; 89) as “too wet, too deep and acid, too often burnt or grazed to have permitted the establishment of anything but ... moorland”. In sub-zone CaN-3c at Callanish Leobag the Cerealia-type pollen maximum occurred sometime between 1400 and 1100 cal BC. Charred barley and cereal plant macrofossil grains found at Calanais were dated to sometime between 1600 and 1400 cal BC, and they may reflect this agriPalaeoenvironment \ 837

Calanais Survey and Excavation, 1979-88

culture. A basal peat pollen sample from Area C at Calanais was attributed to zone 3e and another sample from this basal peat produced a date (GU1403: 2640+/-110) between 1050BC and 400 cal BC. This radiocarbon date places the basal peat or topmost soil or a mixture of them chronologically at the subzone boundary between CaN-3c and 3d in CN1; but the interpolated date for the start of subzone CaN-3e in CN1 at between 920 and 260 cal BC was not very different: most of the first few centimetres of peat at Calanais may indeed have formed in subzone CaN-3e. The transition to zone CaN4 took place at some time between 560 cal BC and 80 cal AD. Ceral-type pollen values disappear briefly but recovered and strengthened during the rest of this zone suggests that the vegetation history of the area including Calanais and Leobag varied locally, although the general story was of sparse tree cover with fairly abundant herbs and heather. There was no evidence from the excavations at Calanais for any human activity.

insula will to a certain extent have influenced the specific composition of the samples. It is to be expected that the majority of the birch wood was located in low-lying areas like the site of Leobag, where birch wood was encountered in the peat. Hence a short regeneration phase of the birch wood would be registered more clearly in the peat columns from Leobag than in samples from the promontory. The above meant difficulties might be encountered in attempting precise correlations between the pollen records of the Kubiena boxes from Calanais and the CN1 and CN3 columns from Leobag. 21.4.2 Area D sampling

21.4 The Kubiena box-samples from the excavation site 21.4.1 Introduction It must be kept in mind that organisms actively mix pollen of different ages in soils. This may not be a significant problem at Calanais, with its acid soils, but the mixing means short term fluctuations will be blurred. Hence short term fluctuations might not show up as clearly as in the peat columns. Moreover due to the intensive human activity at the site of the stone circle a continual registration of the contemporary environment might periodically have been interrupted. Further, some material was dug up and dumped at higher stratigraphic levels without acquiring new pollen, as is for instance suggested by clods of grey clay with a pollen spectrum suggestive of the later 3rd millennium BC in the backfill of Victorian intrusions in the cairn. The differences in palaeo-hydrological conditions between the promontory on which the stone circle is situated and the low lying Leobag pen-

Illus 21.8 Sampling profiles against an interpretative reconstruction plan of cultivation beds. The scale is in metres. In the baulks of Area D black humus rich layers were visible and interpreted as litter or ‘mor’ layers (layers of decayed organic matter), here referred to as turf lines. At most three successive turf-lines were visible in the areas sampled. It was hoped that pollen analysis would reveal the meaning of these turf-lines and secondly it was hoped that the pollen analyses would indicate their dates. For this purpose seven Kubiena box samples were taken from the trench edges. Three boxes were taken in an east facing baulk (Illus 21.8 Section 102a) and four were taken in the main north facing baulk (Illus 21.8 Section 105). In the laboratory, boxes 1 to 7 were stratigraphically sampled to provide three short profiles (Illus 21.9-13). Palaeoenvironment \ 838

Calanais Survey and Excavation, 1979-88

Illus 21.9 Section 102A with Kubiena boxes 1 and 3-2 marked

Illus 21.10 Section 102A with Kubiena boxes 1 and 3-2 [Film PO12]

Illus 21.11 Stratigraphy, phasing and zoning in Area D Section 102a Palaeoenvironment \ 839

Calanais Survey and Excavation, 1979-88

The columns are archaeologically interpreted as spanning the soils of cultivation beds (Illus 21.8), possible upcast from standing stone pit digging and overlying turf lines and soils. To fix the position of these profiles in time their pollen diagrams were compared with the Callanish Leobag columns CN3 and CN1. 21.4.3 Description of the samples: Section 102a 21.4.4 Zonation of the pollen diagrams from Section 102a section

102a box-1

Phase-1, 7-3cm During this phase the Betula values declined to around 10% from a maximum of 19%. The pine curve remained above 1%. The Cerealia pollen curve was similar to that of Ericaceae. Pollen of Compositae liguliflorae was relatively abundant, while the Cruciferae pollen was more frequent in the bottom two samples. Towards the end of this phase Gramineae increased to around the 55% level and the Ranunculaceae become more frequent. Phase-2, 3-0cm A decline in Cerealia was registered which coincided with an increase in Ericaceae and a decline in Gramineae. At the transition to phase-2 Ulmus pollen appeared in the samples and the Betula curve stayed roughly constant, declining a little towards the end. During both phases Plantago lanceolata was relatively abundant. section

102a box 2-3

Phase 1, 13-11cm The tree pollen percentages were high (max 82% at 11cm) mainly consisting of Betula and Corylus and to a much smaller extent Pinus, Alnus and Sorbus. The non-tree pollen percentages were mainly formed by Gramineae. Ericaceae were almost com-

pletely absent. Fern spores were extremely abundant (Polypodium and Monolete spilate type). Phase 2, 11-6cm There was a strong decline in the Betula and Corylus curves, mirrored by a steep increase in Gramineae curve. Cerealia pollen was present during this phase and showed two successive maxima, together with the curve of the Compositae liguliflorae. Fern spores declined strongly in the second half. Furthermore diverse herb vegetation was present including Plantago lanceolata Ranunculaceae, Cruciferae and Trifolium. Phase 3, 6-3cm This phase saw a decline in the Cerealia, Compositae liguliflorae and Gramineae curves in favour of the Ericaceae. The Betula curve declined a little more. Pinus was continuously present in values higher than 1%. Phase 4, 3-0cm Cerealia pollen returned, while Gramineae increased at the expense of the Ericaceae. Compositae liguliflorae pollen did not reach the same frequencies as in the preceding Phases 1 and 2. Betula stayed relatively low and Pinus values did not drop under 1%. 21.4.5 Description of the samples: Section 105 Sampling was undertaken on a cleaned-back face of Section 105 (Technical Note 21.4.5). The two columns were taken about 40cm apart from one another and the section shows that the turf lines in Boxes 4 and 5 did not connect with those in Boxes 6 to 7, suggesting a fairly complex stratigraphic relationship. That complexity was reflected in the results of pollen analysis (Illus 21.17). The reason may be that the section cut obliquely across the slope of a cultivation bed. Boxes 4 to 5 are situated where a cultivation bed was truncated by later ploughing while boxes 6 to 7 were in the soils which had formed over the trough between that bed and Bed 2 to the north.

Palaeoenvironment \ 840

Calanais Survey and Excavation, 1979-88

Illus 21.12 Area D Box 1 pollen

Illus 21.13 Area D Boxes 2 and 3 pollen

Illus 21.14 Part of section 105 with Kubiena boxes 5-4 and 7-6 marked Palaeoenvironment \ 841

Calanais Survey and Excavation, 1979-88

21.4.6 Zonation of the pollen diagrams from Section 105 on Area D section

105 box 4-5

Phase 1, 11-8cm The tree pollen percentages were high (max 82%, 11cm) mainly consisting of Betula and Corylus and to a smaller extent Pinus, Alnus and Sorbus. The non-tree-pollen percentages were mainly formed by Ericaceae and Gramineae. Fern spores were extremely abundant (Polypodium and Monolete spilate type). This description is the same as that for Box 2-3 phase-1. Phase 2, 8-4cm There was a strong decline in the Betula and Corylus curves, followed by a steep increase in the Gramineae curve. Cerealia pollen was present during this phase and showed two successive maxima, together with the curve of the Compositae liguliflorae. Fern spores declined strongly. Furthermore

diverse herb vegetation was present including Plantago lanceolata Ranunculaceae, Cruciferae and Trifolium. This description is much as for Box 2-3 phase-2 except that Ericaceae pollen, towards the end of this phase, were more abundantly present and that the Cerealia curve did not show a maximum immediately after the Betula and Corylus decline. Phase 3, 4-2cm This phase saw a decline in the Cerealia, Compositae liguliflorae and Gramineae curves in favour of the Ericaceae. The Betula curve declined a little. Pinus was continuously present in values higher than 1%. This description is as for Box 2-3 phase-3, except that the decline in Gramineae was not so prominent and the Cerealia did not disappear altogether from the samples. Phase 4, 2-0cm Cerealia pollen returned, while Gramineae increased at the expense of the Ericaceae. Com-

Illus 21.15Â Area D Section 105 with Kubiena boxes 4-5 [Film PO11] Palaeoenvironment \ 842

Calanais Survey and Excavation, 1979-88

Illus 21.16Â Area D Section 105 with Kubiena boxes 6-7 [Film PO10] positae liguliflorae pollen did not reach the same frequencies as in the preceding Phases 1 and 2. Betula recovered a little and Pinus values did not drop under 1%. This description is the same as that for Box 2-3 phase-4. section

105 box 6-7

Phase 1, 14-10cm Straight from the bottom of this column the NAP values dominated, mainly formed by Gramineae and Ericaceae. Cerealia-type pollen was continuously present and the Compositae liguliflorae pollen was relatively abundant. Moreover Plantago lanceolata, Ranunculaceae and Succisa pollen occurred in the samples. Pinus formed a continuous curve and did not drop under 1% Betula fluctuated around 10%.

Phase 2, 10-8cm A decline in the Cerealia, Gramineae and the Compositae liguliflorae was registered followed by an increase in the Ericaceae. Towards the end of this phase the Betula curve increased a little to 17% at the 8cm level. Phase 3, 8-5cm The Betula values declined sharply, while Gramineae increased strongly and Cerealia-type pollen returned, but the Compositae liguliflorae did not reach the high values of the preceding Phase 1. The Pinus curve dropped under 1% values for the first time. Towards the end of this phase Gramineae curve dropped to values around 40%. Phase 5, 5-0cm Ericaceae restored to values around 30% and the Betula curve increased a little.

Palaeoenvironment \ 843

Calanais Survey and Excavation, 1979-88

Illus 21.17Â Stratigraphy, phasing and zoning in Area D Section 105

Illus 21.18Â Area D Boxes 4 and 5 pollen 21.4.7 Interpretation and Correlation of columns D2-3 on Section 102a and D4-5 on Section 105 The pollen records of columns D2-3 and D4-5 both started off with relatively high Betula and Corylus percentages after which a clearance phase

followed. This gives a firm indication for the start of both columns. Straight after the decline in the Betula and Corylus values the first Cerealia-type pollen appeared together with Compositae pollen of the Liguliflorae type and Plantago lanceolata. Subsequently during Phase D-2 Gramineae increased firmly and pollen of the Trifolium-type Palaeoenvironment \ 844

Calanais Survey and Excavation, 1979-88

Illus 21.19 Area D Boxes 6 and 7 pollen and Trifolium repens appeared in the samples, after which the indicators for agriculture decline and Ericaceae increased strongly in phase D-3. Possibly during this phase cattle breeding was still practised since no decline was registered in the Plantago lanceolata, the Ranunculaceae and the Trifolium curves. Gramineae showed a temporary decline during this zone but recovered again in Phase D-4, together with Cerealia. Ericaceae declined, possibly as a consequence of increased agricultural activities. In both columns the Betula curve did not decline below 10% and Pinus and Alnus pollen were relatively abundant. Pinus did not decline below 1% which means that both columns covered a period preceding the regional pine decline (+4000 BP, Bohncke 1988), but following the period with an abundant presence of Betula. This phase D-1 most likely correlates with the “regeneration” phase of birch registered in zone CaN-2d of column CN3. Phase D-2 and D-3 correlate with zone CaN-3a in column CN3 which was characterised as predominantly agricultural in its initial phase but cattle-dominated in its second half. D4, the phase with renewed agricultural influence must correlate with sub-zone CaN-3b. The regional pine decline in column CN3 started shortly after the beginning of zone CaN-3c. Here the Pinus percentages declined under 1%. As already discussed in the introduction to the samples from the excavation site the short term

regeneration phases of the birch at the end of zone CaN-3a and CaN-3b were hardly registered but the curve of the pine pollen, which formed a more regional component in the pollen rain, was in good agreement. In column D4-5 there seems to be a correlation between the occurrence of the so called “turf-line” in the lithology and an increase in Ericaceae pollen. The fact that the curve of Ericaceae in column D4-5 showed two successive maxima but the curve in column D2-3 had only one peak is probably due to local influences. 21.4.8 Interpretation and Correlation of Column D6-7 on Section 105 Column D6-7 lacked the phase with relatively high Betula and Corylus percentages but started with a phase during which Betula values fluctuated around 10% and the Pinus curve did not decline under 1%. Among the herbs, Gramineae and Ericaceae were abundantly present. Pollen of the Cerealia-type and Compositae liguliflorae was present together with Plantago lanceolata, Ranunculaceae and Trifolium pollen, which can be interpreted to mean that both agriculture and cattle breeding was practised during phase-1 in column D6-7. In phase-2 Ericaceae increased at the expense of Gramineae. The Cerealia curve was interrupted but Plantago lanceolata, Ranunculaceae, Trifolium and Potentilla stayed present in Palaeoenvironment \ 845

Calanais Survey and Excavation, 1979-88

more or less equal percentages. The Betula values recovered to 17% at 8cm. At the start of phase-3 in column D6-7 a sharp decline in the Betula values was registered. Cerealia-type pollen reappeared in the samples and Gramineae increased sharply. Ericaceae pollen declined, possibly as a consequence of increased agricultural activity around the site. Together with the birch, Corylus, Alnus and Pinus declined. Pinus fell below the 1% level which coincided with the regional pine decline of around 4000 BP. This decline in the pine pollen marked the beginning of zone CaN-3c in column CN3. During phase-4 in column D6-7 the Cerealia curve increased but the Compositae liguliflorae did not reach the high values of Phase 1 in column D6-7. Ericaceae values recovered while Gramineae declined to around the 40% level. Phase 3 correlated with zone CaN-3b. This fits in nicely with a correlation between Phase 1 and 2 in column D6-7 and zone CaN-3a in column CN3. Column D6-7 contained three â&#x20AC;&#x2DC;turf linesâ&#x20AC;&#x2122; and as in Column D4-5 there seems to have been a marked correlation between the occurrence of these turf-lines and peaks in the Ericaceae curve. Besides Ericaceae the turf lines seem to correlate with temporary declines in the Cerealia curve, which must be explained as temporary declines in agricultural activity around the site. 21.4.9 Interpretation and Correlation of Column D1 The relatively high Betula values and the continuous presence of Pinus pollen at values higher than 1%, together with the abundant presence of Compositae liguliflorae pollen places the start of the pollen registration in box-1 shortly after the birch and hazel decline in zone CaN-3a. Successively a stage during which agriculture was practised (presence of Cerealia and Compositae liguliflorae), a stage during which both agriculture and cattle-breeding (increase in Gramineae and Ranunculaceae) were practised and a stage (phase-2) during which agricultural activities declined and heather spread were registered in the samples. These three stages are also clearly registered in the CN3 column zone CaN-3a. The increase in Ericaceae marked the transition to zone CaN-3b.

21.4.10 Archaeological comment The profiles sampled two cultivation beds, Bed 1 (Section 105) and Bed 4 (Section 102a). Archaeological interpretation of the layers in Section 102a in Boxes 1 to 3 (Phases 2 and 1) was that cultivation bed 4 overlay a pre-cultivation soil. The earliest pollen phase in the boxes correlates with zone CaN-2d during which Cerealia-type pollen was absent at Callanish Leobag and although it is as likely to have been absent there because of increased tree cover as because of a cessation in cereal-growing, that at least does not conflict with, the archaeological interpretation of the lowest soil as not being part of the cultivation bed. On Section 105 (Cultivation bed 1), as mentioned above, the layer at apparently the same level as gingery brown soil 320 which covered much of the ground outside the cairn on Area D was not labelled during section-drawing. It has been labelled 950 during post-excavation for reference purposes. In essence Section 105 cut at a shallow angle along Bed 1. Indeed, that had been my interpretation from other evidence but I had not been able to interpret the detail until I took the pollen report fully into account. Thus the pollen in Boxes 6 and 7 which samples soils overlying the slope down into a trough between cultivation beds was at a higher stratigraphic level than the samples from Boxes 4 and 5. The top 5cm in Box 7 (Phase 5) was taken from the lower part of plough soil 315. The archaeological dating for the plough soil is much the same as the CN3 column dating for late zone 3b to zone 3c, which fits well with the sequence in Box D7. It contrasts slightly with the zonation of a spot sample from 315 on the main part of DI to CaN-3b but variations within a plough soil can be expected. Layer 950 on Illus 21.14 was apparently at the same stratigraphic level as ginger-brown clay 320 on the area south of the cairn, which was crisscrossed with ard marks. However clay 320 was not present close to the section in typical form in plan during excavation. A turf line (or at least a humus-rich band of material) appeared to be lie between the plough soil and the underlying layer 950, although it was not noticed during excavation. Palaeoenvironment \ 846

Calanais Survey and Excavation, 1979-88

The strata sampled in the two box columns below the upper turf line reflect material derived from Ring stone pits which cut through one or more cultivation beds. The sequence appears to reflect the following events. Spoil from the pits was dumped nearby, with, naturally, the material from cultivation beds underlying the spoil from digging through the natural green clay. The green clay is thought to have oxidised to gingery brown. There was some deposition of burial-related material in the nearby part of DI. Then the green or gingery brown clay was spread over that area, and the underlying more humic part of the spoil heaps was levelled out more locally. Without the evidence from the pollen samples the cultivation bed might have been attributed to CaN-2c, the early cultivation phase attested in the pollen column at Callanish Leobag and more precisely it might have been supposed that creation of the cultivation bed started at the same date as formation of the 35cm level in Leobag column CN3, where Cerealia pollen first appeared. This dated approximately to between 3900 and 3600 BC. It might also have been supposed that use of the cultivation bed ended at the CaN-2c to 2d transition at between 3490 and 3020 cal BC. The conclusion from the pollen in the Kubiena box profiles on Area D is however that use of the cultivation beds started right at the end of CaN-2d and thus the cultivation beds must date to the beginning of CaN-3a (starting between 2980 and 2510 cal BC at Leobag but interpreted from the archaeology as having started early in

that range). The correlation of the turf lines with increases in Ericaceae provided powerful support for the idea that each of them represented formation of a turf line and a diminution in human farming activities near the place. The detail of this interpretation could not have been developed without the evidence from the pollen-analysis of the Kubiena box samples from Area D. Thus they did, as hoped, greatly illuminate the interpretation of the archaeology. 21.4.11 Area H sampling Please see illus 21.20 below. 21.4.12 Description of the samples 21.4.13 Zonation of the pollen diagrams of area H box-1 zonation

Phase 1, 6-4cm This phase was characterised by high Betula and Corylus percentages, respectively max 54% and 31%, and relatively low NAP values. The herbs in this phase were dominated by Gramineae. Phase 2, 4-0cm At the transition to Phase 2 Betula showed a strong decline, while Gramineae increased to values around 70%. Towards the top of box-1 Ericaceae began to increase. Furthermore Phase 2 was characterised by a firm increase in Plantago

Illus 21.20Â Area H Section 33 with Kubiena boxes 1 and 2 marked [NMRS-DC38305 Palaeoenvironment \ 847

Calanais Survey and Excavation, 1979-88

Illus 21.21Â Stratigraphy, phasing and zoning in Area H

Illus 21.22Â Area H Box 1 pollen lanceolata, and Ranunculaceae and Compositae liguliflorae were relatively abundant. Cerealia pollen was absent from the samples. box-2 zonation

Phase 1, 6-4cm Betula values were beneath the 10% level and Pinus below the 1% level. Gramineae were abundantly present. Furthermore Plantago lanceolata, Cerealia and to a lesser extent Compositae leg and Ranunculaceae were present in the samples. Phase 2, 4-2cm Betula, Corylus and Pinus increased a little. Cerealia pollen was absent during this phase and Gramineae declined in favour of an increase in the Ericaceae. Sphagnum suddenly appeared in the samples.

Phase 3, 2-0cm The Betula curve declined again, while Corylus maintained values around 10%. Cerealia pollen re-appeared and Gramineae increased a little, as did the Plantago lanceolata curve. 21.4.14 Interpretation and Correlation box h-1

The strong decline in the Betula and Corylus values between Phases 1 and 2 followed by a sharp increase in the Gramineae and Plantago lanceolata curves was a marked phenomenon in the pollen record of CN3 and took place at the transition of zone CaN-2d to zone CaN-3a. The low Pinus values were unexpected for this stage in the vegetation development as was the absence of Cerealia pollen which was expected to increase together with the Compositae liguliflorae curve. Palaeoenvironment \ 848

Calanais Survey and Excavation, 1979-88

Illus 21.23 Area H Box 2 pollen

Illus 21.24 East-west section 8A on the south-facing baulk of Area E [NMRS DC38249A]

Illus 21.25 he sampling area on the west baulk [Film 1981.18.19] Palaeoenvironment \ 849

Calanais Survey and Excavation, 1979-88 box h-2

The low Betula and Corylus values in phase-1 together with Pinus percentages lower than 1% and an absence of Ulmus pollen indicated, for the start of this column, a period after the regional pine decline. The increase in AP values in Phase 2, mainly Betula and Corylus, together with a temporary decline in the Cerealia curve possibly correlated with sub-zone CaN-3d. On the other hand the Corylus curve did not attain values of around 10% during this zone in the CN1 column and also Gramineae were relatively high for this zone. The presence of Sphagnum spores in this zone could have been the result of local wet conditions and cannot be used for more regional correlations. Another possible interpretation is that the increase of Corylus in Phase 2 was the same as that registered in the top of sub-zone CaN-3a, which also coincided with a decline in the Gramineae and Cerealia curves and a slight increase in Ranunculaceae. Ericaceae did not increase to such an extent as in Area D, but local wet conditions, as indicated by the presence of Sphagnum spores, might have prevented this regeneration. The ‘turfline’ between the 3 and 4.5cm level, if this interpretation is right, coincided with the middle turf-line and coincided with a temporary decline in the agricultural activities, as indicated by the decline in the Cerealia curve. The green clay can be dated to the first half of zone CaN-3a.

riod. Slight timber structures were then erected on the clay and the sudden reduction in tree pollen between Box 1 zones 1and 2 might have a very local and anthropogenic explanation in the exploitation of local birch and hazel for wattle walls. The sudden appearance of Sphagnum spores in Box 2 at the 4-2 cm level could, as suggested above, represent water-logging on the old turf line before the green clay was imported. It might even explain why the clay was laid. Spot sample 2061 from the turf line immediately overlying the green clay under the cairn was zoned to CaN-3ai. The implication of the pollen evidence from the box column and spot sample is that the green clay platform was laid late in CaN-3ai and the overlying cairn was built early in CaN-3aii. The pollen zone inversion is thus explicable within the archaeological interpretation. Indeed, although somewhat unexpected it could be said to fit uncannily well. 21.4.16 Area E sampling No drawn record of the precise sampling point in Area E has survived. However it was almost certainly below the highlighted (white) Kubiena box on the timber above the section in Illus 21.25. Judging by the descriptions in Table 21.4 it must have been below both fibrous and amorphous peat, in the top of podzolised soil 1108. 21.4.17 Description of the samples

21.4.15 Archaeological comment When the two pollen diagrams are correlated stratigraphically there appears to be a chronological inversion. The lower part of the green clay appears to contain pollen of the earlier zone CaN-2d but the underlying layers were zoned to CaN-3a in Box 2 (Illus 21.21). But the green clay had been freshly excavated from the native till. Dr Bohncke observed in his introduction the short period within which some deposits were exposed to the pollen rain means that variations need not have shown up in the master columns at Callanish Leobag. If the clay was rapidly spread out it would have acquired pollen representative of a short pe-

Table 21.4 Kubiena box from Area E Box-1 0-3cm 3-7cm

Darker charcoal-rich humic clay perhaps 1112 or a variant

Charcoal-rich humic clay with more fine grit, probably soil 1108.

21.4.18 Zonation of the pollen diagram of area E In the pollen record three phases can be distinguished. Palaeoenvironment \ 850

Calanais Survey and Excavation, 1979-88

Phase 1, 7-4cm The AP percentages were relatively high (75-80%) and consisted mainly of Betula pollen and to a lesser extent Corylus. Fern spores were abundant but the herb vegetation did not seem to be very diverse and consisted mainly of Gramineae and Melampyrum although Cerealia-type pollen and Plantago lanceolata did occur. Towards the end of this phase the Betula curve increased to around 60%. Ulmus pollen was present. Phase 2, 5-2cm This phase was characterised by a slight decline in the Betula values (min 40%) followed by a strong increase in the Corylus curve (max 40%). During this phase Cerealia was absent. The fern spores were still relatively abundant. Ulmus pollen was present.

Phase 3, 2-0cm The Corylus pollen curve declined a little. Gramineae were increasing. Cerealia reappeared in the samples. 21.4.19 Interpretation and Correlation The presence of Ulmus pollen in the samples together with Pinus pollen around 1% and the relatively high Betula frequencies indicates that the registration started before the marked decline in the birch and hazel values, which characterised the beginning of zone CaN-3a. The increase of Corylus at the expense of Betula was a feature that occurred at the transition of zone CaN-2d to CaN-3a and can be explained as the effect of the opening up of the mixed birch and hazel forest by small-scale clearances. Due to the increase in

Illus 21.26Â Area E pollen

Illus 21.27Â Area C section sampled Palaeoenvironment \ 851

Calanais Survey and Excavation, 1979-88

liguliflorae in the forest Corylus flowered more strongly. Following the clearance agriculture was practised judging by the presence of Cerealia pollen. Box E shows the transition of zone CaN-2d to CaN-3a and possibly zone CaN-2c was also reflected in the bottom samples. 21.4.20 Archaeological comment Area E lay at the edge of a cultivated area and a small cache of Hebridean and other pottery was found on the surface of soil 1108 which covered much of the trench. If the soil belonged to the end of CaN-2d rather than the start of CaN-3a, and the various darker patches lying on it (sampled in the 0-3cm level of the Kubiena box) belong to a period when agriculture was practised in CaN-3a that would fit the archaeological interpretation very well: the west edge of area E was just outside the cultivated area. As on Area C the darker deposits could represent the rotted remains of weeds with soil attached. 21.4.21 Area C sampling In Area C in 1980 a baulk was left in the middle of the trench, encapsulating the fallen standing stone and related archaeology. It was sampled using 5 Kubiena boxes in a continuous column which covered the whole profile apart from the upper peat and turf (Illus 21.27). Box 2 was not available for pollen analysis. 21.4.23 Zonation of the pollen diagram Throughout column C the Betula values stayed relatively low (e.g. below the 10% level) Pinus pollen was relatively abundant in box-1 but dropped under the 1% level in box-3 and subsequently declined slowly to form a discontinuous curve. Ulmus pollen occurred incidentally in the samples. The fluctuating presence of the different herb species was used as the main guideline to separate phases.

21.4.22 Description of the samples

Phase 1, 35-28cm This phase was characterised by the presence of high Plantago lanceolata, Cerealia and Compos-

Illus 21.28Â Stratigraphy, phasing and zoning in Area C Palaeoenvironment \ 852

Calanais Survey and Excavation, 1979-88

Illus 12.29Â Area C Boxes 1 to 5 pollen itae liguliflorae pollen percentages. Potentilla-type pollen was present and Gramineae were relatively abundant. Pinus pollen did not decline under 1%. Towards the end of this phase Ericaceae pollen increased strongly. Phase 1 was followed by a period which cannot be assessed (the absent box 2 28-21cm). It included formation of a turf line Phase 2, 21-15cm During this phase Cerealia pollen was completely absent, while the Compositae liguliflorae pollen curve was relatively low. Both Ericaceae and Gramineae were dominant in the NAP. Plantago lanceolata pollen was relatively abundant while the Ranunculaceae were less frequent than in the preceding phase. Trifolium pollen remained present in the samples. Phase 3, 15-13cm A temporary increase in the AP percentages was registered, together with a sharp increase in Er-

icaceae. Ulmus pollen was present. The Plantago lanceolata curve declined, while the Potentilla curve increased. Phase 4, 13-8cm A marked increase in the Cerealia curve and a further increase in the Potentilla curve, combined with a decline in the AP percentages formed the start of this phase. Gramineae and Plantago lanceolata recovered a little, while Ericaceae show a firm decline. The Plantago maritima curve was more prominent during this phase. Rumex pollen was continuously present but Trifolium species disappeared from the samples. Phase 5, 8-2cm The Cerealia curve declined again followed by a sharp increase in the Ericaceae. The Plantago lanceolata and the Potentilla curve showed a reduction. Gramineae, after an initial decline recovered again towards the end of this phase. The Cyperacea showed a sudden peak at the 2cm level. Palaeoenvironment \ 853

Calanais Survey and Excavation, 1979-88

Phase 6, 2-0cm In this phase the Cyperaceae dropped suddenly and also Ericaceae showed a further decline. The Cerealia curve and the Rumex curve increased. Furthermore Plantago maritima was more frequent. 21.4.24 Interpretation and Correlation Some traces of the Betula and Corylus decline were visible in the bottom samples Pinus pollen was relatively abundant. The sequence started off straight after the clearance phase registered in zone CaN-3a in the CN3 column. Also the presence of relatively high Gramineae values together with Plantago lanceolata and Compositae liguliflorae was characteristic of this zone. Within this zone CaN-3a the pollen assemblage of phase-1 was comparable with the stage during which both agriculture and cattle breeding was practised. The top of this phase was comparable with the Ericaceae increase at the transition of zone CaN-3a to CaN-3b. This increase in the heather in the samples from area D coincided with the middle turf-line. Due to the absence of box-2 a complete registration of the successive zones within zone CaN-3 was impossible. It is difficult to establish how far zones CaN-3b and CaN-3c were represented in box-2; hence phase-2 was difficult to place in time. Anyhow the temporary increase in the AP percentages in phase-3 could possibly be compared with zone CaN-3d in CN1 column, starting between 1070 and 420 and ending between 920 and 260 cal BC. The comparison of phase-4 and phase-5 with zone CaN-3e in column CN1 was more striking. The increase in agricultural activity as indicated in the Cerealia curve, together with the strong increase in the Potentilla and a more pronounced presence of Rumex and strong fluctuations in Ericaceae curve were features which also characterise zone CaN-3e. Moreover from this level onwards the Pinus curve became discontinuous in both column C and CN1. Another mutual feature was the increase in Rumex and Cerealia pollen and a decline in the Ericaceae. This feature in column

CN1 marked the transition to zone CaN-4, perhaps between about 500BC and 200AD, and in Column C this increase formed the transition to phase-6. Phase-6 possible correlated with the beginning of zone CaN-4 but other interpretations were possible. If the top part of column C represented the upper part of the remaining peat after the peat-cuttings in the 19th century AD than the sudden increase in the Cyperaceae could well have represented the level to which cutting took place, and after which peat accumulation started again. If so, the upper phase, phase-6, reflected the recent situation in the Calanais area. The cultivation bed system that was practised up to pre-modern times could have contributed to the presence of Cerealia pollen in the samples. 21.4.25 Archaeological comment The Phase 1 context might have been a mound of peaty material brought from elsewhere. But the preferred archaeological interpretation is that it had been a mound of weeds with attached soil. The abundance of Plantago and Cerealia-type pollen in it supported that interpretation. The mound contained Food Vessel pottery, which is somewhat unexpected given the palynological dating to sometime between the start of CaN-3a between 2980 and 2500 cal BC and its end between 2560 and 2200 cal BC. The Food Vessel tradition is supposed to have arisen slightly later, in the 22nd century BC. The zoning of basal peat to the end of CaN-3d is supported by the zoning of basal peat pollen spot sample 275 from Area C to Zone 3e. The transition from 3d to 3e at between 920 and 260 cal BC seems likely to have occurred when peat started to grow without restraint. Along with a basal peat date of 1050 to 400 cal BC from sample Call80/54/244 (GU-1403 2640+/-110 BP) from the Area C east baulk it helps to narrow down the date of peat formation to between 920 and 400 cal BC (Chapter 8: Area C). The pollen analysis suggests that all of the peat grew in prehistory; but the field interpretation was that peat 201c had been cut in a relatively modern period and that 201a and 201b represented rePalaeoenvironment \ 854

Calanais Survey and Excavation, 1979-88

Illus 21.30Â Contexts spot-sampled for pollen in Area D newed growth. Local phases 5 and 6 have a pollen spectrum not very different from that in CaN-6 in pollen column CN2 at Callanish Leobag which may have a medieval to modern date (see discussion of Sample CN3-119 in 21.3.3 above). Perhaps a match with a much later zone in the CN2 column at Callanish Leobag or in the pollen columns from Calanais Fields should be explored. 21.4.26 Archaeological comment on Kubiena box samples as a whole With few exceptions there was a good match between archaeological and palynological interpretations of the layers sampled in Kubiena boxes on all of the sites. The pollen provided considerable additional information, sometimes helping to discriminate between alternative archaeological hypotheses. The soil characteristics noted in Dr Bohnckeâ&#x20AC;&#x2122;s report were a useful supplement to the archaeological record.

21.5 Pollen analyses of the spot samples The stratigraphic and descriptive information originally supplied to Dr Bohncke contained many errors. This section has therefore been rewritten extensively. Any resulting errors are the responsibility of PJA. The intention was to fit the spot samples from the excavation site into the sequence on the base of their pollen characteristics. For this purpose, the successive stages as found in the Kubiena box-samples were used for comparison. 21.5.1 The stratigraphy of the spot samples from Area D The spot samples from Area D came from a variety of contexts (Illus 21.30). The analyses have been grouped in roughly stratigraphic order from early to late. They show that the cultivation beds were open in early CaN-3a. Some of the subsequent old

Palaeoenvironment \ 855

Calanais Survey and Excavation, 1979-88

turf lines south of the cairn pre-dated the cairn. The cairn itself was probably built in CaN-3aii which started between ended between 2470 and 2200 cal BC. The topmost old turf line continued to grow in the area south of the cairn after the cairn had been built. The upper plough soil was still in use in CaN-3c. 21.5.2 Spot sample from early cultivation beds on Area D Table 21.5 Spot sample from 379 Area

Sample Context Field interpretation

2031

379

A displaced turf originally growing on clean green clay, in cultivation bed in Area DII

Pollen characteristics: The AP values were comparatively high (16%) and dominated by the presence of Corylus pollen (10.6%). 1.8% Pinus pollen was present. Furthermore the pollen from Plantago lanceolata and Compositae liguliflorae was relatively abundant, respectively 12 and 7.18%. Cerealia pollen was absent. Correlation: The pollen assemblage of this sample can be correlated with levels from the beginning of zone CaN-3a. Here, after an initial clearance in the birch wood, there was a short term increase in the hazel pollen curve. At the same time the Plantago lanceolata and the Compositae liguliflorae increased firmly and there was a development towards the agriculture stage of zone CaN-3a. Archaeological comment: This was one of the earliest turfs from the site, if its interpretation as a component of a cultivation bed is accepted. It may have been cut from turf which grew on an area previously stripped of soil. That stripping may have been associated with the digging of the curving ditch or (unknown) related features. 21.5.3 Spot sample from between cultivation

beds and old turf lines on Area D Table 21.6 Spot sample from 377 Area

Sample Context Field interpretation

2087

377

Chocolate brown, fine silty clay over 388 and under 365 but also upcast lying on 365.

Pollen characteristics: The AP percentage formed 11.2% of which 7.8% was Corylus. The NAP was dominated by Gramineae (47%) while Ericaceae were relatively low. Plantago lanceolata and Compositae liguliflorae were abundantly present. Cerealia was present. Correlation: The above characteristics together with pine values of 1% points pointed to a correlation with an early stage in zone CaN-3a, where the Corylus was still high but where agriculture was practised, as indicated by the presence of Cerealia pollen and the relative abundant presence of Compositae liguliflorae. Archaeological comment: The label 377 was used for a layer overlying the old cultivation beds, although as described in Chapter 9.3 pits had been dug in 377 and patches of clay of a very similar texture and colour lay over the top of the prevalent old ground surface 365. The correlation with an early stage in CaN-3ai is in accord with the archaeological interpretation 21.5.4 Spot samples from pits on Area D Table 21.7 Spot sample from 916 Area

Sam- ConDescription ple text 2017

916

Field interpretation: Post hole with stone packing and infill of brown clay with an ambiguous relationship to old turf lines 365, which it may have cut.

Palaeoenvironment \ 856

Calanais Survey and Excavation, 1979-88

Pollen characteristics: Relatively low AP values, but Pinus remained at levels over 1% and Corylus was relatively abundant. The sample contained high Gramineae and Plantago lanceolata values. Correlation: This sample belongs in the second half of CaN-3a during which there was an emphasis on cattle breeding. Archaeological comment: The posthole lay on the northern side of cultivation bed 2, which was represented by 388. Its top had been disturbed and it is not clear whether it originally cut the main prevalent turf lines in Area D, but the pollen zonation suggests that its fill was roughly contemporary with use of the chambered cairn. Probably, then, it had been cut through the prevalent turf lines.

OGS 365 was ambiguous; the post hole may have cut it. However the post hole definitely underlay layer 320, which was in Zone 3b, and the tentative correlation with CaN-3c cannot be correct unless layer 320 was wholly a result of post-depositional soil processes, which is not the preferred interpretation. Instead the lack of particular phase characteristics, perhaps caused by mixing of soils with pollen from different zones, should be emphasised. An alternative interpretation is that the upper part of the post hole was filled with material indistinguishable from 320 and was missed during excavation, but given the ambiguity of the pollen signature it is not preferred. Table 21.9 Spot sample from posthole 385 Area Sample Context Field interpretation

Table 21.8 Spot sample from posthole 392 Area

Sample Context Field interpretation

2032

392

Post hole, full of black, charcoal impregnated with an ambiguous relationship to old turf lines 365, which it may have cut.

Pollen characteristics: The AP value was low of which 0.8% was Pinus pollen. Gramineae and Ericaceae were both relatively high. Furthermore the Plantago lanceolata, the Ranunculaceae and the Compositae liguliflorae percentages were relatively low. Cerealia pollen was present along with 7% Potentilla pollen. Correlation: The pollen contents of this sample were not very characteristic of any phase. The combined occurrence in relatively high frequencies of Ericaceae and Gramineae, together with the presence of Cerealia pollen and low pine values tentatively suggested a correlation with zone CaN-3c but other correlations were possible. Archaeological comment: The relationship to

2024

385

Post hole with infill of black, greasy clay with a lens of OGS in the fill half way down (15cm deep). The post hole underlay 369 (base of 315). Over 388, possibly bank material.

Pollen characteristics: Low AP value. NAP was dominated by Gramineae (50%) and Ericaceae (33%). Cerealia pollen was present. Plantago lanceolata pollen was comparatively low. Correlation: This sample was most likely to belong in zone CaN-3c. Archaeological comment: This pit was found once the basal layer 369 of plough soil 315 was cleared away. Its attribution to CaN-3c contrasts with the attribution of a sample 2078 from the overlying plough soil 315 to CaN-3b. The pollen contents were also in contrast to those of Pit 386 which was in an apparently similar stratigraphic position. However these apparent anomalies fits adequately within the favoured interpretation that the area south of the cairn was used for placing small amounts of burial-related material (sometimes in pits and sometimes in small dumps and Palaeoenvironment \ 857

Calanais Survey and Excavation, 1979-88

in one case in a small stone setting) over a long period during which the ground was worked-over several times, ending in a period of cultivation. The lens of old ground surface in the backfill of the pit need only mean that it was cut through an old ground surface and re-filled with the material through which it had cut.

Sample 2010 showed extremely high Gramineae values while the Plantago lanceolata values were low and Ranunculaceae were absent. Sample 2041 contained AP values of 9.5% and the Gramineae were comparatively low (54%). Table 21.11 Spot samples from layer 360

Table 21.10 Spot sample from posthole 386 Area Sample Feature Field interpretation

2028

386

Post hole with black, greasy clay fill. Under plough soil 369 (315) Cuts chocolate brown layer 377 and cultivation bed 388

Pollen characteristics: Comparatively high AP values (13%), dominated by the Corylus. Pinus pollen was relatively abundant (2.2%). Both Ericaceae and Gramineae plentiful, 36% and 34% respectively Cerealia pollen was present together with relatively high Compositae liguliflorae values (3.4%). Correlation: This sample correlated with the beginning of zone CaN-3a and was comparable with sample 2031 except for the presence of Cerealia pollen. Archaeological comment: This zonation is at first sight surprisingly early, but the pit fill was very probably material dug up to create the pit and may have included part of a turf line equivalent to that which produced sample 2031. 21.5.5 Spot samples from D under the cairn Pollen characteristics: Samples 2010 and sample 2041 differed appreciable from the others. The following characteristics are valid for most of the samples. The AP values were relatively low. The NAP percentages were dominated by Gramineae, and Ericaceae were relatively low. Plantago lanceolata pollen was abundantly present. Furthermore Ranunculaceae and sometimes Trifolium pollen were present.

Area

Field interpretation: A layer of greasy orange/ brown clay with quartz chips in the cairn near the outer Sample Context face of the chamber wall which appeared to have been deliberately laid for the overlying course of cairn boulders.

2042

2043

360

underneath stone U

underneath stone Y

360

2045

360

2052

360

2056

360

2057

360

2010

360

2041

360

related to sample 2056

underneath stone Y

underneath stone Q

underneath stone V

sealed by a stone in the cairn

Correlation: The low AP values occurred from the clearance in phase CaN-3a onwards. More specifically, the low Ericaceae values, combined with extremely high Gramineae values and an abundant presence of Plantago lanceolata and Compositae liguliflorae pollen, pointed to correlation with the second stage of zone CaN-3a. During this stage there was an emphasis on cattle breeding (see also the presence in the samples of Ranunculaceae and Trifolium pollen) while agriculture was less important or practised further away from the site. The absence or near-absence of Pinus pollen in the samples was remarkable, but in the main column CN3 Pinus was less frequent in the samples Palaeoenvironment \ 858

Calanais Survey and Excavation, 1979-88

ascribed to the second half of zone CaN-3a and only recovered during zone CaN-3b. The regional Pinus decline was registered in zone CaN-3c. Sample 2010 possibly belonged to the same stage of zone CaN-3a but due to the over-representation of Gramineae pollen the percentages of the other species were suppressed and comparison was made difficult. Sample 2041 possibly lay somewhat earlier in time because Corylus pollen was relatively abundant, which was characteristic of the first half of zone CaN-3a. Archaeological comment: this clay was presumably scraped up from earlier clays in the surrounding area early in CaN-3aii to provide a bedding layer for the basal stones of the cairn. The two seemingly anomalous samples may reflect different earlier deposits. Table 21.12 Spot sample from Context 910 Area Sample Context Description

2018

910

Silty clay from the base of the cairn, above the green clay. Excavated out of sequence and excavation not completed.

Pollen characteristics: relatively low AP values, but Pinus remained at levels over 1% and Corylus was relatively abundant. Correlation: the sample can be placed straight after the tree clearance in the middle of zone CaN-3a where Ericaceae increased at the expense of Gramineae. Archaeological comment: This fits well with the zonation of most of the samples from 360 and increases the likelihood that the cairn was built in CaN-3aii Table 21.13 Spot sample from clay 356 Area

Sample Context Field interpretation

2059

356

Taken from the clay between the stones at the base of the chamber wall

Pollen characteristics: The AP values were comparatively high (13%), of which 2.6% was Pinus pollen and 4.2% was Corylus pollen. Within the NAP Gramineae were dominant (54%) and Ericaceae were relatively low (14%). Cerealia pollen occurred in the samples. Plantago lanceolata and Compositae liguliflorae were plentiful. Correlation: The relatively high Pinus values in this context pointed to an early stage within zone CaN-3a, but later in time than sample 2031 because Corylus pollen was less frequent. By way of contrast, the presence of 3.6% Ranunculaceae pollen and Trifolium pollen suggested a correlation with the second half of zone CaN-3a during which there was an emphasis on cattle breeding and the curves of Gramineae, Plantago lanceolata, Ranunculaceae and Trifolium increased. Archaeological comment: This is interesting because it could be used to argue that the chamber wall was built significantly earlier than the outer cairn where basal layer pollen from 360 belonged in CaN-3aii. However the many samples from layer 360 included a sample 2041 belonging to CaN-3ai so it would not be safe to assume that sample 2059 proves that the wall and chamber were built at different times. Table 21.14 Spot sample from cairn layer 362 Pollen characteristics: The AP percentage was very low and the NAP value was dominated by Gramineae (63%) and Ericaceae (28%). Compositae liguliflorae pollen was relatively abundant in this sample. Correlation: The very low AP value together with an abundant presence of both Ericaceae and Gramineae pointed to a correlation with the CaN-3b to CaN-3c transition, where Cerealia Palaeoenvironment \ 859

Calanais Survey and Excavation, 1979-88

pollen was sometimes absent from the samples and where Gramineae were relatively high.

2073

365

as 2065

Archaeological comment: This is by far the latest zonation for any layer associated with the cairn. The transition from CaN-3b to 3c was dated to between 1900 and 1600 cal BC at Callanish Leobag. Particularly given its proximity to modern topsoil it seems most likely that it should be associated with the dilapidation of the cairn which took place during the period when the area south of the cairn was being used for deposition of burial material and ploughing.

2039

365

as 2065

Area Sample Context Field interpretation

2083

362

This context underlay topsoil. It consisted of a dark brown gritty clay with much charcoal in fill of cairn overlying a layer of dark brown/black clay bedding (370), which in turn overlay the green clay capping (373) over stone paving around stone 29 clay near to central monolith.

21.5.6 Spot samples from old turf lines on Area D Table 21.15 Spot sample from bifurcating old ground surfaces 365 Area Sample Context Field interpretation

2065

365

This label was given to a sequence of (commonly two) turf lines found to the north of cultivation bed 388.

D D

2069

2072 2075

365

365 334

as 2065

Very similar old turf lines to the north of 388

Pollen characteristics of all the samples except 2039: The AP values were low but varied around 10%. Pinus pollen was relatively abundant (between 1.2 and 2.0%). Both Betula and Corylus pollen was comparatively frequent. Ericaceae was high, while Gramineae varied between 23 and 30%. Cerealia pollen was absent. Plantago lanceolata, Ranunculaceae and Compositae liguliflorae were present. Potentilla was present in sample 2072. Pollen characteristics sample 2039: The AP percentage was low (4.6%). Pinus pollen occurred but did not reach more than 1%. Ericaceae pollen was abundantly present. Gramineae reached 27 to 30%. Cerealia pollen was present and Plantago lanceolata, Ranunculaceae and Compositae liguliflorae occurred. Correlation of all the samples except 2039: The presence of Pinus pollen in percentages more than 1%, together with an absence of Cerealia pollen and a relative abundant presence of Ericaceae pollen, while Plantago lanceolata, Ranunculaceae and Compositae liguliflorae were present was characteristic of the pollen assemblages occurring at the end of zone CaN-3a, transition to zone CaN-3b. At this level the agricultural activities declined and Ericaceae increased. Sample 2039 from the lower part of the sequence of OGS also seemed to date to the end of zone CaN-3a but perhaps slightly earlier than the other samples Archaeological comment: The (commonly two) turf layers ran together in places and were very hard to distinguish during excavation. The turf lines underlay ginger layer 320 and the samples Palaeoenvironment \ 860

Calanais Survey and Excavation, 1979-88

in Kubiena boxes 4-5 and 6-7 suggested that they should date to CaN-3a (some to CaN-3ai and some to CaN-3ii). In places the two turf lines were separated by a thin layer of clay (see fig 21.9 above). The turf lines in the middle of Area DI were truncated by later ploughing. The pollen zonation suggested that it was the upper old ground surface which was sampled, and that the upper turf line continued to grow after the cairn was built until between 2560 and 2200 cal BC. 21.5.7 Spot sample from the late cultivation soil on Area D Table 21.16 Spot sample from plough soil 315 Area Sample Context Field interpretation D

2078

315

Plough soil with ard marks

2077

359

A sample from an ard mark cut into reddish orange gritty clay 359 (=320) under 344 (=315) late ploughing level. Several ard marks were visible in this layer.

Pollen characteristics: AP percentage was low (5.4%) Both Ericaceae and Gramineae were abundant (45% respectively 36%). Cerealia pollen was absent but Plantago lanceolata (5.2%) and Compositae liguliflorae (1.4%) were present. Correlation: This sample possibly correlated with zone CaN-3c where both Ericaceae and Gramineae were high and the Pinus curve declined under 1%.

Pollen characteristics: The AP values were low (7%) and dominated by Corylus pollen (4.6%). The NAP was dominated by Gramineae and Ericaceae was relatively low. Cerealia pollen was absent. Compositae liguliflorae formed 1.8% of the pollen sum correlation: The high Gramineae values combined with low AP percentages, and the fact that the Plantago lanceolata reached 5.4%, suggested it most likely that the sample correlated with zone CaN-3b. During zone CaN-3c Ericaceae values were much higher and correlation with this zone was impossible.

Archaeological comment: This correlation suggests that that ploughing continued after the transition from CaN3b to 3c at some time between about 1880 and 1520 cal BC.

Archaeological comment: This zonation fits well with the interpretation of the general sequence on Area D. The lack of cereal pollen fits the archaeological interpretation that the ground working on Area D, which included ploughing with an ard, was not related to cultivation but was part of a tradition of dealing with small ritual deposits made in the southern half of the Ring outside the chambered cairn.

Table 21.17 Spot sample from ard mark 359 Area Sample Context Field interpretation

21.5.8 Spot samples from late contexts on Area D Table 21.18 Spot sample from Victorian backfill 331 Area Sample Context Field interpretation 2015

331

OGS Clods of grey humus within a Victorian intrusion.

Pollen characteristics: The AP percentage was low (4.6%). Pinus pollen occurred but did not reach more than 1%. Ericaceae were abundantly present. Gramineae reached 27 to 30%. Cerealia pollen was present and Plantago lanceolata, Ranunculaceae and Compositae liguliflorae occurred. Correlation: These samples had many characteristics in common with the samples of the middle “turf-line”; but Pinus pollen was less frequent and Cerealia pollen was present. A correlation Palaeoenvironment \ 861

Calanais Survey and Excavation, 1979-88

Illus 21.31 Contexts spot-sampled for pollen in Area B with the upper turf line seems unlikely due to the comparatively low Gramineae values and the relatively high presence of Compositae liguliflorae pollen. Possibly this sample correlates with the lower turf-line which should occur at the end of zone CaN-3a Archaeological comment: These clods were dug up by Victorians to help back-fill a sondage into the cairn. No traces of the characteristics of turf lines were noted in the field records. Perhaps a match with a much later zone in the CN2 column at Callanish Leobag or in the pollen columns from Calanais Fields should be explored. Table 21.19 Spot sample from chamber clay 355 Area Sample Context Field interpretation D

2081

355

Redeposited green clay. in the chamber

Pollen characteristics: The AP value was very low (1.8%). The NAP percentage was formed mainly by Gramineae (26%), Ericaceae (35%) and Cyperaceae (25%). The relative abundant occurrence of Cyperaceae had no comparison with other samples from the excavation site, but could have resulted from the very local occurrence of sedges in the vegetation. Correlation: The very low AP value together with the low values for Compositae liguliflorae and Plantago lanceolata, made a correlation with zone CaN-3c likely. Archaeological comment: The 1857 clearance of peat led to removal of black (burial) deposits from the chamber and it looked during excavation as if the green clay 355 was redeposited after the Victorian intrusion. The suggested correlation is difficult to explain. Perhaps a match with a much later zone in the CN2 column at Callanish Leobag or in the pollen columns from Calanais Fields should be explored. Palaeoenvironment \ 862

Calanais Survey and Excavation, 1979-88

21.5.9 The stratigraphy of the spot samples from Area B Please see illustration above. 21.5.10 Spot samples from Area B Table 21.20 Spot samples from slot 883 Area Sample Context Field interpretation

2009

883

2011

883

2020

883

Dark grey/green gritty charcoal rich material with burnt bone and rotted stone underneath the bottom course of the stones in the passage. This was part of the slot cut into the green clay platform under the cairn and contained charcoal much older than the date of the context.

As sample 2009

Table 21.21 Spot sample from passage pit 886 Area Sample Context Field interpretation

2070

889

Round pit and green clay upper fill from the middle of the passage in an ambiguous relationship with the green levelling under the cairn

Table 21.22 Spot sample from passage pit 881 Area Sample Context Field interpretation

2022

882

Brown loamy soil with charcoal and poss. hazelnut shells within a dark feature (881) under the north passage wall.

Pollen characteristics: The AP values were relatively low and Pinus pollen was absent or only present in low percentages (0.2%). The NAP values in all the samples were dominated by Gramineae. Ericaceae varied between 7 and 15% and the Cyperaceae were virtually absent. Plantago lanceolata pollen was not very frequent in the samples except for sample 2020 (11.4%). Compositae liguliflorae pollen was relatively abundant in the samples but again sample 2020 formed an exception. Here by way of contrast the Ranunculaceae were frequently encountered in the sample. Furthermore sample 2011 contained relatively high Corylus percentages. In sample 2022 the AP values were low and NAP was dominated by Gramineae. Ericaceae was low and Plantago lanceolata was present (respectively 4% and 5%). Furthermore Compositae liguliflorae was relatively abundant. Correlation: The relatively low AP and Ericaceae values together with an abundant presence of Gramineae made a correlation with two different zones possible, zone CaN-3a and zone CaN-3b. During zone CaN-3b the Pinus values generally recovered to more than 1% and Compositae liguliflorae pollen was less frequent. Also in the spot samples from area D that are correlated with zone CaN-3a low Pinus values do occur. The samples hence reflect a stage straight after the clearance of birch wood in the middle of CaN-3a. In some of the samples the decline in the Corylus was still visible (e.g. 2011). Sample 2020 should possibly be placed somewhat later in time during zone CaN-3aii where there is a tendency towards an emphasis on cattle breeding (high Plantago lanceolata and Ranunculaceae values). In sample 2022 the extremely high Gramineae values combined with the presence of Plantago lanceolata and Compositae liguliflorae made a Palaeoenvironment \ 863

Calanais Survey and Excavation, 1979-88

correlation with zone CaN-3a most likely, although the AP percentages were low. The sample even contained Cerealia pollen which would place it in the first half of zone CaN-3a, where there was an emphasis on agriculture.

Pollen characteristics: The AP values were low and the NAP was dominated by Gramineae. Ericaceae values were low and Plantago lanceolata was present (respectively 4 and 5%). Furthermore Compositae liguliflorae was relatively abundant.

Archaeological comment: Most of the samples came from a slot 883 dug through the green clay platform under the cairn and associated with a slight wooden structure. The dating to the start of CaN-3aii is entirely credible given that the cairn itself seems to have been built after 2500 cal BC. Sample 2022 came from shallow pit fill 882 overlying slot 883 in the passage. The pit should date to zone CaN-3aii, but the feature may have been partly or wholly filled with the older material through which it was dug. Sample 2070 came from a pit 889 dug in the clay forming the base of the passage. It was filled with turfs and stones. Its stratigraphic relationship to 883 was ambiguous, as was its relationship to the green clay platform under the cairn because of erosion in the passage. The similarity of its pollen (sample 2020) to that in 883 suggests it too should post-date the green clay levelling into which 883 was cut, although the higher value for Plantago lanceolata and the lower value for Compositae liguliflorae pollen suggest that Pit 889 and slot 883 were not of exactly the same date as each other. One possible interpretation is that the pit was dug first, and rapidly back-filled, while the slot was not finally filled until the slight structure based in it had decayed or been removed. Another is that, given the by then ancient charcoal incorporated in slot 883 the pollen in it was of mixed age.

Correlation: The extremely high Gramineae values combined with the presence of Plantago lanceolata and Compositae liguliflorae make a correlation with zone CaN-3a most likely, although the AP percentages were low.

Table 21.23 Spot sample from sandy layer 812

Correlation: The abundant presence of Compositae liguliflorae and Gramineae, while Ericaceae were relatively high, was a feature of zone CaN-3a. The high Ranunculaceae and the absence of Cerealia pollen point to a correlation with zone CaN3aii, when agriculture was declining around the site and cattle breeding became more important.

Area Sample Context Field interpretation

2054

812

A distinct even sandy green layer with some pieces of charcoal which appeared to stretch right across trench BV.

Archaeological comment: Layer 812 seems to represent spreading of green clay between two episodes of deposition of burial-related material and ground-working. Its pollen contents contrast with those of an overlying dark plough soil 141, which was much richer in heather, but the underlying dark layer 160 was not analysed for pollen. Table 21.24 Spot sample from kerbstone pit in passage 859 Area

Sample Context Field interpretation

2050

859

redeposited material in facade stone pit in passage

Pollen characteristics: AP value was low. Gramineae were relatively abundant (53%) and Ericaceae reached 16%. Plantago lanceolata, Compositae liguliflorae and Ranunculaceae were relatively abundant. Cerealia were absent from this sample.

Archaeological comment: The zonation fits the archaeological interpretation.

Palaeoenvironment \ 864

Calanais Survey and Excavation, 1979-88

Table 21.25 Spot sample from charcoal spread 139 Area

Sample Context Field interpretation

2021

2030

139

Charcoal spread over soil, under a rough line of field stones (134) running from Ring stone 44 alongside the East Row

as Sample 2021

Pollen characteristics: The AP values were low (4%). The NAP values were dominated by Ericaceae and Gramineae. The Plantago lanceolata reached high values (10.8 and 11.9%). Compared to the Ranunculaceae and Compositae, liguliflorae, Cruciferae pollen was relatively abundant. Correlation: Both samples could be compared with zone CaN-3a but Ericaceae values were relatively high for this zone. A combination of both cattle breeding indicators (Plantago lanceolata, Ranunculaceae and Cruciferae) together with indicators for agriculture (Cerealia and Compositae liguliflorae) as is the case in sample 2021, points to a correlation with zone CaN-3c. Sample 2030 lacked Cerealia pollen but contained more Gramineae and Pinus pollen. This difference between the two samples can be explained by assuming that sample 2030 lay somewhat earlier in time, possibly the transition of zone CaN-3b to zone CaN-3c, before the regional pine decline. Archaeological comment: The ascription of 2021 and 2030 to a relatively late pollen zone in CaN-3c suits the archaeological interpretation well, and there is no reason why the layer should not include diachronic material if, as suspected, it was not thoroughly mixed but a partially groundworked set of dumps. Table 21.26 Spot samples from turf line and ard marks on Area B Area Sample Context Field interpretation

2013

806

2063

806

2029

809

2023

809

2019

800

Turf line (grey lens) in ard-marked area BINX ard mark X

ard mark A in a group of ard marks in grey sand cut into yellow sand in BINX. ard mark B under F117

ard mark from BIII, a different area from the others here.

Pollen characteristics: Of these samples 2013, 2023 and 2029 and 2063 can be compared with sample 2021 from the late (CaN-3c) charcoal-rich soil spread 139. Sample 2029 contains relatively high Pinus values and lacks Cerealia pollen, which makes this sample more comparable with sample 2030 from soil 139. Correlation: Sample 2029 is likely to correlate with levels shortly after the transition of zone CaN-3b to zone CaN-3c, clearance at a stage just before the regional pine decline. The samples 2023 and 2013 both correlated with zone CaN-3c and possibly with an early stage during zone CaN-3c because of the relatively high Corylus values that are present in these samples. Sample 2063 contained material from an ard mark cut into the same feature as 2013 but Ericaceae and Gramineae values differed appreciable. Also sample 2063 lacked Cerealia pollen and Cruciferae, which made it difficult to compare with any of the other ard mark samples. Sample 2019 from context 800 lacked both the high Pinus value and relatively high Corylus percentage of the others but contains Cerealia pollen up to 0.8% and relatively high Cruciferae value. It too correlates with zone CaN-3c, but reflects a stage somewhat later in time where the Cruciferae have increased (see also column CN3). Archaeological comment: Ard mark 800 was from a completely different part of the area and Palaeoenvironment \ 865

Calanais Survey and Excavation, 1979-88

may have a completely different history from the ard marks cut into 806 and 809. The ard marks may be of various archaeological phases, but their general assignation to CaN-3c fits the overall interpretation of Area B. Table 21.27 Spot sample from groove 121 Area Sample Context Field interpretation B

2014

121

Linear groove under 117 = natural brown fibrous layer. Within 117.

Pollen characteristics: The AP values were very low. Gramineae and Ericaceae formed the major contribution to the NAP values. Cerealia pollen was present in all the samples. Plantago lanceolata was relatively high (9.8%) as were Cerealia (0.8%). Cruciferae, Ranunculaceae and Compositae liguliflorae were present. Correlation: Sample 2014 shows many comparisons with sample 2019 and can be placed in zone CaN-3c, after the regional pine decline. Archaeological comment: Context 121 is from under or within brown fibrous layer 117 interpreted as a modern soil development within prepeat soils, and an allocation to CaN-3c or later is entirely credible. Table 21.28 Spot sample from plough soil 141 Area Sample Context Field interpretation B

2085

141

2086

141

From a plough soil in Area BIV to the east of the cairn. As Sample 2085.

Pollen characteristics: The AP values were very low. Gramineae and Ericaceae formed the major contribution to the NAP values. Cerealia pollen was present in both samples. Cruciferae, Ranunculaceae and Compositae liguliflorae were present. Compositae liguliflorae was very high.

Correlation: The high Compositae liguliflorae percentages present a problem with the proposed correlation with zone CaN-3c. Archaeological comment: Context 141 was interpreted as equivalent to the late plough soil 315 on Area D and its assignation to CaN-3c matches the pollen evidence from Area D and fits the archaeological evidence from Area B. 21.5.11. The spot samples from Test Pit J and from Area C Test Pits G2 and J were dug in 1981. Although there were some ambiguities in labelling of the samples it is most likely that no samples from G2 or the 1980 Test Pit G1 were analysed for their pollen. Test pit J was in an area where no cultivation beds were visible. Area C was excavated to investigate a fallen stone, the pit in which it had been erected and any related archaeology. The samples taken for pollen analysis were a Kubiena box column from a temporary baulk across the fallen stone and the basal peat spot sample from the east baulk described below Table 21.29 Spot sample from soil test pit J Area

Sample Context Field interpretation

Probably G3=J; see 2060 Technical Note 13.4.2)

1981 Soil test Layer-3 pit horizon NW profile.

Pollen characteristics: The AP value reached 20.6%. Betula pollen formed the major contribution to the AP percentage (13.4%), Corylus 5.4%. Gramineae were relatively high (56%), while Ericaceae only reach 5%. The Compositae liguliflorae were relatively abundant (6.2%). Plantago lanceolata and Ranunculaceae pollen were present, the Cerealia were absent. Palaeoenvironment \ 866

Calanais Survey and Excavation, 1979-88

Correlation: The relatively high AP percentages, mainly birch, and high Gramineae values indicate a correlation with an early stage in zone CaN-3a during which the Betula values were still declining and hazel was present. This sample falls straight after the clearance which marks the transition of zone CaN-2d to CaN-3a. Archaeological comment: If this is from G2 Layer 3 as both Bohncke and the original Soil Sample list have it the sample is from black peat. The overlying layer was black peat with some quartz grains and charcoal. The assignation to just after the CaN-2d to CaN-3a is slightly surprising given the !st millennium BC date of the basal peat on the other side of the stone setting and at the Calanais Fields project where peat grew, in round terms, from 500 BC to 800 AD and from 250 BC to after 1700 AD. But Pit G2 was on what appeared from air photographs CUP RA84 and RA85 of 1955 to have been cut into a squarish pre-modern cultivation area. Peat and other material may have been dug up elsewhere and dumped on this area and modified in the 19th century AD. If this sample is actually from G3 (=J) as the AOC Macroplant Bag record has it, it was from a very compact green brown silty clay. But that would mean an inversion because the pollen in Sample 2036 from Layer 4 is assigned to the transition of zone CaN-3a to CaN-3b. That too would be somewhat surprising.

Area

Sample Context

G3 = J

2036

Layer-4

Layer 4 is described as uncemented green subsoil.

Field interpretation

B1g horizon in the pit.

Pollen characteristics: Very low AP percentage. NAP was dominated by Ericaceae while Gramineae were relatively low (29%). Plantago lanceolata pollen occurred in low frequencies. Ranunculaceae, Compositae liguliflorae and Cerealia pollen were present. Correlation: The strong increase in Ericaceae in this sample suggests a correlation with the transition of zone CaN-3a to CaN-3b when the human impact temporarily declines. Archaeological comment: A B1g layer (sample 2036) is the gleyed form of an illuviated subsoil in which fine material has derived from organic-rich higher layers. That is not incompatible with the field description of Pit J Layer 4 as uncemented green subsoil. Table 21.31 Spot sample from Area C

Table 21.30 Spot sample from soil test pit J Area

Illus 21.32 The location of the test pits and Area C

Sample

275?

Context

201

Field interpretation See Archaeological comment below

Pollen characteristics: 6.2% AP. NAP was dominated by Ericaceae and Potentilla-type pollen. 0.4% Cerealia was present. All the other herbs occur in low percentages. Correlation: The high Potentilla values, in combiPalaeoenvironment \ 867

Calanais Survey and Excavation, 1979-88

nation with the other pollen characteristics make this sample fit in zone CaN-3e (see also Area C Kubiena box pollen diagram). Archaeological comment: This sample, recorded as Sample 201 in the draft pollen report, is very probably sample 275 from context 201, basal amorphous peat on Area C, taken as a possible C14 sample. Another basal peat sample Call80/54/244 produced a date (GU-1403 2640+/-110 BP) between 1050 and 400 cal BC from humic acid. The basal peat sample age from Area C is not significantly different from two Zone 3e ages from the Leobag CN1 column when they are considered as a group. The two Zone 3e peat sample ages were GU-1289: 2440+/-112 BP and GU-1170: 2355+/-110 BP which dated to between 850 and 200 cal BC and to between 800 and 150 cal BC respectively. Interpolation between those dates and earlier dates suggests that the 3d to 3e transition took place between 920 and 260 cal BC (Table 21.2). Combining this information with that from the basal peat date suggests unrestricted peat growth started between about 920 and 400 cal BC. 21.5.12 Spot samples from Area H Area H was dug in 1981. It was designed to explore the cairn and allow repairs to the chamber wall. The old ground surfaces to the north of the cairn were complex and truncated, and the stratigraphic correlation between those at the north end of the site and those close to the cairn was ambiguous. Table 21.32 Spot sample from token burial 736 Area Sample Context Field interpretation

2026

736

Reddish-yellowish clay with a few fragments of burnt bone (Day Book 6) in an area of discoloured stone in the outer cairn.

Pollen characteristics: 15.6% AP of which 10.8% was Corylus. Gramineae dominated in the NAP and Ericaceae reached 25% Cerealia pollen was absent from the sample. 2.2% Pinus was present. Correlation: The relatively high Corylus and Pinus percentage, together with a dominant occurrence of Gramineae in the NAP are characteristics of the beginning of zone CaN-3ai. Archaeological comment: This is an extremely interesting context, a possible token burial deposit’. It contained a rim sherd of AOC beaker, a ‘neolithic’ sherd and a single piece of hazel charcoal (sample 2243) produced an age AA-24967 (4050+/-45 BP) which calibrated to between 2860 and 2470 cal BC. That is somewhat too early for any beakers in Britain or on nearby parts of the continent. The analysis suggests a zone, CaN-3ai, dated to much the same period as the cairn but too early for the sherd. The archaeological interpretation of the complex evidence from stratigraphy, radiocarbon and artefacts is that the cairn must have been built fairly soon after 2500 BC and the token burial must have contained pollen-bearing material old at the time. Table 21.33 Spot sample from slot fill 730 Area Sample Context Field interpretation H

2048

730

Charcoal-rich greyish green brown sandy clay within slot 773 cut into the green clay platform.

Pollen characteristics: 9.4% AP of which 3.2% was Corylus. The NAP was dominated by Gramineae (62.8). Ericaceae was at 13.8%. Cerealia were present. Correlation: quite like sample 2027 from context 735; probably CaN-3ai, like sample 2027 Archaeological comment: The context was a slot cut by erosion in the chamber. It was filled before the cairn was built.

Palaeoenvironment \ 868

Calanais Survey and Excavation, 1979-88

Illus 21.33 Contexts spot-sampled for pollen in Area H Table 21.34 Spot sample from slot fill 778 Area Sample Context Field interpretation

2051

778

Slightly gritty, humic rich clay with ochre from a slot underneath the cairn; cutting the green clay platform. The fill included charcoal of earlier age.

Pollen characteristics: The AP was relatively high (12.8%) the majority derived from Corylus and Betula. Pinus pollen was present (1.4%) within the NAP Ericaceae were present in relatively low values (14.4%). Gramineae reached 58%. Cerealia pollen was absent. Correlation: The relatively high Corylus and Pinus values point to a correlation with the middle part of zone CaN-3a close in time to sample 2026 (CaN-3ai).

Archaeological comment: The pollen could go with the date of filling of the context or with clay of the same date as the charcoal (from the same sample 2051 and context 778: A charred hazel nut shell (sample 2051) provided a date SUERC-11617; (4425+/-35 BP) of between 3330 and 2920 cal BC. Although the zone CaN-3i is only loosely dated it does appear to be significantly later than the dates quoted above. Therefore the pollen in the clay may have been dominated by pollen of the period when the context was filled. One possible implication, if the cairn was, as argued above, built after 2500 cal BC, is that the light structure built on the platform was significantly earlier than the cairn. In that case the clay in the token burial deposit sampled by Sample 2026 might have included clay picked up from the area of the (by then destroyed) light timber structures. Table 21.35 Spot sample from kerb trench 735 Palaeoenvironment \ 869

Calanais Survey and Excavation, 1979-88

Area

Sample Context Field interpretation

2027

735

Dark grey brown gritty clay in later kerb slot

Pollen characteristics: 9.8% AP of which 7% was Corylus. The NAP was dominated by Gramineae (69%). Cerealia were present in this sample (0.8%). Correlation: This sample contains a pollen assemblage with all the characteristics of the first half of zone CaN-3a, the stage during which there was an emphasis on agriculture. The only problem with this correlation is that the Compositae liguliflorae could be expected to have shown higher values. Archaeological comment: The interpretation of the context was fairly ambiguous and the pollen zonation does not resolve the ambiguities. The pollen reflects the material through which the slot was cut rather than pollen characteristic of the period when the slot was dug. The implication is that the work was done when there was little pollen rain, as for example during winter. Table 21.36 Spot sample from turf line 751 Area

Sample Context Field interpretation

2012

751

turf line north of the cairn

Pollen characteristics: Low AP value, Pinus pollen absent equal percentage Ericaceae and Gramineae (43%). The Compositae liguliflorae are relatively high. Correlation: The relatively high Ericaceae value together with an absence in the Cerealia and in the indicators for pasture land (Ranunculaceae and high Plantago values) indicates a correlation with the middle turf-line at the transition of zone CaN-3a to zone CaN-3b. Archaeological comment: This zonation is inexplicable unless the sample was taken from an area where the turf line was not covered by imported green clay 750 in the middle of CaN-3a and con-

tinued to grow until the transition to CaN-3b at some time between 2560 and 2200 cal BC. This will be discussed further with Samples 2076, 2080 and 2082 below. Table 21.37 Spot sample from chamber wall fill 770 Area

Sample Context Field interpretation

2071

770

Very dark grey clay lower infill of chamber wall.

Pollen characteristics: This sample showed relatively high AP values of which Betula and Corylus contributed the most. Only 0.6% Pinus pollen was present. 47% Gramineae and 40% Ericaceae were present in this sample. The relatively high Ericaceae percentages are remarkable; sample 2071 would be very comparable to sample 2061 from the turf line immediately below 771 if it wasn’t for them. Correlation: Likely to correlate with an early stage in zone CaN-3a but the high Ericaceae possibly indicated a development towards the middle “turf-line” at the transition of Zone CaN-3a to CaN-3b. Archaeological comment: The chamber wall was built at the start of CaN-3aii. The clay in the wall contained highly residual potsherds. The high Ericaceae values represent an evanescent event or use of imported material zoned to CaN-3ai. Table 21.38 Spot sample from turf line 77 Area

Sample Context Field interpretation

2061

771

Turf line underneath the chamber wall and above the green clay platform

Pollen characteristics: Relatively high AP value (13.4%) mainly formed by Betula and Corylus pollen. Gramineae were abundantly present (55%) while Ericaceae values were low (16%). Palaeoenvironment \ 870

Calanais Survey and Excavation, 1979-88

Correlation: The relatively high AP value combined with the high Gramineae percentage and low Ericaceae percentage points to an early stage in zone CaN-3a. Archaeological comment: This thin OGS-like layer is of considerable interest because it was either an incipient turf line before the cairn was built or composed of material which silted / washed down through the cairn. The overlying clay layer 770 of the chamber wall was zoned to CaN-3a to 3b in spot sample 2071 and although that was probably later than the date of cairn building it was sufficiently different from this sample to suggest that 771 was indeed an incipient turf line forming on the green clay platform under the cairn. Spot sample 2061 came from the turf line immediately overlying the green clay under the cairn. The implication of the pollen evidence from the box column and spot samples is that the green clay platform was laid late in CaN-3ai and the overlying cairn was built early in CaN-3aii which fits the archaeological interpretation that the cairn was built soon after the middle of the millennium fairly well. Table 21.39 Spot samples from ard mark 734 and plough soil 707 Area Sample Context Field interpretation H

2076

734

2082

707

2080

707

Ard mark 3; Broad ard mark light green gritty sandy clay. Cuts 707= plough soil (ard mark layer) plough soil a slightly greasy clay with charcoal flecks and rotted stones at a higher level than OGS 771

as 2082

Pollen characteristics of 2076 and 2082: The relatively high AP values were mainly caused by the relatively high Corylus percentage. Both Gramineae and Ericaceae were relatively abun-

dant, although Gramineae were twice as frequent. Cerealia was absent from both samples. Compositae liguliflorae was relatively abundant like the Ranunculaceae. Pinus pollen was present in low values. By way of contrast sample 2080, from the same plough soil as 2082, had: low AP values, and 1.2% Pinus. But as in 2082 Corylus pollen was relatively frequent (4.6%). Gramineae pollen was very abundant (64%). Compositae liguliflorae pollen was frequently encountered in the sample. Cerealia pollen was present. Correlation of 2076 and 2082: The high AP values point to a correlation with the earlier half of zone CaN-3ai where Corylus showed a brief increase. The relatively high Compositae liguliflorae values seem to confirm this correlation. Sample 2082 seems somewhat later than 2076 judging from the declining Corylus values. Correlation of 2080: After the successive Betula and Corylus decline near the middle of zone CaN-3a a stage follows during which Gramineae increase together with the Compositae liguliflorae and to a lesser extent the Plantago lanceolata. This sample fits nicely in mid zone CaN-3a. Archaeological comment: Mid CaN-3a is surprisingly early for the ploughing which should be late Can 3a, CaN-3b or later. The ard mark 734 cut the plough soil 707 in a way that suggests that it was significantly later than the plough soil, but the pollen from the ard mark had earlier characteristics. The pollen must have come from the layers cut by the ard rather than being contemporaneous with ploughing. A likely source for mid CaN-3a pollen is the clay fills of the cairn. It was robbed of stones and one scenario supported by the pollen inversion is that cairn material was pulled down onto this area and the clay fill component left when the stones were removed formed part of the plough soil 707. Perhaps also the pollen may have derived partly from green clay 750 which was imported to this area towards the end of CaN-3ai (Illus 21.34). These contexts are stratigraphically higher and significantly later than turf line 751 (sample 2012), Palaeoenvironment \ 871

Calanais Survey and Excavation, 1979-88

which is placed at the CaN-3a /3b transition (Illus 21.33, 21.34). The zonation of the turf line appears anomalous. However a clue to possible resolution comes from the analysis of Kubiena box sampling on Area D. There what appeared during excavation as a pair of turf lines was shown to be far more complex because some of the turf lines lay on an old cultivation bed and others on deposits which had filled the trough between two cultivation beds. There were parts of two cultivation beds north of the cairn on Area H with a trough between them. Clay 750 lay in the trough and sealed most of turf line 758 which also survived best there. The turf lines in the north section may, as on Area D, have been more complex than the record suggests. There is thus an abstract possibility that the sample actually came from the higher turf line 771 (Illus 21.34). But only future excavations and spot and Kubiena box pollen sampling of the various turf lines and related strata north of the cairn can resolve this problem. 21.5.13 Discussion of the spot samples It is a pity to end on that note, because by and large the spot samples fitted in well with the interpretation of the excavation results. However, it is fair to say that the spot samples, unlike the Kubiena box profiles, rarely tested the archaeological interpretations or refuted alternative interpretations. For instance those relating to ploughing on Area H can be interpreted as representing old material pulled down from the cairn, which would incidentally help to explain where the soil came from. But a result in CaN-3b could have been explained equally easily and probably in much the same way but with more ploughing occurring. As Dr Bohncke remarked in his original draft, spot samples can reflect a very short period of time during which the pollen rain could have been different from the average for the century or so represented by each cm of the reference pollen column CaN-3a. There was also always a potential problem of residual pollen where human activities had disturbed earlier layers or had led to dumping of clay derived from early levels above â&#x20AC;&#x2DC;laterâ&#x20AC;&#x2122; strata. Nevertheless

the spot samples did often make a very significant contribution to interpretation, for instance showing that turf line 365 had continued to grow outside the cairn after it had been built. 25.6 Discussion of the pollen analysis from an archaeological perspective P J Ashmore Before the pollen analysis was conducted I and others were concerned that old pollen might have been mixed into later contexts, vitiating analysis. The truth probably is that the earlier contexts, including the early turf lines represented in the Kubiena boxes, did contain pollen representing the contemporaneous vegetation fairly accurately; but many of the later contexts dating to after the green clay platform under the cairn, and particularly the ground-worked soils of the late 3rd and 2nd millennia BC probably contained a significant component of residual pollen. One of the several examples which allow this fairly optimistic conclusion about the pollen in the earlier strata is provided by archaeological consideration of the pollen from a turf line751 in Area H, another part of which (758) was immediately below the green clay platform in a Kubiena box column. Here I quote from Chapter 12.7.4. â&#x20AC;&#x153;A pollen sample (2012) from turf line 751 had a high proportion of heather pollen (43%) which, together with an absence in cereals and in the indicators for pasture land led to an interpretation that it went with the transition of zone CaN-3a to zone CaN-3b at some time between 2560 and 2200 cal BC (Chapter 21: Pollen). No pollen samples from the overlying green clay 750 were analysed, but a spot sample from the overlying plough soil 707 had pollen characteristic of sub-zone CaN-3ai. This suggests either a reversal, with older pollen in plough soil 707 than in turf line 758, or mixing of pollen of various zones in the soil (highly likely, given that it was a plough soil) or local conditions producing a very different mix of plants to that found at the same time at Calanais Leobag. Of course, all three effects may have been present. But inspection of the pollen diagrams from Leobag suggests that there was a peak in heather of about the same percentage as Palaeoenvironment \ 872

Calanais Survey and Excavation, 1979-88

Illus 21.34Â Perspective section 13 [NMRS DC38285] in turf line pollen sample 2012 during CaN-3aii close to the junction with CaN-3ai, and that cereal was absent at the junction between CaN-3ai and 3aii [Chapter 21 Illus 21.3). It is entirely credible that the pollen in sample 2012 belonged around that time. Such a reinterpretation would accord with the pollen evidence from under the cairn where 758, the same turf line as 751, belonged to the junction between CaN-3ai and CaN-3aii.â&#x20AC;? What I found encouraging in analysing this and some other apparent problems from an archaeological perspective is that re-inspection of the reference pollen column at Leobag immediately allowed a natural interpretation where previously there seemed to be problems. Any future excavation at Calanais should employ a much more extensive and systematic sampling strategy. Pollen counts should be much higher. Degradation of pollen should be recorded (Tipping 2000). Very detailed recording and characterisation of turf lines, with input from soil scientists, will be essential. If possible, samples should be analysed before completion of a campaign of excavation to allow feedback to archaeological interpretation, and to allow supplementary sampling. However as it turned out, by and large, the pollen did not seem to share the deep residual-

ity which was such a common occurrence with pottery and charcoal at Calanais, with Phase 3 material in Phase 7 contexts. Archaeological interpretation and dating was usually straightforward. The Kubiena box samples were particular useful in discriminating between possible dates for the near-basal cultivation beds. More generally the zonal dating of these short profiles and of the spot samples was more in line with artefactual dating of the chambered cairn to after 2500 BC than were the radiocarbon dates. It might have been possible to relate the Kubiena box profiles to each other and the spot samples to the profiles without the nearby reference pollen columns. But they obviously helped dating and interpretation. Analysing more than one column seems essential to show how much local variation there was. The strategy for future columns should be analysis of contents and dates of ranging-samples followed if the preliminary data is promising by pollen sampling and radiocarbon dating at closer intervals than in the existing columns so that short term variations in pollen rain and peat growth rates can be captured better. If possible several columns covering the same period should be analysed to allow a better assessment of the catchment area.

873

Palaeoenvironment \ 873

21.7 Pollen tables for the spot samples Table 21.40Â Spot samples from B TRENCH NUMBER

BIV

BIV WX

BIVWX

B IV WX

CONTEXT

139

141

812

889

883

SAMPLE

2021

2030

2085

2086

2054

2070

2009

2011

dark soil

sandy layer

deep passage pit

PINUS SYLV

0.4

1.4

0.8

0.2

0.4

0.2

ALNUS GLUT

0.6

0.4

0.8

0.4

0.6

0.4

CONTEXT TYPE TAXA

BETULA

CORYLUS

QUERCUS TILIA

FRAXINUS ULMUS SALIX

SUM TREES % TREES

GRAMINEAE

0.8 0

0.2 0

20 4

26.6

1.8

0.2 0

0.2 20 4

48.4

0.4

3.6 0

5.6 56

0.2

1.8 0

0.2 0

2.8 58.4

1.4 1

3.2 82.2

3.8 1.5 0 0

5.3 65.5

passage slot passage slot

1.2

4.6 0.2 0

0 0

6.6 77.6

4.4 8.8 0 0

13.8 60.6 Palaeoenvironment \ 874

CYPERACEAE

5.4

4.2

0.2

0.5

0.2

0.4

CEREALIA

0.6

0.4

0.2

ERICACEAE

MELAMPYRUM

PLANTAGO LANC PLANTAGO MAR POTENTILLA ROSACEAE

CRUCIFERAE

RANUNCULACEAE COMPOSITAE TUB COMPOSITAE LIGULIFLORAE

TRIFOLIUM TYPE SUCCISA PRAT

22.6 0

10.8 2

1.8 0

2.6

2.2 0

2.4 1

0.4

0.8

1.2 0

0.4

0.6 0

5.8 0

0.4 0

0.2

3.2 0

3.2

0.2

1.8 0

14.9 0

1.3 0

6.6 0.2

2.6 0

0.4 0

0.8

0.4

0.8

10.8 0.2

3.2

0.2 0.2 0

1.6 0.2

0.4

5.6

2.8

4.5

3.6

7.4

0.2

0.8

1.4

0.2

0.6

SPERGULARIA

FILIPENDULA

0.8

3.8

3.2

7.8

ALISMA

11.8

22.6

2.8

CHENOPODIACEAE

CHAMENERION

22.2

CARYOPHYLLACEAE 0.8 RUMEX

19.2

0 0 0 0

1.2 0

0.2 0 0 0 0

0.8 0 0 0 0 0 0

0 0 0 0 0 0.4

0 0 0 0 0 0

0.4 0.2 0 0 0 Palaeoenvironment \ 875

SPERGULA

URTICA

0.2

UMBELLIFERAE

LILIACEAE

SUM HERBS

480

472

486

484

521

467

431

% HERBS

94.4

97.2

96.8

94.7

93.4

86.2

SPHAGNUM

0.2

0.8

3.8

1.8

0.5

4.4

6.4

PTERIDIUM

0.2

1.2

MONOLETE PSI

1.2

4.4

2.6

16.8

2.2

1.8

1.6

POLYPODIUM

0.6

0.8

0.6

1.3

1.4

0.6

LYCOPODIUM

0.2

OTHER TRILETE

0.2

SELAGINELLA

ZYGNEMATACEAE

0.2

SPYROGYRA

0.4

0.2

MOUGEOTICA

0.2

0.4

0.2

COPEPODA

0.6

CHARCOAL

ZONE

3a?

3a or 3b

3a or 3b Palaeoenvironment \ 876

Table 21.41 Spot samples from B TRENCH NUMBER

BIV WX

BIVWX

B III

BIII

BINX

B INX

BINX

CONTEXT

883

882

859

121

800

806 ard X

806

809

SAMPLE

2020

2022

2050

passage slot passage ‘pit’

passage kerb pit

PINUS SYLV

0.2

0.4

ALNUS GLUT

0.4

CONTEXT TYPE TAXA

BETULA

CORYLUS

QUERCUS TILIA

FRAXINUS ULMUS SALIX

SUM TREES

1.4

5.6 0

2.2 3.6 0

0.2 0

4.6

3.4 0

2014

linear groove

2019

2063

2013

2023

2029

ard mark

ard marked ard mark soil

0.6

0.2

0.6

1.2

0.2

0.4

0.8

0.2

0.4

1.8 2

0.2

0.2 24

1.2

1.8

0.2 0

1.6 2.2 0 0

1.6

3.8 0

0 0

1.6 5

0 0

0.2 38

ard mark

0.4

5.2 0

% TREES

7.6

7.4

8.2

4.8

3.6

4.2

6.8

7.6

7.2

GRAMINEAE

60.2

71.4

53.4

42.8

42.2

59.8

30.6

39.6

54.4 Palaeoenvironment \ 877

CYPERACEAE

0.2

1.6

3.8

0.8

2.6

3.2

0.6

CEREALIA

0.2

0.8

0.2

0.4

ERICACEAE

MELAMPYRUM

PLANTAGO LANC PLANTAGO MAR POTENTILLA ROSACEAE

CRUCIFERAE

RANUNCULACEAE COMPOSITAE TUB COMPOSITAE LIGULIFLORAE

TRIFOLIUM TYPE SUCCISA PRAT

12.8 0

11.4 0.2

0.6 0

1.2

0.6

RUMEX

SPERGULARIA

CHAMENERION

0.2 1

0.8

1.4 0

16 0

7.8 0

0.8 0

0.8

6.2

0.4

26.2 0

9.6 1

6.6 0

25 0

10.4 0.6

3.8

0.2

3.6

3.8 0

20.8 0

8.8

0.2 0.6 0

0.6 0

48 0

8.6

0.4 2

0.2 2

29.8 0

0.2 4

0.2

0.4 4

0.2

17.4 0

0.4 2 0

0.4 2

1.8

5.2

2.4

2.8

0.4

3.6

0.2

0.4

0.2

0.4

0.6

0.2

CARYOPHYLLACEAE 0

CHENOPODIACEAE

0.2

0.4

ALISMA

0.2

FILIPENDULA

Palaeoenvironment \ 878

SPERGULA

UMBELLIFERAE

467

458

URTICA

LILIACEAE

SUM HERBS % HERBS

462

92.4

0.2 0

93.4

91.6

476

95.2

482

96.4

479

95.8

0 0

466

93.2

0.2 0

462

92.4

464

92.8

SPHAGNUM

3.8

3.4

11.2

0.2

0.4

0.2

MONOLETE PSI

2.6

3.2

2.6

0.8

1.2

2.6

0.4

0.8

PTERIDIUM

POLYPODIUM

LYCOPODIUM

OTHER TRILETE SELAGINELLA

ZYGNEMATACEAE SPYROGYRA

MOUGEOTICA COPEPODA

CHARCOAL ZONE

0.2

3.6 0

0.2 0

0.6 0 0 + 3ai or 3aii

0.4

1.8 0

0.2 0

0.2 0 0 0 + 3ai

0.2

0.2 0

0.2 +

3aii

0.6 0

0.8 0

0 0

0.2 0

0 + 3c?

0 + not zoned

0 + 3c early

1.4 0

0.2

1.2 0 0

0.2 + 3c early

0 + 3c early

Palaeoenvironment \ 879

Table 21.42Â Spot samples from D TRENCH NUMBER

D II

D0I

CONTEXT

379

377

916

386

385

392

360

331

910

SAMPLE

CONTEXT TYPE TAXA

2031

2087

turf in cul- clay under tiv. bed OGS

2017

2028

2024

2032

2010

2015

2018

post hole

pit

cairn fill

cairn clay

PINUS SYLV

1.8

0.6

2.2

0.2

0.8

0.6

ALNUS GLUT

0.4

0.2

0.8

0.5

0.4

0.2

1.6

BETULA

CORYLUS

QUERCUS TILIA

FRAXINUS ULMUS SALIX

SUM TREES

2.8 10.6 0

7.8 0

1.2

4.4 0

0.2 0

2.8 7

0.2 0

0.2 2.8 0

0.2 0

1.3 2.5 0 0

0.3

0.3 22

1.2

1.2 0

0 0

1.4 2.4 0 0

2.8

5.6

0.2 0

% TREES

16.2

11.2

6.6

3.4

5.5

2.8

4.6

10.8

GRAMINEAE

46.6

56.6

33.8

50.2

44.5

80.8

26.6

38.6 Palaeoenvironment \ 880

CYPERACEAE

4.8

0.4

5.2

6.5

7.4

0.2

CEREALIA

0.2

0.6

0.4

0.3

0.2

ERICACEAE

MELAMPYRUM

PLANTAGO LANC PLANTAGO MAR POTENTILLA ROSACEAE

CRUCIFERAE

RANUNCULACEAE COMPOSITAE TUB COMPOSITAE LIGULIFLORAE

23.2 0

12 0

1.2 0

0.4

1.8 0

13 0

6.6

0.6

5.4 0

0.4

2.8 0

17 0

10.2 0.8

0.8 0

0.6

0.3

0.8

36.4 0

4.6 0

2.4

0.6

1.2 0

0.2

32.6 0

3.4 0

1.8 1

0.4 0

30 0

3.5

0.5 7

0.5 0

9.6 0

1.2 0

2.2 0

0.2 0

0.4

46.2 0

6.4

0.4 6

0.2 0

30.2 0

15.2 0.2

1.2 0

0.6

2.2 0

7.8

6.2

2.2

3.4

1.2

0.3

2.6

1.8

0.6

0.2

0.8

0.4

0.6

0.3

0.2

CARYOPHYLLACEAE 0.8

0.4

0.2

0.4

0.5

0.2

TRIFOLIUM TYPE SUCCISA PRAT

CHENOPODIACEAE RUMEX

SPERGULARIA

CHAMENERION

0.5

ALISMA

FILIPENDULA

0.2

0.3

Palaeoenvironment \ 881

SPERGULA

0.2

UMBELLIFERAE

0.2

URTICA

LILIACEAE

SUM HERBS % HERBS

419

83.8

444

88.8

467

93.4

0.2 0

435 87

483

96.6

0.3 0

378

94.5

486

97.2

477

95.4

446

89.2

SPHAGNUM

1.6

0.2

1.2

0.4

0.5

0.2

0.6

MONOLETE PSI

0.2

0.8

0.6

0.8

0.2

0.6

PTERIDIUM

POLYPODIUM

LYCOPODIUM

OTHER TRILETE SELAGINELLA

ZYGNEMATACEAE SPYROGYRA

MOUGEOTICA COPEPODA

CHARCOAL

ZONE

0.8 0

0.4

early 3a

1.8 0

0.2 0

early 3a

late 3a

1.6 0

0.2 0

0 0

0.6 0

0.5 0

0.2 0 0

0.2

3c?

0.4 0

0.2 0

1.6 0

0.2 0

?late?3a

0 +

Palaeoenvironment \ 882

Table 21.43Â Spot samples from D TRENCH NUMBER

CONTEXT

360

SAMPLE

CONTEXT TYPE TAXA

2041 cairn clay

2042 cairn clay

2043 cairn clay

2045 cairn clay

2052 cairn clay

2056 cairn clay

D IV

362

357

2057

2083

2077

cairn clay

chamber slot

PINUS SYLV

0.4

0.2

1.2

0.2

0.4

0.8

ALNUS GLUT

0.4

0.2

1.2

0.6

BETULA

CORYLUS

QUERCUS TILIA

FRAXINUS ULMUS SALIX

SUM TREES % TREES

GRAMINEAE

CYPERACEAE

2.1

6.7 0

9.5 53.6 0

0.2

1.6 0

0.2 0

2.6 68.4 0

1.2 0

3.4 74.4 0.8

3.4

2.8 0

7.6 63.8 0.6

5.2 73.8 0

1.2 2.8

0.2 0

5.4 67.8 0.8

1.2

2.4 0

0 0

0.2 22

4.4 54.8 0.8

0.2 1.2 0 0

1.8 63.2 0.2

0.8 2.8 0

0 0

5.4 44.6 1.4

Palaeoenvironment \ 883

ERICACEAE

19.9

11.6

7.8

7.4

9.2

27.8

MELAMPYRUM

CEREALIA

PLANTAGO LANC PLANTAGO MAR POTENTILLA ROSACEAE

CRUCIFERAE

RANUNCULACEAE COMPOSITAE TUB COMPOSITAE LIGULIFLORAE

10.6 0.2

0.2 0

0.2

3.2

0.2

4.8

0.2 1

0.2

2.2 0

0.2

0.8 0

0.4 1

0.2 7

0.2 0

0.4

3.6

0.2

0.2 6.2 0

0.4

0.2 0

3.2

0.4

0.2 11.8 0.4

1.2 0

0.2

2.6 0.8

0.2 18.4 1.6

4.2

0.2

0.4 4.2 0

2.2 0

0.6 0

0.2 1 0

5.2 0

0.4

0.4 0

2.5

1.2

4.2

2.2

1.4

0.2

0.4

0.2

0.6

CARYOPHYLLACEAE 0

0.2

0.4

0.2

TRIFOLIUM TYPE SUCCISA PRAT

CHENOPODIACEAE

RUMEX

SPERGULARIA

CHAMENERION

ALISMA

FILIPENDULA

SPERGULA

0.4

0.2

0.6 0

Palaeoenvironment \ 884

URTICA

0.2

LILIACEAE

UMBELLIFERAE SUM HERBS % HERBS

513

90.5

487

97.4

483

96.6

462

92.4

474

94.8

473

94.6

478

95.6

491

98.2

473

94.6

SPHAGNUM

2.5

0.4

0.8

4.6

1.2

0.4

1.2

MONOLETE PSI

1.2

0.4

2.6

0.8

16.8

1.4

0.8

PTERIDIUM

POLYPODIUM

LYCOPODIUM

OTHER TRILETE SELAGINELLA

ZYGNEMATACEAE SPYROGYRA

MOUGEOTICA

COPEPODA

0.5

1.8 0

0.5

1.2 0

0.2 0

0.2

1.2 0

0.2 0

0 0

1.8 0

late 3a

?early 3a

late 3a

0 0

ZONE

2.4

CHARCOAL

0.2

0.4

0.2 1 0

0.2 0

0.2

0.2 0

0.2

late 3a

bound 3b/3c

0.2 3c

Palaeoenvironment \ 885

Table 21.44Â Spot samples from D TRENCH NUMBER

CONTEXT

355

356

365

344

315

SAMPLE

CONTEXT TYPE TAXA

2081 chamber clay

2059

2039

2065

2069

2073

2072

2075

2078

wall clay

OGS

plough soil

PINUS SYLV

0.2

2.6

0.6

1.8

1.4

1.2

0.8

0.2

ALNUS GLUT

0.4

1.4

0.6

0.2

0.6

0.2

0.8

BETULA

CORYLUS

QUERCUS TILIA

FRAXINUS ULMUS SALIX

0.2

0.6

0.6 0

4.8

4.2 0

2.8 0

5.2

3.4 0

5.8

2.6 0

0.2 0

1.8 0 0

0.2

0.2 0

5.4

2.4 0

0.2 0

3.2 4.2

0.2 0

0.4

4.6 0

0 0

SUM TREES

GRAMINEAE

25.8

53.8

31.6

23.4

32.8

29.6

28.8

27.6

65.6

% TREES

1.8

10.6

Palaeoenvironment \ 886

CYPERACEAE

25.2

1.6

2.8

4.4

4.2

2.6

2.8

CEREALIA

0.2

ERICACEAE

MELAMPYRUM

PLANTAGO LANC PLANTAGO MAR POTENTILLA ROSACEAE

CRUCIFERAE

RANUNCULACEAE COMPOSITAE TUB COMPOSITAE LIGULIFLORAE

35.4 0

2.2

0.8

5.4

0.6

0.8

0.2

14.2 0

8.2

0.8

0.6

0.2 0

3.6

0.2

47.2 0

3.6

0.2

4.8 0

0.2

0.4 0

50.8 0

5.8

0.2 4

0.4

0.8 0

38.2 0

0.4

5.2 0

0.2 0

55.8 0

2.8 0 6

0 1

49.8 0

3.6 0

0.2

0.2 0

55.8 0

7.2 0

3.2 0

0.2 0.6 0

18 0

5.4

0.4

0.4 0

0.8 0

0.2

1.4

2.8

1.8

1.4

1.6

0.2

0.6

CARYOPHYLLACEAE 0

0.2

TRIFOLIUM TYPE SUCCISA PRAT

CHENOPODIACEAE RUMEX

SPERGULARIA

CHAMENERION ALISMA

FILIPENDULA

0.2

0.4

1.4

0.2

0.8

0.4

0.2

Palaeoenvironment \ 887

SPERGULA

UMBELLIFERAE

URTICA

LILIACEAE

SUM HERBS % HERBS

491

98.2

435 87

465 93

447

89.4

447

89.4

460 92

455 91

445 89

470 93

SPHAGNUM

0.2

5.4

0.2

0.4

MONOLETE PSI

0.4

1.4

0.2

2.2

PTERIDIUM

POLYPODIUM

LYCOPODIUM

OTHER TRILETE SELAGINELLA

ZYGNEMATACEAE SPYROGYRA

MOUGEOTIA COPEPODA

0.2

0.2 0

CHARCOAL

ZONE

1.2 2

0.2 0

0.2

mid/ late 3a

?upper 3a

0.4 0

0 0

0.8 0

0.2 0

0.4 0

3a/3b bound

0.6 0

0.2

6.6

0.2 0 0

0.2

1.2

middle turf 3a/3b line? bound

0 +

Palaeoenvironment \ 888

Table 21.45Â Spot samples from H TRENCH NUMBER

CONTEXT

751

771

736

770

735

730

778

SAMPLE

CONTEXT TYPE TAXA

2012

2061

OGS N of OGS on cairn green clay

2026 token burial

2071

2027

cairn clay

late kerb slot

2048

2051

fill in slot fill in slot below cairn below cairn

PINUS SYLV

0.2

2.2

0.6

1.4

ALNUS GLUT

0.6

0.4

1.2

BETULA

CORYLUS

QUERCUS TILIA

FRAXINUS ULMUS SALIX

SUM TREES % TREES

GRAMINEAE

0.8 2.2 0.4 0 0 0 0

20 4

43.8

6.8 6 0 0 0 0 0

13.4 55.2

1.6 10.8 0 0 0 0 0

15.6 41.6

6.4 7.8 0 0 0 0

0.2 81

16.2 46.6

1.2 7 0 0 0 0 0

9.8 68.8

5.2 3.2 0 0 0 0 0

9.4 62.8

6.6 0.4 0

0.4 0 0

12.8 58.2 Palaeoenvironment \ 889

CYPERACEAE

0.8

0.6

1.4

0.6

1.2

0.2

CEREALIA

0.8

0.2

ERICACEAE

MELAMPYRUM

PLANTAGO LANC PLANTAGO MAR POTENTILLA ROSACEAE

CRUCIFERAE

RANUNCULACEAE COMPOSITAE TUB COMPOSITAE LIGULIFLORAE

TRIFOLIUM TYPE SUCCISA PRAT

43 0

3.4 0.2 0.6 0 0 0 0

1.2 0

0.6 2.8 0

0.4 1.8 0

0.2 2.4 0.2

7 0

2.2 0

0.6 3.8 0.2

6.2 0.6 0.2 0 1

1.2 0.8

13.8 0 7

0.2 0.2 0

0.4 2.8 0

14.4 0

7.2 0.2 0.8 0 1

3.2 0.2

1.8

0.8

1.6

0.2

SPERGULARIA

FILIPENDULA

0.2

8.8

7.4

0.2

ALISMA

8.8

39.8

1.2

CHENOPODIACEAE

CHAMENERION

24.8

3.6

CARYOPHYLLACEAE 0 RUMEX

15.8

0 0 0

0 0 0 0

0 0 0

0 0 0 0

0 0

0.2

0 0

0.2

0 0

0 0.2

0 0

Palaeoenvironment \ 890

SPERGULA

UMBELLIFERAE

URTICA

LILIACEAE

SUM HERBS % HERBS

0 0

480 86

0 0

433

86.6

0 0

422

84.4

0 0

419

83.8

0 0

451

90.2

0 0

453

90.6

0 0

436

87.2

SPHAGNUM

0.2

6.2

5.6

7.6

6.2

2.4

4.6

MONOLETE PSI

0.6

1.4

0.2

1.8

4.8

PTERIDIUM

POLYPODIUM

LYCOPODIUM

OTHER TRILETE SELAGINELLA

ZYGNEMATACEAE SPYROGYRA

MOUGEOTIA COPEPODA

CHARCOAL

ZONE

0.6 0 0 0 0 0 0 0

0.8 0 0 0 0 0 0 0

trans 3a 3b early 3a

0.2 1 0

0.2 0 0 0 0 0

very early 3a

0 1 0

0.2 0 0 0 0

0.4 +

3a/3b

0.2 0 0 0

early 3a

1.8 0 0 0 0 0 0

0 +

early 3a

0.2 4.8 0 0 0 0 0 0

very early 3a

Palaeoenvironment \ 891

Table 21.46Â Spot samples from H, C and G TRENCH NUMBER

G III

GII

CONTEXT

734

707

201

1y4

lyr 3

SAMPLE

CONTEXT TYPE TAXA

2076 ard mark

2080

2082

275

plough soil plough soil basal peat

2036 soil

2060 soil

PINUS SYLV

0.2

1.2

0.4

0.2

0.8

ALNUS GLUT

0.6

0.4

0.2

0.8

BETULA

CORYLUS

QUERCUS TILIA

FRAXINUS ULMUS SALIX

SUM TREES % TREES

GRAMINEAE

CYPERACEAE

4.8 14.2 0 0 0

0.2 0

101

20.2 45

0.2

0.6 4.6 0 0 0 0 0

6.8 64

0.4

3.2 8 0 0 0 0 0

11.6 49.6 1

4.2 0.6 1 0 0 0 0

6.2 3.8 3

1.2 0 0 0 0 0

3.6 28.8 0.2

13.4 5.4 0 0 0 0 0

103

20.6 56.4 1.2

Palaeoenvironment \ 892

ERICACEAE

17.2

14.2

57.8

61.2

5.2

MELAMPYRUM

CEREALIA

PLANTAGO LANC PLANTAGO MAR POTENTILLA ROSACEAE

CRUCIFERAE

RANUNCULACEAE COMPOSITAE TUB COMPOSITAE LIGULIFLORAE

TRIFOLIUM TYPE SUCCISA PRAT

0 7

0.2 0.6 0

0.4 5.4 0.2

2.6 0.4

0.2 0.8 3.4 0.2

26.4 0.2 0 0 0

0.6 0

0.2 1.6 0

5.6 0.4 1.2 0

0.4 1.6 0

6.2

0.4

0.2

0.4

0.2

0.8

0.2

SPERGULA

0.8

1.6

0.8

SPERGULARIA

FILIPENDULA

1.4

0.4

0.2

2.8

ALISMA

0.6

3.8

0.4

3.4

CHENOPODIACEAE

CHAMENERION

2.4

2.6

CARYOPHYLLACEAE 0 RUMEX

0.2

0 0 0 0 0

0 0 1 0 0 0 0 0

0 0

0.4 0 0 0 0 0

0.2 0

0.2 0 0 0 0 0

0 0 0 0 0

0.2 0 0

0.2 0 0 0 0 0 0 0 Palaeoenvironment \ 893

URTICA

1.4

0.2

LILIACEAE

0.2

0.4

UMBELLIFERAE SUM HERBS % HERBS

SPHAGNUM PTERIDIUM

MONOLETE PSI POLYPODIUM

LYCOPODIUM

OTHER TRILETE SELAGINELLA

ZYGNEMATACEAE SPYROGYRA

MOUGEOTIA COPEPODA

CHARCOAL ZONE

399

79.8 48.8 0

4.4 4.8 0 0 0 0 0 0 0 + early 3a

466

93.2 0 0 0 0 0 0 0 0 0 0 0 +

0 442

88.4 0 0 0 0 0 0 0 0 0 0 0 +

first half of earlyish 3a 3a

0 469

93.8 0.2 0

0.4 0 0 0 0 0 0 0 0 +

482

96.4 0

0.2 0.8 2.2 0

0.2 0 0 0 0

0.2 +

397

79.4 0.2 0 5 5 0

0.8 0 0 0 0

0.2 +

Palaeoenvironment \ 894

Calanais Survey and Excavation, 1979-88

22. Macroplant Robin Inglis and Anne Crone with additional material from P J Ashmore 22.1 Sample collection, storage and preparation 22.1.1 Introduction During excavation a few pieces of charcoal were treated as small finds. The vast majority of the pieces of charcoal considered here were however from samples taken for soil or pollen analysis. In 1983 Mr R McCullagh was commissioned by Historic Scotland to analyse a set of samples primarily for radiocarbon dating. In 2005 AOC Archaeology Ltd was commissioned by Historic Scotland to undertake processing and charcoal identification of the remaining samples retrieved during the excavations. Some of the samples had been processed soon after the excavation but the majority had been stored unprocessed in sealed plastic bags and cardboard boxes, at Historic Scotland stores at Croft an Righ in Edinburgh. During Phase 1 of this programme, the unprocessed samples were variously sieved and sorted, the nature of the sieving, wet or dry, depending on the size of the sample (see Methodology below). The data from Phase 1 was tabulated and following selection of key contexts by the excavator, Phase 2, a programme of charcoal and macroplant identification and preparation of samples for radiocarbon dating, was undertaken. This report draws together the results of Phases 1 and 2 and integrates them with the results of the earlier programme of charcoal analysis (McCullagh 1983). 22.1.2 Methodology Slightly different methods of sample processing were used in each programme. McCullagh (1983) processed dried samples of varying sizes. The sediment was treated in a bath of hydrogen peroxide (H2O2), which was used to break out the carbon component of the sample. This also results in bleaching of the organic material,

and releases carbon dioxide which forms a white froth on top of the bath containing the charcoal fraction. Once treated the samples were poured over a bubbled water surface to separate the sediment fraction, the organic material that floated off being caught in a 2.00 mm and a 0.5 mm sieve. In the recent programme undertaken by AOC the samples were assessed for their suitability for wet or dry sieving. A comprehensive inventory of all of the samples and their volumes was produced. Those with volumes greater than 2 litres were chosen for wet sieving. The exception to this was one 1.5 litre sample which was judged to contain large quantities of charcoal. The total assemblage consisted of 265 samples, of which 39 bulk samples were wet sieved. All sieving was undertaken using a Siraf style flotation tank which differed from the methodology applied in 1983 in that there was no pre-treatment. The samples were subjected to the manual manipulation of the sediment in a bubbling water bath, floating the organic material through sieves of 1 mm and 0.3 mm. The remaining samples, i.e. those less than 2 litres, were dry sieved through a sieve stack containing a 4mm and a 1mm sieve. The different portions were then fully sorted for all artefacts and ecofacts. Any material recovered was then counted, and weighed by material type. The same process was applied to the dry residues resulting from the wet sieving process. All information was collated on a Microsoft Excel spreadsheet. Following selection of key samples their charcoal contents were analysed. As many fragments as possible were identified as to species, up to a maximum of ten. Species identifications were made with reference to Schweingruber (1978). Single fragments of each species were weighed and then stored in gel capsules for subsequent radiocarbon dating. The condition of the charcoal was noted and recorded, together with size and number of growth-rings where relevant. This latter may provide insights into the size of wood available.

Macroplant \ 895

Calanais Survey and Excavation, 1979-88

22.2 Results The results from AOCâ&#x20AC;&#x2122;s sieving and sorting programme are presented in Table 1. Charcoal was present in 121 of the samples processed. However, only 14 samples have more than 4 g of charcoal present, with the bulk of the samples (82) containing significantly less than 1 g of charcoal. The largest quantity of charcoal retrieved was from sample 2007. It probably came from context 315 on Area D; the preceding sample 2006 which produced 9.5 gm of charcoal, was from this context and although sample 2007 was unlabelled apart from the sample number, 315 is one of the few contexts likely to have produced so much charcoal which yielded a total of 30.17 g. Despite the relatively small quantities of charcoal retrieved during this round of processing they compare favourably with those studied previously by McCullagh (1983), whose brief was to concentrate on potentially datable material. Other material types retrieved during the sorting programme include fragments of flint, quartz, bone (including small mammal and fish bone), charred macroplant remains and some ceramics. The charred macroplant material consisted pri-

marily of hazelnut shell, but one context, secondary capping of the cairn, contained three cereal grains, identified as barley and wheat (Table 4). At least three of the ceramic sherds have some form of linear decoration. Context 356, [2059], also contained a single insect fragment. Of the 265 samples processed 135 were sterile, in that they contained no significant artefacts or ecofacts. The combined results of both programmes of charcoal analysis are presented in Table 2. The information is referenced by sample number and context/site number. The information is not presented by site as previous attempts at spatial analysis have not proved useful (McCullagh 1983). In all, 320 individual identifications have been made, totalling 57.69 grams. Figure 1 shows the overall species composition of the assemblage: 22.2.1 Species composition Eight species were present, the most abundant of which were birch (33.44%), alder (23.44%) and hazel (21.25%). The only variation between the range of species identified by McCullagh (1983) and those identified during the AOC work is the absence of heather. Heather can be identified mac-

Illus 22.1Â Species composition Macroplant \ 896

Calanais Survey and Excavation, 1979-88

roscopically but none was found despite close observation of each sample for condition, etc. It should be noted that in many instances identification beyond genus has not been possible. This is due mainly to the small size and poor condition of many of the fragments under investigation (see below). So, for instance, we cannot identify the species of the pine. Pomoideae covers a wide range of tree/shrub type, from large fruit trees, such as apple, to small bushy shrubs, such as hawthorn. 22.2.2 Condition The condition survey of the charcoal is presented in Appendix 3. The overall condition of the charcoal was very poor, and in some cases made species identification impossible. However, this provides some insight into the taphonomy of the charcoal, the degree of abrasion reflecting the nature of the context from which it was derived. For instance, those contexts with a concentration of highly abraded fragments may often represent plough soils (Ashmore pers comm). This not only gives an indication of the depositional environment but suggests that some of the charcoal may have come from the earlier layers and deposits disturbed by the plough. Post-depositional processes had an effect on some charcoal. The different depositional and post-depositional environments altered the condition of the charcoal considerably. The charcoal in sample 277 from posthole 093 on Area A, dating to the first half of the fourth millennium BC, shows consistent amounts of heavy mineralization and abrasion, as does that in sample 278 from a dark layer 160 on Area B considerably confused by iron-panning. On the other hand while the charcoal in sample 2365, from gritty sandy clay 732 in the basal infill of the cairn on Area H, indicates a different depositional or post-depositional environment with consistent light mineralization and abrasion. The weight of the majority of the charcoal pieces in the assemblage, despite the species, is 0.1 g or under, with a consistently small number of rings; with only 9 identifications having more than 10 visible rings. This points to small and young

specimens being utilised on-site, and possibly indicates the poor growing environment for larger species / specimens on the Western Isles. 22.2.3 Radiocarbon-dated charcoal at Calanais (P.J Ashmore) The selection of pieces of charcoal for dating reflected the survival of large fragments in interesting contexts. Hence the number of dates for each species (Illus 23.2) does not match their frequency of occurrence at Calanais (Illus 22.1). Alder was the only species (with more than one sample dated) restricted to a single short period, around the transition from pollen sub-zone 2c to 2d in the last third of the 4th millennium. Pollen evidence shows that it had begun to form a continuous curve at the end of sub-zone 2b (which occurred sometime between 4020 and 3750 cal BC) but was not particularly prominent in 2c or 2d (ending at some time between 2980 and 2510 cal BC). It has continued to grow in the area. Pine, which it had been thought might prove either all very early or all very late, occurred in both the late 4th and the mid 3rd millennia BC. However, as described by Ann Crone below, many sub-peat pine stumps found under peat in Lewis dated to the 3rd millennium cal BC. While the general guideline that most Scottish archaeological pine turns out to be very early or late may remain true for most of Scotland it is not true for the area around Calanais. No attempt has been made to correlate identified charcoal with pollen zonation of individual contexts. It was clear from analysis of some contexts that the pollen in them was of a different date from the charcoal because when soil and clay were moved about so was the charcoal in them; but sometimes the pollen reflected the period when the material was moved. 22.3 Discussion The information gained from the charcoal analysis in this study reinforces the conclusions drawn from other palynological and carbonised assemblages for this region.

Macroplant \ 897

Calanais Survey and Excavation, 1979-88

Illus 22.2Â Sample materials by species and date Macroplant \ 898

Calanais Survey and Excavation, 1979-88

22.3.1 Carbonised assemblages from other Neolithic sites in the Western Isles (B A Crone) Crone (1999) has previously summarised the current data for the Western Isles; her report is copied here in full apart from omission of the paragraph relating to Calanais (McCullagh 1983). At the domestic site of Bharpa Carinish, North Uist, a series of adjacent hearths and hearth deposits produced predominantly hazel (62%) and birch (35%) with a few fragments of willow and rowan (Sorbus aucuparia) (Crone 1993, 376). Much of the charcoal was small roundwood, between 5-40 mm in diameter. At Screvan Quarry, North Uist, a pit containing Neolithic pottery and other artefacts produced a charcoal assemblage which was predominantly birch (91%) (Downes & Badcock unpublished). Hazel, Ericaceae (ling and heather), Prunoideae (cherry) and Salicaceae (willow and poplar) were also present and it was noted that the charcoal appeared to originate from roundwood. The bulk of the charcoal assemblage from the Neolithic features at Allt Chrisal, Barra came from a kiln, or pottery clamp (Boardman 1995). Some 77% of the charcoal in this feature was birch, 10% was pine and alder, while hazel and Pomoideae (this group includes hawthorn, crab-apple and rowan) were also present. The charcoal assemblages from several other Neolithic features on the site were similar in composition although much smaller. It was noted that, throughout the entire assemblage, there was very little evidence for timber charcoal, small roundwood being the norm (ibid 153). The records of earlier excavations tend to list the wood species simply on a presence/absence basis. At Eilean an Tighe, North Uist, an extensive ashy deposit produced primarily birch and hazel together with a few pieces of willow, all described as ‘pieces of stem’, or roundwood (Scott 1951a, 24). A single piece of Scots pine came from the ‘byre’ structure. At Northton, Harris, the Neolithic II deposits produced birch, hazel, willow and bog myrtle (Myrica gale) (Cowie pers comm). Scott recorded birch charcoal from the chamber floor of the cairn at Clettraval, North Uist (1935, 521) and willow, hazel, oak, pine and possibly birch charcoal from the cairn at Unival, North Uist (1948, 1)

22.3.2 Palynological and other evidence for the woodland resource of the Western Isles Various travellers and writers have recorded the macroscopic remains of wood throughout the Western Isles, in intertidal peats and under blanket peat (Angus 1987). While these testify to the former presence of trees on the islands it is not possible, on these observations alone, to ascertain the extent and nature of the woodland or the chronology of its disappearance. A systematic study of the macroscopic remains of birch, pine and willow at forty sites on Lewis has provided some answers (Wilkins 1984). The pine was found as large in situ stumps, up to 0.5 m in diameter, and in two locations, a ‘virtual forest of stumps’ was recorded (ibid 254). Radiocarbon-dating placed many of the stumps in the 4th and 3rd millennium cal BC (Dickson and Dickson 2000, 37-39, 67). The birch usually occurred as small branches, no more that 80 mm in diameter, lying horizontally near the base of the peat. The birch was deposited in the peat at a much earlier date, in the 6th millennium cal BC. Ritchie (1985) has also radiocarbon-dated birch from intertidal deposits at Borve, Benbecula to 5700 ± 170 uncal BP. Roots, trunks and branches thought to be birch have been observed in inter-tidal peats closer to Eilean Domhnuill, at Vallay, North Uist but these have not been dated (Beveridge 1911, 6). Early palynological work in the Western Isles tended to reinforce the belief that the islands were largely devoid of woodland from a very early date. However, what may be a treeless landscape to the palynologist may still contain sufficient exploitable woodland resources for the inhabitants. For instance, at Little Loch Roag, a low arboreal pollen count was interpreted as signifying a ‘forestless’ landscape throughout the Flandrian although a small amount of birch and hazel scrub was also postulated (Birks & Madsen 1979). Only 11 km away at Calanais Leobag, Bohncke found much higher arboreal pollen values and concluded that pockets of birch woodland, also containing hazel, willow, rowan and poplar, probably existed in sheltered valleys in the area (1988, 461); see also the next section 23.3.3 and Chapter 21: Palaeoenvironment. If such differences can exist between pollen sites so close together then the low arboreal Macroplant \ 899

Calanais Survey and Excavation, 1979-88

pollen counts from sites on or near the edge of the machair in North Uist, at Balemore (Hirons forthcoming), Balelone Farm and Loch Scolpaig (Mannion & Moseley forthcoming) may signify similar pockets of scrub woodland further inland. In contrast, studies in South Uist suggest that up to half of the available landscape around Loch Lang was wooded in the Neolithic period and that the woodland was also species-rich (Bennett et al 1990). At Loch Lang the pollen record is dominated by tree and shrub pollen until 4000 BP. Birch and hazel dominated the woodland but oak (Quercus sp.), elm (Ulmus sp.), alder and ash (Fraxinus excelsior) were also present. Pine pollen was recorded but its status within the woodland is uncertain. At Askernish high pollen counts indicated the presence of birch-hazel scrub in the locality (Hirons 2003) and high proportions of tree pollen are also recorded for the earlier part of the Postglacial at the nearby sites of Loch an t-Sil and Loch Airigh (Edwards et al 1995). In their reconstruction of the former woodland cover of Scotland McVean and Ratcliffe (1962 Map B) postulated birch woodland all along the eastern half of the Outer Hebrides. The palynological record tends, on the whole, to support this general picture but indicates greater species diversity on the southern islands, while the macroscopic evidence indicates that stands of pine probably punctuated the landscape. This summary of the woodland resource available in the Western Isles presents the background against which the charcoal assemblage from Calanais should be viewed. Analysis of the Calanais assemblage indicates that the local woodland was similar in composition to that found elsewhere in the Hebrides, i.e. birch-dominated with significant quantities of hazel and alder. 22.3.3 The evidence from palynological analysis at Leobag (P J Ashmore summarising part of Chapter 21) This section is mainly a summary of the results of palynological analysis at Leobag undertaken by Sjoerd Bohncke (Bohncke 1988 and Chapter 21). It includes some information from radiocarbon dating at Calanais (see Chapter 23)

From some ill-dated time between the late 7th and the early 5th millennium BC the lower areas round Calanais, some now submerged by the sea, contained pockets of birch woodland with subsidiary hazel, willow and rowan. In its understory were species such as Melampyrum, Lonicera and ferns. Sphagnum moss was abundant at Leobag (Pollen zone CaN-1). From some ill-dated time between the second quarter of the 6th millennium BC and the middle of the 5th millennium BC there was some disturbance, perhaps caused by hunter gatherers burning woodland, leading to a patchwork of woodland and grassland mixed with heather. Towards the end of this phase herbs, dwarf shrubs and ferns showed a firm increase. Another remarkable feature of this period at Leobag was the abundance of willow straight after the birch decline (Pollen zone CaN-2a). From some ill-dated time between the last half of the 6th millennium BC and the last half of the 5th millennium BC birch regenerated somewhat. Heather increased. Elsewhere wet grassland developed and sphagnum mosses were fairly abundant. There were some signs of continuing hunter-gatherer activity (CaN-2b). At some time between about 4020 and 3750 cal BC elm began to form a discontinuous pollen curve. At most a few centuries later arable farming began to be practiced on a small scale, while pockets of birch woodland continued to flourish. Grasses increased and more diverse herb vegetation appeared (CaN-2c). At some time between 3490 BC and 3020 cal BC until between 2980 and 2510 cal BC birch woodland regenerated strongly and it is thought that the area was not used for farming. Potentilla varieties and grasses declined. This period (CaN-2d) immediately preceded the building of cultivation ridges at Calanais itself. The next period (pollen zone CaN-3a) saw the building of the cultivation beds, the stone setting and the chambered cairn. It began at a date between 2980 and 2510 cal BC with a strong decline in birch and the reappearance of cereals, followed by an increase in grasses. Hazel, oak, willow and rowan reacted with a slight increase. At the end of this period birch temporarily recovered. Macroplant \ 900

Calanais Survey and Excavation, 1979-88

During analysis of the excavation samples at Calanais it was often subdivided into CaN-3ai, in which there were indications of agriculture, and CaN-3aii in which there were stronger indicators of pastoralism. The transition from 3ai to 3aii may have coincided with building of the chambered cairn. The period ended at some time between 2560 and 2200 cal BC. Between then and some date between 1900 and 1500 BC there was a phase (CaN-3b) of mixed arable and stock farming ending with a short term regeneration of birch woodland. Rowan disappeared from the area. It may be that this or the next period saw despoliation of the cairn. In the next period (CaN-3c) birch pollen declined under 10% for the first time. Shortly afterwards elm pollen disappeared from the samples and pine showed a marked decline. Other tree species slowly declined. Grasses and heather increased again, and cereals were continuously present with a maximum at some time between 1500 and 1200 cal BC. Towards the end of this sub-zone the Betula percentages increased a little together with the Alnus and Pinus values, while the Ranunculaceae, Cruciferae, Caryophyllaceae and Plantago lanceolata declined. This period possibly ended between about 1000 and 350 cal BC. It seems likely that the end saw peat growth initiation at higher altitudes as well as an increased in peat growth at Leobag. It marked the end of archaeologically recorded activities at Calanais.

22.4 Appendices 22.4.1 Appendix 1 – Samples examined for dating material 22.4.2 Appendix 2 – Combined charcoal identifications 22.4.3 Appendix 3 – Condition survey from recent charcoal analysis 22.4.4 Appendix 4 – Macroplant identifications 22.4.5 Appendix 5 – Samples sent for radiocarbon 22.4.1 Appendix 1 Samples examined for dating material

Macroplant \ 901

Charcoal from Retent

Total charcoal content

<0.1

1.27

Callanish 2340 - 2364

<0.1

0.08

377

CSS 354 - 381

<0.2

271

CSS 264-284

<0.5

282

CSS 264-284

<0.5

753

CSS 752-773

<0.5

0.22

754

CSS 752-773

<0.5

0.2

755

CSS 752-773

<0.5

0.49

756

CSS 752-773

<0.5

757

CSS 752-773

<0.5

0.05

758

CSS 752-773

<0.5

0.05

759

CSS 752-773

<0.5

0.07

Sample

Box

2347

Callanish 2340 - 2364

2348

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0 0.86

0.86

0.9

1.3

Macroplant \ 902

Charcoal from Retent

Total charcoal content

<0.5

0.91

CSS 752-773

<0.5

0.26

704

CSS 752-773

<0.5

0.42

765

CSS 752-773

<0.5

0.16

771

CSS 752-773

<0.5

0.31

2057

CSS 2049-2068

<0.5

2059

CSS 2049-2068

<0.5

2436

CSS

<0.5

2456

CSS 2453-2553

<0.5

2466

CSS 2453-2553

<0.5

2477

CSS 2453-2553

<0.5

2489

CSS 2453-2553

<0.5

2494

CSS 2453-2553

<0.5

Sample

Box

760

CSS 752-773

762

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0.02 0.12

0 0.32

0.32

0.04

insect remains - n/a

0 0.07

0.07

Macroplant \ 903

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Sample

Box

2498

CSS 2453-2553

<0.5

2511

CSS 2500etc

<0.5

2512

CSS 2500etc

<0.5

286

CSS 285-299

<0.5.

0.14

752

CSS 752-773

0.5

0.24

761

CSS 752-773

0.5

0.04

767

CSS 752-773

0.5

4.26

2037

CSS 2023-2048

0.5

2056

CSS 2049-2068

0.5

2491

CSS 2453-2553

0.5

763

CSS 752-773

0.75

5.39

769

CSS 752-773

0.75

0.46

768

CSS 752-773

2.41

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0.08

Macroplant \ 904

Charcoal from Retent

Total charcoal content

1.74

CSS 752-773

1.29

773

CSS 752-773

2.18

2018

CSS 2009-2022

2.63

2047

CSS 2023-2048

0.1

2367

Callanish 2365-2367

0.61

2457

CSS 2453-2553

0.55

2473

CSS 2500etc

2499

CSS 2500etc

2547

CSS 2500etc

2552

CSS 2453-2553

2553

CSS 2453-2553

267

CSS 264-284

1.5

Sample

Box

770

CSS 752-773

772

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

0.12

Charred Rmns

Lithic

Pot

Bone

Qutz

0.15

0.33

Other:

0.09

0.12

Macroplant \ 905

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Charred Rmns

CSS 2453-2553

1.5

Yes

6.65

0.31

6.96

0.02

2554

CSS 2500etc

1.5

284

CSS 264-284

Yes

2068

CSS 2049-2068

Yes

3.92

2161

Callanish 2365-2367

Yes

0.03

2367

Callanish 2365-2367

Yes

1.62

0.32

1.94

2356 - 2350

Callanish 2340 - 2364

Yes

3.05

0.26

3.31

2458?

CSS 2453-2553

Yes

1.95

0.15

2.1

2455

CSS 2453-2553

2.5

Yes

3.83

4.83

2462

CSS 2500etc

Yes

12.33

1.02

13.35

381

CSS 354 - 381

2.75

Yes

2007

Callanish? 2007

Yes

Sample

Box

2480

Description

Lithic

Pot

Bone

Qutz

Other:

0 0 0.57

4.49

0.04 0.04

0.39

0.03 3.18 0.07

0.2

0.24

0 4.08

26.09

30.17

0.12

3.06

Macroplant \ 906

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Sample

Box

2483

CSS 2453-2553

n/a

2488

CSS 21 2453-2553

<0.5

2550

CSS 1 2453-2553

0.12

374

CSS 354 - 381

10/1 F15

<0.2

0.11

358

CSS 354 - 381

10/1 NE, Middle Sample 1 Top Bank 331

<0.2

0.05

357

CSS 354 - 381

10/1, NE, 331/2

0.2

0.06

354

CSS 354 - 381

10/1, NE, F-336, Sample 6

<0.5

0.08

375

CSS 354 - 381

10/1, NE, F-337

<0.5

0.05

2049

CSS 123, Rab2049-2068 bit Root

2433

CSS

20/5, 851, <0.1 B1 NWx

379

CSS 354 - 381

24/07 C, Sub samples, 16-32

0.5

<0.1

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0.07

11.81

Macroplant \ 907

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Sample

Box

Description

Vol (L)

2426

CSS

36, Black Layer, C2

0.5

2425

CSS

36, C/1 W, Bottom of <0.5 Peat Layer

2428

CSS

36/02

2427

CSS

36/C2, Black Layer

0.5

2429

CSS

36/C2, Black Layer

0.75

2083

CSS 362 2069-2090

<0.1

2359

Callanish 2340 - 2364

362

0.5

2358

Callanish 2340 - 2364

365, D1&5, 3 of 4

2.5

2351

Callanish 2340 - 2364

383

3.02

2354

Callanish 2340 - 2364

383

0.5

1.29

2508

CSS 2500etc

40/4, 6,

<0.5

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0 1

Dessicated organic remains - 2.19

1 0

0.1

Yes

0.1

11.48

0.15 0.02 Macroplant \ 908

Charcoal from Retent

Total charcoal content

<0.5

0.36

CSS 40/7, 20 2453-2553

<0.5

0.2

2053

CSS 846, B8 2049-2068

<0.5

2432

CSS

887, B4 Wx

<0.5

2070

CSS 889 2069-2090

360

CSS 354 - 381

A 2, 19, 22 <0.5

290

CSS 285-299

A1, 21, <0.5 infill of 34

366

CSS 354 - 381

A1, 31, 15, under infill <0.2 Layer 41

372

CSS 354 - 381

A1, 31, 16&17, Layer 42

<0.5

296

CSS 285-299

A1, 4, layer <0.5 40, 19

280

CSS 264-284

A1, F-87, 24

<0.5

281

CSS 264-284

A12, 23

<0.5

Description

Sample

Box

2471

CSS 40/7, 15 2453-2553

2472

Vol (L)

1.5

Wet Sieved

>4mm Charcoal from flot

0.02

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

1.36

0.04

Macroplant \ 909

Charcoal from Retent

Total charcoal content

Box

Description

365

CSS 354 - 381

A2, 46, 14, <0.2 Layer 47

295

CSS 285-299

A2, 9, 79

0.5

368

CSS 354 - 381

A-F3, 26, Layer 10 Top

<0.2

Callanish - Peter Strong

Ard mark Fill, Call 88, 13/05/88

0.5

Callanish - Peter Strong

Ard mark Fill, Call 88, 13/05/88

0.5

364

CSS 354 - 381

Area A, Fr 3, 27, Layer 10 bottom

<0.5

2063

CSS B/Wx 2049-2068

<0.5

2029

B/Wx, CSS 809, Ard 2023-2048 mark A

<0.5

2023

B/Wx, CSS 809, Ard 2023-2048 mark B

<0.5

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Sample

0.1

0.59

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0.1

0.59

0.1

Macroplant \ 910

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Sample

Box

2025

CSS B/Wx, Pit 0.5 2023-2048 154

2013

CSS B1 NW 2009-2022 Ex, 806

0.5

2033

B1 NW, CSS 861, 2023-2048 Charcoal

<0.1

2034

CSS B1 Nw, 2023-2048 862

<0.1

264

CSS 264-284

2019

CSS B3, F-800 0.5 2009-2022

0.48

Pot

Bone

Qutz

Other:

B3, F-142, 1 43

0 0

CSS

<0.5

2431

CSS

B3, Post hole 179, post pipe B

<0.5

2050

CSS B4 Wx, 2049-2068 859

2021

CSS B4, 139 2009-2022

0.5

278

CSS 264-284

B4, F-160, 2 34

Lithic

0.48

B3, Post hole 179, post pipe A

2435

Charred Rmns

2.66

2.66 0

Yes

0.53

0.24

0.77

0.19

Macroplant \ 911

Description

Sample

Box

2022

CSS B4Wx, 2009-2022 882

2011

CSS B4Wx, 2009-2022 883

2020

CSS B4Wx, 2009-2022 883

1.5

2054

CSS B5, 812 2049-2068

0.5

2085

CSS B5, Sam2069-2090 ple 141

<0.5

CSS

2165 2030

Bag 4, 13/05/88

Vol (L)

Wet Sieved

Yes

>4mm Charcoal from flot

2.93

Total charcoal content

Charred Rmns

0.07

0.03

0.09

3.02

0.57

0.15

1.6

CSS BH, 139 2023-2048

<0.5

CSS

Black Layer at S 1 3/4

CSS

Black Layer at S4

1.5

0.04

Qutz

0.07

Other:

0.44

0.63 fish bone - n/a

0.03

Yes

Bone

0.1

0.15

Callanish - Peter Strong

10.19

0.15

Callanish Barvas? 2365-2367

Pot

0.11

0.07

Yes

Lithic

Black Charcoal Filled Feature, Call 88, 13/05/88

1.45

Charcoal from Retent

0.6

0.12 3.67

0.08

16.47

0.05

Fish Bone - 0.02 / shell - 38.86

0.04

Macroplant \ 912

Charcoal from Retent

Total charcoal content

Box

Description

Vol (L)

CSS

Black Layer, 14/05/88

1.5

Callanish - Peter Strong

Black Layer, Call 88, 13/05/88

Callanish - Peter Strong

Black Layer, Call 88, 13/05/88

Callanish - Peter Strong

Black Layer, Call 88, 14/05/88

1.5

Callanish - Peter Strong

Black Layer, Call 88, 14/05/88

2430

CSS

C/1 W, Organic Remains

<0.5

0.34

CSS

Charcoal in root mess

<0.1

2.28

2017

CSS D1 2009-2022

3.5

0.19

362

CSS 354 - 381

0.74

D1, 3/5, 13

Wet Sieved

>4mm Charcoal from flot

Sample

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

Yes

0.02

Yes

0.02

0 Dessicated Organic Remains - 0.84

0.07

0.03

Macroplant \ 913

Charcoal from Retent

Total charcoal content

<0.5

0.31

CSS 285-299

D1, 315, 7 1.25

0.64

289

CSS 285-299

D1, 326, 1 0.5

0.09

2015

CSS D1, 334, 2009-2022 1of2

2073

CSS D1, 334, 2 2.25 2069-2090 of 2

2084

CSS D1, 352 2069-2090

<0.5

2081

CSS D1, 355 2069-2090

<0.2

2069

CSS D1, 365, 2069-2090 2of 4

2.5

Yes

0.13

2039

CSS D1, 365, 2023-2048 4of4

2.5

Yes

0.13

2090

CSS D1, 374, <0.5 2069-2090 ashy patch

2028

CSS D1, 386, 2023-2048 292

<0.1

2032

CSS D1, 392 2023-2048

<0.1

Description

Sample

Box

2078

CSS D1, 315 2069-2090

293

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Yes

0 0.16

0.16

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0.06

0 0.01

0.14 0.13 0

0.03

0.03 0

Macroplant \ 914

Charcoal from Retent

Total charcoal content

<0.5

0.34

D1, Layer 315, 11

1.16

D1, layer 315, 2

1.25

0.42

2031

CSS D2, 290, 2023-2048 379

<0.5

288

CSS 285-299

D2, 316, 3 1

0.03

274

CSS 264-284

D2, F-316, <0.5 5

0.13

2077

D5, 359, CSS Ard mark 2069-2090 A-B

<0.5

2441

CSS

E1, F-15, Ard mark Fill

<0.1

2088

CSS E21, Sam<0.2 2069-2090 ple 2

2439

CSS

F 334, 302 <0.1

2567

CSS

F 369. D1 4

355

CSS 354 - 381

F. 335/5, 10/1 NE

Description

Sample

Box

2024

D1, fill CSS from 385, 2023-2048 291

287

CSS 285-299

292

CSS 285-299

Vol (L)

0.5

Wet Sieved

>4mm Charcoal from flot

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

1.33

Yes

2.29

0.82

3.11

0.47

0.02

9.13

Macroplant \ 915

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Sample

Box

2014

CSS F121, G3 2009-2022

0.5

291

CSS 285-299

F130, B1, 35

1.25

361

CSS 354 - 381

F-139, 42, Charcoal 0.5 Spread

273

CSS 264-284

F-145, B3, 1.5 37

367

CSS 354 - 381

F-172, 33

1.25

299

CSS 285-299

F20, Kilpatrick?, 16/3

0.5

2089

CSS F26, Tr E 2069-2090

<0.1

270

CSS 264-284

F-31, 50

<0.5

356

CSS 354 - 381

F-336, 3, NE 10/1

0.5

359

CSS 354 - 381

F-388

<0.2

297

CSS 285-299

F424

0.5

294

CSS 285-299

F5, 8

0.75

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0 0.05

0.05

0.23

4.98

0.01

0.02

Macroplant \ 916

Charcoal from Retent

Total charcoal content

Box

Description

283

CSS 264-284

F-627, 30, <0.5 Ard mark

276

CSS 264-284

F-704, 51

<0.5

369

CSS 354 - 381

Fv - 644, 12, Post hole Fill

<0.5

2060

G3, layer CSS 3, B?, NW 1 2049-2068 profile

2058

G3, layer CSS 4, B2, NW 1 2049-2068 profile

2005

CAL 80/ BL/001

2175

numerCallanish ous small n/a 2365-2367 subsamples

2500

CSS 2500etc

numerous small n/a subsamples

2501

CSS 2500etc

numerous small n/a subsamples

2502

CSS 2500etc

numerous small n/a subsamples

Half Sieved

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Sample

0.3

Yes

0.9

0.29

Charred Rmns

Lithic

Pot

0.3

1.19

Bone

Qutz

Other:

1.09

1.57

Macroplant \ 917

Charcoal from Retent

Total charcoal content

Box

Description

2505

CSS 2500etc

numerous small n/a subsamples

2062

Pit G3, CSS layer 1, SE 1 2049-2068 profile

2055

Pit G3, CSS Layer 3, 1 2049-2068 SE profile

2036

Pit G3, CSS Layer 4, 1 2023-2048 SE profile

2046

Pit G3, CSS Sample 2023-2048 A1, NE profile

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Sample

Lithic

Pot

0.12

1.49

Bone

Qutz

Other:

2440

CSS

Pollen Sample, Site H, <0.1 F-772, Sample 28

871

CAL 80/ BL/001

Pollen Un1 sieved, 1

0.05

871

CAL 80/ BL/001

Pollen Un1 sieved, 2

0.38

2497

CSS 2500etc

RC6

0.5

Charred Rmns

Macroplant \ 918

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Sample

Box

Description

Vol (L)

2496

CSS 2500etc

RC6, 7

0.5

378

CSS 354 - 381

Rest of label worn <0.5 off

380

CSS 354 - 381

Rest of label worn <0.2 off

2347

Callanish 2340 - 2364

Rest of label worn 2 off

2348

Callanish 2340 - 2364

Rest of label worn 1.75 off

2448

CSS 2447-

RO 15

1.75

2453

CSS 2447-

RO 15

1.5

2451

CSS 2447-

RO 15, 14 1

2450

CSS 2447-

RO 15, 16 2.5

2452

CSS 2447-

RO 15, 19 1.5

2454

CSS 2447-

RO 15, 20 1.75

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0 6.01

6.01

Yes

9.32

0.24

9.56

5.15

0 Yes

Macroplant \ 919

Charcoal from Retent

Total charcoal content

Box

Description

2449

CSS 2447-

RO 15, 21 2

Yes

2447

CSS 2447-

RO 15, 24 2.5

Yes

2086

CSS Sample 2069-2090 141, B5

0.5

376

CSS 354 - 381

Sample 4, 020

0.5

CSS

Second Bag, 13/05/88

268

CSS 264-284

Site A1, F-A30, Layer 35, 20

<0.1

2442

CSS

Site E, F-12

<0.1

0.14

2360

Callanish 2340 - 2364

Site E, F-30, Sample 3

0.75

0.03

371

CSS 354 - 381

Site F, F 629, 31

<0.5

0.05

2080

CSS Site H, 2069-2090 707

2079

Site H, CSS 714, Sam- <0.5 2069-2090 ple 1

Vol (L)

0.5

Wet Sieved

>4mm Charcoal from flot

Sample

0.25

Charred Rmns

0.25

Lithic

Pot

Bone

0.11

Qutz

Other:

0.49

0 Yes

0.24

0.04

1.58

0.18

0 0

0.85

Macroplant \ 920

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charcoal from Retent

Total charcoal content

Charred Rmns

0.06

0.01

Sample

Box

2027

Site H, CSS 735, Sam- <0.5 2023-2048 ple 6

2040

Site H, CSS 769, sam2023-2048 ple 22

2367

Site H, Callanish F.738, 2 2365-2367 Sample 15

2082

Site H, CSS F-707, 2069-2090 sample 4

2067

Site H, CSS F-726, <0.5 2049-2068 Sample 21

2048

Site H, CSS F-730, 1.5 2023-2048 Sample 32

1.65

2353

Callanish 2340 - 2364

Site H, F-732, <0.5 Sample 10

0.92

2363

Callanish 2340 - 2364

Site H, F-733, Sample 9

<0.5

0.04

2076

Site H, CSS F-734, 2069-2090 sample 3

0.75

0.84

2.5

<0.5

Yes

1.2

Yes

3.03

Lithic

Pot

Bone

Qutz

Other:

1.02

1.2

1.03

4.06

0.3

10.28

0.25

0.02

0.03

0.02

0.57

Macroplant \ 921

Charcoal from Retent

Total charcoal content

0.5

0.16

Callanish 2340 - 2364

Site H, F-746, 0.5 Sample 13

0.56

2352

Callanish 2340 - 2364

Site H, F-746, <0.5 Sample 14

0.11

0.03

2355

Callanish 2340 - 2364

Site H, F-747, 1 Sample 11

0.71

0.05

2362

Callanish 2340 - 2364

Site H, F-757, Sample 17, 0.5 Possible Chamber Floor

2434

CSS

Site H, F-766, <0.2 Sample 33

4.54

2438

CSS

Site H, F-766, <0.1 Sample 33

1.25

2061

Site H, CSS F-771, <0.5 2049-2068 Sample 30

0.55

Description

Sample

Box

2026

Site H, CSS F-736, 2023-2048 No.7

2364

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Charred Rmns

Lithic

Pot

Bone

Qutz

Other:

0.36

Macroplant \ 922

Charcoal from Retent

Total charcoal content

Charred Rmns

Site H, CSS F-772, 1 2023-2048 Sample 29

1.64

0.16

2051

Site H, CSS F-778, 2049-2068 sample 31

3.2

0.12

2361

Callanish 2340 - 2364

2016

CSS Site H, 3 2009-2022 Sample 24

2071

CSS 2071

13 13

Sample

Box

2038

Description

Vol (L)

Wet Sieved

>4mm Charcoal from flot

Site H, 0.75 Sample 18

Lithic

Pot

Bone

Qutz

Other:

0 Yes

5.35

0.12

5.47

0.28

1.22

Site H, 7.5 Sample 27

Yes

10.72

0.74

11.46

0.11

10.64

CSS

Third Bag, 3 13/05/88

Yes

1.09

0.35

1.44

Callanish - Peter Strong

Top Block, Cal 88, 4 13/05/88

Yes

0.66

0.07

0.73

2.14 0.07

0.98

5.54 0.05

0.19

See Technical note 22.4.1 for errors 22.4.2 Appendix 2: combined charcoal identifications Sample

Context

370

A17

Alnus glutinosa

Betula sp

Calluna vulgaris

Corylus avellana

Pinus sp

Pom-oideae

Salix sp

Sorbus sp

Unid

Weight (g) 0 Macroplant \ 923

Alnus glutinosa

Betula sp

Calluna vulgaris

Corylus avellana

Sample

Context

Pinus sp

Pom-oideae

Salix sp

Sorbus sp

Unid

269

A18

370

A18

277

A93

0.25

4277

A93

3.17

164

B111

B123

132

B139

279

B139

361

B139

391

B141

686

B141

2255

B142

243

B149

278

B160

215

B167

702

B167

2225

B806

143

B815

2247

B846

2257

B854

694

B881

698

B883

1 1

Weight (g) 1.2

0.09

* 1

* * not found

0.7 1

0.33 1

* 2

0.4 *

Macroplant \ 924

Alnus glutinosa

Betula sp

Calluna vulgaris

Corylus avellana

Sample

Context

Pinus sp

Pom-oideae

2009

B883

2011

B883

678

B885

500

B889

248

D300

604

D300

213

D305

685

D315

687

D315

2006

D315

2006

D315

347

D318

226

D333

2075

D334

684

D336

D344

666

D344

690

D352

236

D359

257

D360

2010

D360

0.7

Salix sp

Sorbus sp

Unid

Weight (g) 1.1

0.82

* * * 1

4 1

1.08

9.5

1 1

2 2

1 1

0.52

* 1

Macroplant \ 925

Sample

Context

Alnus glutinosa

Betula sp

Calluna vulgaris

Corylus avellana

2041

D360

2042

D360

2043

D360

2044

D360

2045

D360

2064

D360

c4.0

2066

D360

0.9

2068

D360

273

D361

2246

D361

0.9

689

D362

2039

D365

2065

D365

2069

D365

2045

D366

672

D369

678

D369

688

D369

693

D369

2567

D369

3688

D369

266

D375

Pinus sp

Pom-oideae

Salix sp

Sorbus sp

Unid

Weight (g) 6.5

c2.0 c2.0 0

0.93

0.02 0

0.06

0.6

* 1 1

1 6

* 1.5 *

3 1

0.97

* *

Macroplant \ 926

Alnus glutinosa

Betula sp

Calluna vulgaris

Corylus avellana

Sample

Context

692

D375

2087

D377

2249

D379

2250

D379

592

D398

592

D398

2052

D360

0.9

2035

D908

253

F615

266

F630

259

F631

0.2

272

F639

258

F643

373

F645

256

F648

256

F648

257

F649

260

F649

F708

H708

H712

H724

Pinus sp

Pom-oideae

Salix sp

Sorbus sp

Unid

Weight (g) *

0.07

* 3

1.1

c1.4

0.6

2.6 0 1

* *

Macroplant \ 927

Sample

Context

Alnus glutinosa

Betula sp

2366

H728

116

H732

2365

H732

2365

H732

674

H733

2243

H736

2367

H738

2367

Calluna vulgaris

Corylus avellana

Pinus sp

Pom-oideae

Salix sp

Unid

Weight (g) c10 *

0.38 1

6.1 *

1 2

Sorbus sp

2.9 2

0.68

0.28

123

H739

671

H740

2012

H751

2256

H795

230

H767

229

H768

324

H769

332

H769

2040

H769

2106

H769

2357

H769

321

H769

2356/0

H770

Third Bag

S13

0 2

2.1 2

* * 0.16

* 1

0.59

* 2

0.7 0.24

Macroplant \ 928

Sample

Context

Alnus glutinosa

Top Block

S13

0.26

Bag 4

S13

0.42

S33

TOTAL

Betula sp

Calluna vulgaris

Corylus avellana

Pinus sp

Pom-oideae

Salix sp

Sorbus sp

Unid

Weight (g)

0.17 107

57.69

22.4.3 Appendix 3: Condition survey from recent charcoal analysis Sample

Context

Species

Weight (g)

Description

Bag 4 13/05/88

Alnus

0.17

Not roundwood, very twisted fragment, unclear over ring no.

Bag 4 13/05/88

Corylus

0.04

2-3 rings visible, not roundwood, abraded and rounded

Bag 4 13/05/88

Alnus

0.06

3-4 rings visible, not roundwood, very mineralised and abraded around fringes

Bag 4 13/05/88

Pomoidea

0.04

Small fragment with only 2 rings visible, well preserved, not roundwood

Bag 4 13/05/88

Alnus

0.06

2-3 rings visible, not roundwood, abraded and rounded

Bag 4 13/05/88

cf. Pomoidea

0.03

Very mineralised throughout, very abraded, unclear over ring no.

Bag 4 13/05/88

Corylus

0.02

2-3 rings visible, not roundwood, abraded and rounded, mineralised around fringes

Macroplant \ 929

Sample

Context

Species

Weight (g)

Description

Third Bag 13/5/88

Betula

0.04

Slightly mineralised, 3-4 rings visible, rounded and abraded

Third Bag 13/5/88

Alnus

0.07

Well preserved lightly abraded, not mineralised, 3 rings visible

Third Bag 13/5/88

Alnus

0.05

8-9 rings visible, not roundwood, mineralised throughout

Third Bag 13/5/88

Betula

0.03

5-6 rings visible, roundwood - 8-9mm diameter

Third Bag 13/5/88

Betula

0.05

Very mineralised, very abraded and rounded, 2-3 rings visible, not roundwood

Top Block Cal 88

Corylus

0.15

4-5 rings visible, very rounded and abraded, not roundwood

Top Block Cal 88

cf. Alnus

0.02

1-2 rings visible, mineralised around fringes, abraded and rounded

Top Block Cal 88

Pomoidea

0.05

4-5 rings visible, very rounded and abraded, mineralised throughout

Top Block Cal 88

Alnus

0.04

1 ring visible, well preserved, mineralised around fringes

Alnus

0.06

3-4 rings visible, not roundwood, very mineralised and abraded around fringes

Alnus

0.02

Very small fragment, very abraded, 3-4 rings visible

Alnus

0.01

1-2 rings visible, not roundwood, very abraded

Alnus

0.03

3-4 rings visible, not roundwood, very mineralised and abraded around fringes

Alnus

0.05

2-3 rings visible, not roundwood, abraded and rounded

D344

Alnus

0.05

Well preserved fragment, 5-6 rings visible.

D344

Pinus

0.09

Up to 7 rings visible, well preserved, not mineralised.

D344

Betula

0.06

Up to 3 rings visible, very abraded and mineralised. Macroplant \ 930

Sample

Context

Species

Weight (g)

Description

D344

Corylus

0.03

Possible roundwood - up to 5mm radius - 5 rings visible, heavily abraded

D344

Alnus

0.03

Possible roundwood - 6-7 rings - 8mm diameter, abraded and rounded.

D344

Alnus

0.04

Possible roundwood - 5-6 rings - 8mm radius, abraded and fragmentary

D344

cf. Salix

0.06

Heavily mineralised fragment, possible roundwood - 5-6 rings - 8-9mm diameter

D344

Alnus

0.06

2-3 rings visible, heavily abraded and rounded, mineralised around fringes

D344

Pomoidea

0.03

3-4 rings visible, heavily abraded and mineralised

D344

Corylus

0.07

3-4 rings present, heavily abraded and rounded

256

F648

277

A93

Betula

0.06

2-3 rings visible, mineralised around fringes, heavily abraded and rounded

277

A93

Betula

0.05

Heavily abraded and rounded fragment, mineralised around fringes, 2-3 rings visible

277

A93

cf. Betula

0.08

3-4 rings visible, heavily mineralised, very twisted fragment

277

A93

Betula

0.02

2-3 rings visible, very friable, mineralised around fringes

277

A93

Betula

0.04

1-2 rings visible, very degraded and rounded, friable

278

Betula

0.07

1-2 rings visible, very soft and abraded, large parent fragment

278

Corylus

0.08

6-7 rings visible, very abraded and soft, not roundwood

278

Alnus

0.02

5 rings visible, possible roundwood - 5-6mm diameter, very soft and abraded

278

Corylus

0.02

Up to 3 rings visible, very soft and abraded fragment,

278

Corylus

0.04

2 rings visible, mineralised throughout, very abraded and soft

278

Corylus

0.03

Abraded around fringes, well preserved with up to 2 rings visible.

278

Corylus

0.04

*, very small fragment, up to 3 rings visible, very soft and abraded

278

Corylus

0.03

*, small and extremely abraded, 3-4 rings visible, not roundwood.

No fragments were identifiable

324

H769

Betula

0.11

8 rings visible, mineralised around fringes, very abraded

324

H769

Betula

0.03

5-6 rings visible, lightly abraded and mineralised Macroplant \ 931

Sample

Context

Species

Weight (g)

Description

324

H769

Betula

0.02

Possible the same as fragment above

370

A18

Alnus

0.02

Extremely abraded and mineralised, 13 rings visible

370

A18

Betula

0.05

Up to 10 rings visible, extremely abraded and mineralised

370

A18

Corylus

0.02

3-5 rings visible, lightly abraded and mineralised around fringes

592

D398

Alnus

0.03

2-3 rings present, very rounded and abraded

592

D398

Alnus

0.02

2 rings present, mineralised around fringes, very abraded

592

D398

Alnus

0.02

*, 2 rings visible, mineralised around fringes, very abraded

688

D369

Betula

0.2

4-5 rings visible, very degraded and soft.

688

D369

Alnus

0.29

4-5 rings visible, very degraded and soft.

688

D369

Pinus

0.34

Up to 6 rings visible, quite angular fragment, some abraded edges.

688

D369

Alnus

0.32

5-6 rings visible, abraded and mineralised

688

D369

Pinus

0.05

3-4 rings present, mineralised around fringes,

688

D369

Pinus

0.08

3-4 rings visible, mineralised throughout

688

D369

Pinus

0.11

Possible from same parent fragment as above.

688

D369

Pinus

0.05

Possible from same parent fragment as above.

688

D369

cf. Corylus

0.06

Poor condition, soft and abraded, 2-3 rings visible, mineralised along fringes

2006

D315

Corylus

0.28

16-17 rings visible, possibly roundwood - 7mm radius, angular and mineralised

2006

D315

Betula

0.1

Mineralised around fringes, rounded and abraded, 3-4 rings visible.

2006

D315

Corylus

0.09

3-4 rings visible, abraded and rounded, very twisted internal structure.

2006

D315

Corylus

0.13

26 rings visible, Roundwood of 10mm radius, angular and mineralised around fringes

2006

D315

Corylus

0.11

25-26 rings visible, large and abraded, same as above?

2006

D315

Salix

0.08

11-13 rings visible, rounded and abraded, mineralised around fringes

2006

D315

Betula

0.11

2-3 rings visible, mineralised around fringes, heavily abraded and rounded Macroplant \ 932

Sample

Context

Species

Weight (g)

Description

2006

D315

Betula

0.09

4-6 rings visible, mineralised around fringes, abraded and rounded

2006

D315

Salix

0.04

Possible roundwood - 3-4 rings visible - 8mm diameter, angular but abraded

2006

D315

Betula

0.05

11-12 rings visible, no mineralisation, angular but visible abraded

2011

Corylus

0.1

2 rings visible, mineralised around fringes, abraded and rounded

2011

Alnus

0.09

Not roundwood, 3 rings visible, mineralised and abraded, larger parent fragment

2011

Alnus

0.07

More than 7 rings visible, possibly roundwood - 7mm diameter, lightly abraded

2011

Corylus

0.13

Extremely abraded and mineralised fragment, approx 3 rings visible

2011

Alnus

0.1

Roundwood fragment - 5-6 rings visible - 10mm diameter, lightly abraded and mineralised

2011

Alnus

0.19

very abraded and rounded fragment, 1-2 rings present, mineralised around fringes

2011

Alnus

0.05

2 rings visible, possible roundwood - 3mm radius, well preserved

2011

Corylus

0.02

Possible roundwood fragment, 2 rings visible, 7mm diameter, mineralised throughout

2011

Alnus

0.04

Roundwood fragment - 5mm radius, 5-6 rings visible, mineralised throughout.

2011

Corylus

0.03

Not roundwood, 6-7 rings visible, mineralised throughout, heavily abraded.

2035

D908

No fragments were identifiable

2039

Corylus

0.02

Only fragment from context identifiable, roundwood - 6.5mm diameter - 3 rings

2040

Alnus

0.12

1-2 rings visible, mineralised around fringes, abraded and rounded

2040

Alnus

0.11

Up to 4 rings visible, not roundwood, mineralised along one edge towards centre

2040

Betula

0.08

Very soft and abraded fragment, 3-5 rings visible, extremely mineralised.

2040

Alnus

0.04

1-2 rings visible, very mineralised around fringes,

2040

Betula

0.03

2-3 rings visible, not roundwood, abraded and rounded

2040

Alnus

0.03

5-6 rings visible, not roundwood, very abraded and fragmentary

2040

Corylus

0.02

Possible roundwood - 4-5 rings visible - 7mm diameter, mineralised around fringes

2040

Betula

0.07

3-4 rings visible, not roundwood, very mineralised and abraded around fringes Macroplant \ 933

Sample

Context

Species

Weight (g)

Description

2040

Alnus

0.05

Large and abraded fragment, 3-4 rings visible

2040

Alnus

0.04

*, Up to 6 rings visible, extremely abraded and mineralised

2045

D366

Betula

0.06

Rounded and abraded fragment, 3-4 rings visible

2045

D366

Corylus

0.11

large angular fragment, 3-4 rings visible, lightly mineralised

2045

D366

Betula

0.07

Lightly mineralised, 2-3 rings visible, rounded and abraded

2045

D366

Betula

0.04

Lightly mineralised, only 1 ring visible, lightly abraded and rounded

2045

D366

Betula

0.06

*, 9-10 rings visible, lightly mineralised and rounded, quite abraded.

2045

D366

Corylus

0.04

1-2 rings visible, angular but abraded fragment, lightly mineralised.

2045

D366

Corylus

0.05

Possibly same as above fragment.

2045

D366

Betula

0.07

*, 4-5 rings visible, angular but abraded fragment.

2045

D366

cf. Betula

0.04

3-4 rings visible, very rounded and abraded, mineralised along one edge.

2045

D366

Alnus

0.06

2-3 rings visible, very friable, abraded and rounded, mineralised around fringes.

2068

Alnus

0.14

Possible roundwood - 1mm diameter, & rings visible, abraded and soft.

2068

Pomoidea

0.11

3 rings visible - not Roundwood, edges abraded and mineralised, very soft.

2068

Corylus

0.11

10-11 rings visible, possibly roundwood - 8mm diameter, abraded and rounded.

2068

Corylus

0.1

Roundwood fragment - 5 rings visible - 12mm diameter, mineralised and rounded.

2068

Corylus

0.07

Possible roundwood - 6mm radius, extremely mineralised

2068

Pomoidea

0.08

Not roundwood, up to 6 rings visible, very abraded and rounded

2068

Betula

0.08

Not roundwood, up to 6 rings visible, extremely friable and soft, much abraded.

2068

Betula

0.04

Not roundwood, 3-4 rings visible, mineralised only round fringes.

2068

Corylus

0.06

9 rings visible, not roundwood, mineralised and abraded

2068

Pomoidea

0.14

2 rings visible, not roundwood, extremely mineralised,

2069

Corylus

0.03

Extremely abraded and soft fragment. 4 rings visible, mineralised along one side Macroplant \ 934

Sample

Context

Species

Weight (g)

Description

2069

Betula

0.01

Very small and soft fragment, 1-2 rings visible.

2069

Betula

0.02

Very small and very mineralised fragment, 1-3 rings visible

2365

H732

Alnus

0.05

Possible roundwood of 4mm radius, unclear over no. of rings. Abraded and mineralised

2365

H732

Alnus

0.06

Possible roundwood of 5mm radius, well preserved and not mineralised.

2365

H732

Alnus

0.03

5 rings visible, mineralised around fringes, rounded and abraded.

2365

H732

Alnus

0.06

*, 8-9 rings visible, mineralised around fringes, not roundwood

2365

H732

Betula

0.03

3-4 rings visible, not roundwood, very mineralised and abraded around fringes

2365

H732

Corylus

0.04

5-6 rings visible, mineralised around fringes, well preserved

2365

H732

Alnus

0.02

3 rings visible, lightly mineralised, very abraded and rounded.

2365

H732

Corylus

0.03

Possible roundwood - 6mm radius - 4-5 rings visible, well preserved

2365

H732

Betula

0.04

Roundwood - 3-4 rings visible - 6mm diameter, mineralised throughout

2365

H732

Corylus

0.02

Roundwood - 2-3 rings visible - 3mm radius, mineralised around fringes

2367

H738, sample Pomoidea 15

0.26

Very soft and mineralised, 10-12 rings visible, roundwood of 7mm radius

2367

H738, sample Pomoidea 15

0.14

Possibly same as above fragment.

2367

H738, sample Alnus 15

0.06

Roundwood fragment, 5-6 rings visible, 6mm radius, well preserved.

2367

H738, sample Alnus 15

0.08

2-3 rings visible, well preserved, mineralised around fringes

2367

H738, sample Betula 15

0.14

2-3 rings visible, not roundwood, well preserved, mineralised around fringes

2367

Corylus

0.08

7-8 rings visible, mineralised around fringes

2367

Corylus

0.07

7-8 rings visible, mineralised in places

2367

cf. Alnus

0.05

5-6 rings visible, very mineralised and abraded. Macroplant \ 935

Sample

Context

Species

Weight (g)

Description

2367

Alnus

0.03

1-2 rings visible, not roundwood, abraded and mineralised around fringes

2367

Alnus

0.05

3 rings visible, very abraded and rounded, mineralised throughout.

2567

Alnus

0.26

3 rings visible - not roundwood, quite fragmented and abraded

2567

Alnus

0.14

4 rings visible, not roundwood, very abraded and soft

2567

Alnus

0.1

1-2 rings visible, mineralised around fringes, abraded and rounded

2567

Corylus

0.1

8 rings visible, well preserved and solid piece of charcoal, not roundwood,

2567

Corylus

0.12

Up to 4 rings visible, very abraded and twisted fragment.

2567

Alnus

0.02

2 rings visible, not roundwood, lightly mineralised

2567

Pinus

0.11

Well preserved lightly abraded fragment, 7 rings visible, not roundwood.

2567

Corylus

0.08

Possible roundwood, 2 rings visible - 10mm diameter, abraded, and soft

2567

Alnus

0.04

Not roundwood, 2 rings visible, extremely mineralised and abraded

2356/0

Alnus

0.11

6-7 rings visible, roundwood of 7mm radius, little mineralisation

2356/0

Betula

0.11

Extremely mineralised, 2 rings visible, large sections too distorted for clear view.

2356/0

Alnus

0.11

Very mineralised, up to 4 rings visible very abraded and rounded.

2356/0

Alnus

0.08

Very twisted and distorted fragment, very mineralised, up to 2 rings visible.

2356/0

Alnus

0.06

Possible roundwood fragment, up to 7 rings visible - 5mm radius,

2356/0

Betula

0.05

Extremely mineralised and abraded fragment. Up to 3 rings visible.

2356/0

Corylus

0.06

mineralised around fringes, up to 4 rings present, wide ring pattern

2356/0

Alnus

0.07

Very poorly preserved fragment, with up to 4 rings visible, very mineralised

2356/0

Corylus

0.02

Mineralised around fringes, up to 2 rings visible, rounded and abraded

Macroplant \ 936

Sample 2356/0

Context

Species

Weight (g)

Description

Alnus

0.03

Very mineralised and abraded fragment, 2-3 rings visible.

22.4.4 Appendix 4: macroplant identifications (Allan Hall) The following material was submitted for identification and to check its suitability for AMS dating. In the â&#x20AC;&#x2DC;Plant material presentâ&#x20AC;&#x2122; column *indicates suitable material and quantity for dating, (*) those cases which are more marginal in quantity). Context

Other labelling

Plant material present

Weight for dating (mg)

Bag 4 13/5/88, Callanish #24,28,33 (labelled 131 by ARH)

charred hazel (Corylus avellana L.) nutshell*

Top block Cal 88, 13/5/88, Callanish Peter Strong (labelled 132 by ARH)

ditto*

Second bag 13/5/88 Callanish #24, 28â&#x20AC;Ś

ditto (*)

336

356, 3, NE 1011 CSS 354-381

Specimen labelled as charred seed but material present comprised two fragments of (the same?) charred plant structure which may be an underground organ; there is probably not enough to date in any case

<10

374

10/1 F1 CSS 354-381

12 grains, probably all barley (Hordeum); weight is for 9 selected for dating*

704

CSS 752-773; in derivative lists supplied to specialists as from context 764, which must be a faulty transliteration of 704. 704 is the general label for the 3 barley and one wheat (Triticum) grain (*); also two charred ?seeds 12 cairn outside the chamber wall. I distinctly remember of unknown identity and a trace of wood charcoal putting grain from the late secondary capping of the cairn in sample bags (PJA).

883

2011 B4Wx 883 CSS 2009/2022

charred hazel nutshell*

115

missing

2016 Site H, Sample 24 CSS 2009-2022

charred hazel nutshell*; also little wood charcoal

182

916

2017 D1, CSS 2009/2022

charred hazel nutshell (*); also a little wood charcoal

910

2018 DI CSS 2009-2022

charred hazel nutshell*

96 Macroplant \ 937

Context

Other labelling

Plant material present

Weight for dating (mg)

883

2020 B4Wx, CSS 2009-2022

charred hazel nutshell*

883

2022 B4Wx CSS 2009-2022

charred hazel nutshell (*)

735

2027 Site H, Sample 6 CSS 2023-2048

charred hazel nutshell (*)

772

2038 Site H, Sample 29 CSS 2023-2048

charred hazel nutshell*

151

778

Site H, 2051, Sample 31 CSS 2049-2068

charred hazel nutshell*

114

356

2059 D CSS 2049-2048

charred hazel nutshell (*)

360

2068 D CSS 2049-2068

charred hazel nutshell (*); also two fragments of a broken cereal grain, probably wheat

889

2070, CSS 2069-2090

a very little wood charcoal

770

Site H, Sample 27, 2071 CSS 2071

charred hazel nutshell*

122

352

D1, 2084, CSS 2069-2090

charred hazel nutshell*

746

Site H, 2352, Sample 14 CSS 2340-2364

charred hazel nutshell*

732

Site H, 2353, Sample 10 CSS 2340-2364

charred hazel nutshell (*); also a little wood charcoal

747

Site H, 2355 Sample 11 CSS 2340-2364

charred hazel nutshell (*); also a little wood charcoal

missing

2480 CSS 2453-2553

Charred plant material, perhaps seaweed? (perhaps not suitable for dating, but if needed, weight is 21 mg)

missing

2508 RL â&#x20AC;&#x2DC;82 40/4 6 CSS 2500 etc.

charred hazel nutshell (*)

369

D2567

charred hazel nutshell (*)

770

2356 H /0? CSS 2340-2364

charred hazel nutshell*

22.4.5 Appendix 5: Charcoal found in pottery bags in 2011 The following is a list of charcoal fragments found in pottery bags by Dr Sheridan; they have now ( July 2011) been separated out. Macroplant \ 938

Calanais Survey and Excavation, 1979-88

Find no.

Area

Context

80.168

BIII

160.1

80.172

315

81.234

315

81.315

DIV

316

81.154

746

81.229

HII

768

81.323

HII

769B

22.4.6 Appendix 6: Samples sent for radiocarbon dating Sample No.

Context No.

Species

269

A18

Betula

4277

A93

Betula

2006

D315

Pinus

D344

Pinus

2045

D360

Pinus

2045

D360

Corylus

688/2567

D369

Pinus

688/2567

D369

Alnus

688/2567

D369

Corylus

592

D398

Alnus

256

F648

Corylus

Top Block

Corylus

2011

B4 Wx 883

Hazelnut Shell

2020

B4 Wx 883

Hazelnut Shell

Bag4

Strong 13

Hazelnut Shell

374 10/1 F1

704

Barley Seed

764

704

Wheat Grain

764

704

Barley Grain

2084

D1 352

Hazelnut Shell

2352

H764

Hazelnut Shell

2038

H772

Hazelnut Shell

2051

H778

Hazelnut Shell

The report and catalogue prepared by R McCullagh in 1983 is archived in the folder â&#x20AC;&#x2DC;Calanais final resource filesâ&#x20AC;&#x2122; in NMRS. Macroplant \ 939

Calanais Survey and Excavation, 1979-88

23. Radiocarbon Dating 23.1 Introduction 23.1.1 Measurement and calibration The 39 radiocarbon dates discussed here were obtained in three batches. A basal peat date was measured around 1982 by the Glasgow University radiocarbon laboratory. 15 dates for charcoal, identified by Rod McCullagh, were obtained in 1997 from the Accelerator Mass Spectrometer at Arizona (AA-) (Ashmore 1997) and a further 23 from charcoal identified by Robin Inglis in 2006 were measured on the Scottish Universities Environmental Research Centre Mass Spectrometer at East Kilbride (SUERC-). Historic Scotland paid for all of the dates. The age obtained around 1982 dated the humic acid from a slice of basal peat to the east of the main stone setting. Each of the 1997 and 2006 ages was obtained from a single piece of charcoal or charred hazel nut shell, to avoid the possibility of mixed charcoal of different periods producing a chronologically meaningless average age (Ashmore 1999, 2000). The charcoal came from short-lived species, apart from the pine (Crone, pers comm). The samples were all prepared at Scottish Universities Research and Reactor Centre in East Kilbride. The ages and dates are listed with their calibrations in Table 23.3 and plotted on Illus 23.1. The calibrations here use the 2004 data set published by Reimer et al (2004) and were calculated using OxCal 3.10 (Bronk Ramsay 2005), set to cub r:5 sd:12 prob usp [chron]. 23.1.2 Cautionary words It is important to remember two things. Firstly, the dates were not for human activities such as the filling of a pit, but for a few yearsâ&#x20AC;&#x2122; growth of a plant, in most cases a tree or a shrub, charcoal from which was incorporated in the pit at some later date. The measured ages were not necessarily from the latest rings of the plant, and they averaged the radiocarbon age of each annual ring in the sample.

Apart from the pine, the species dated at Calanais had natural lifetimes of no more than a century and in most cases significantly less (although coppicing can lead to survival of the bole of a tree or shrub for longer than normal). Only three of the samples had unusually negative dC13 values, which might have been a sign of having been from wood preserved in peat bogs prior to being burnt, but they were from hazel which unlike birch does not seem to occur often in deep peat. The high d13C levels may have another explanation. Given the damp local climate the dated charcoal from Calanais probably did not result from wild fires. In that sense each date probably reflected human activity shortly after the death of the plant from which the charcoal was produced. But in most cases the charcoal probably lay somewhere other than its find spot for an appreciable time, before being turned up by ground-working or moved with its surrounding soil into a new place from which we sampled it. In what follows such samples have been called â&#x20AC;&#x2DC;residualâ&#x20AC;&#x2122; (Ashmore 1999a). Secondly, the 39 dates obtained reflect only a tiny fraction of the activities at Calanais. Given that several dates are indistinguishable from one another, the 38 dates (not including the peat date), spread over about 2000 years of goings-on, might represent one or two charring episodes every 2 to 10 generations. Keep in mind also that although the one sigma limits have been included in the tables, to abide with the conventions of SAIR, about a third of the dated samples will have had true ages outwith them; and one or two of the ages probably have true dates lying outside the 2 sigma limits quoted here; there is only a 16.7% chance that all of them lie between the 2 sigma limits quoted. 23.2 Samples and their treatment 23.2.1 Selection strategies There are basically two possible strategies for a dating programme. One is to select only samples Radiocarbon dating \ 940

Calanais Survey and Excavation, 1979-88

which relate directly to what one is interested in; for instance, collagen or bone apatite to measure the age of skeletal material, or a charred cereal grain to show when a particular species was being grown. The other is to obtain many dates with where possible more than one from each context, to get a set of termini post quem. In this second approach the main danger is that we have a natural tendency to forget that a sample may have been be very much older than the context in which it was found. Nevertheless termini post quem can be very useful. For instance, had the charred cereal grains in the cairn capping at Calanais produced modern dates they would have shown that the capping was modern. In the end this dating programme included a mixture of both strategies. The dates of the charred cereals were of interest as part of the history of cereal growing irrespective of their stratigraphic position. Another part of the thinking was that it would be valuable to see whether the tendency for Scottish pine to have either a pre-4th millennium date or a fairly recent one applied in the Calanais area (Crone pers comm). But a major consideration in obtaining the 2006 age set was that not enough ages were measured in 1997 to allow an understanding of patterns of residuality. 23.2.2 Residuality It is clear that some pieces of charcoal were found in contexts formed much later than the death of the plants providing the charcoal. The contexts with irrefutably residual charcoal, because they also contained significantly later non-intrusive charcoal, were: —— D315, a plough soil; —— D344, a plough soil; —— D398, green clay under the cairn. —— D360, a dirty clay under the cairn —— D369, a soil under plough soil 315. There were also contexts with charcoal for which residuality was demonstrated by the existence of much later dates from stratigraphically earlier contexts. They included:

—— H772 and H778, fills of slot 773 cutting the green clay under the chamber wall; —— B883, a gritty soil in what was probably the same slot; Then there were contexts where the presence of residual charcoal was interpreted from general stratigraphy but could not be proved unambiguously. —— H746, the uppermost fill of a depression, immediately below turf line 771. —— D352, sandy clay outside the cairn; 23.2.3 Modelling and Bayesian analysis In an early draft of this report Bayesian models were created based on a subset of the dates in the belief that each of them was of much the same date as its context. The models are listed in Technical Note 23.2.3. Using the inbuilt Bayesian methods of OxCal the conclusion was that the Ring was set up and the chambered cairn built between 2900 and 2600 cal BC. That estimate was widely publicised. But further consideration of the contexts and other dating information shows that even more of the charcoal fragments than supposed in selecting them for modelling were centuries old at the time that the contexts in which they were found were sealed. Amongst the most important examples is the piece of hazel charcoal (AA-24967, sample 2243/81) in a context H736 forming part of the primary build of the cairn. The context included a beaker rim sherd which cannot have been earlier than 2500 cal BC, at least a hundred years later than the Bayesian dating model suggested (and rather late for the radiocarbon date (2860 to 2470 cal BC) from H736 considered in isolation). And another date (2910 to 2630 cal BC) from a single piece of birch charcoal from the closely related context H732 (AA-24966, sample 116/81) was significantly different from that for sample 2243/81. If the piece of charcoal from 732 was residual it seems very possible that the one from 736 was too. In essence it is impossible to guarantee that any samples from Calanais were of much the same

Radio carbon ating \ 941

Calanais Survey and Excavation, 1979-88

date as their contexts, except possibly a fragile heather twig (sample 132/80; AA-24957) in a late context. The simple Bayesian model was invalid, depending as it did on the assumption that a sample in a later context would always be younger than a sample in an earlier context. Attempts to use Bayesian techniques were abandoned. 23.3 Indirect dating Analysis of pollen samples from Calanais and their comparison with radiocarbon-dated reference columns at Callanish Leobag produced a dating scheme which fitted well with a late date for the cairn and less well with the Bayesian dating model produced in the draft report. The reference columns there were zoned and radiocarbon-dated (Table 23.1). Then short Kubiena box columns from Areas C, D, E and H at Calanais were zoned by comparison with the Leobag columns; and spot samples were compared with these short profiles. That allowed dates from the Leobag columns to be applied (cautiously) to the pollen-sampled layers at Calanais.

The interpolated and projected dates were estimated from their nearest measured neighbours in the same peat column. Two different interpolation methods were used for some important subzone boundaries. The projected dates for subzone boundaries 1 to 2a-2b are less trustworthy than the others because of increased uncertainties about variations in rates of peat growth in the lower parts of CN3. The evidence from Leobag and the calculations underlying the dates in Table 23.1 are explained in Chapter 21: Palaeoenvironment. Several of the most interesting layers sampled for pollen at Calanais were turf lines. These need to be interpreted with great care because a single turf line can grow for centuries; it reflects a process rather than an event. So stratigraphically a single turf line can have had a life interrupted by being covered by clay in one place, but ongoing elsewhere. Just to re-emphasise this point, because it is important in assessing the date of the Ring and the green clay platform under the cairn, a pollen sample in one part of a visually coherent turf line does not necessarily provide a date for a different part of that turf line.

Table 23.1 Pollen zone transition dates from Calanais Leobag Peat Descrip- Pollen depth tion sub-zones CN1 Projected Peat

Peat

Interpolated Interpolated Peat

Peat Peat

Calculated

Peat depth CN3

Code

late 2 σ cal BC

late 1 σ cal BC

early 1 σ early 2 σ cal BC cal BC

516

564

400

760

850

3e to 4 CN1

-86

3e CN1

GU-1289 200

GU-1170 150

200

750

800

262

340

859

921

3c-3d CN1

424

519

993

1072

3c CN1

GU-1986 1110

1130

1380 1380

1410

1500

GU-1171 1200

1380

1640

1800

1520

1750

1880

3e CN1

3d-3e CN1

3c CN3

3b-3c CN1

3b-3c CN1/ CN3

50 -

GU-1290 800

1010

Radiocarbon dating \ 942

Calanais Survey and Excavation, 1979-88

Peat Descrip- Pollen depth tion sub-zones CN1

Peat depth CN3

Code

Peat

3a-3b CN3

3ai-3aii

Peat Interpolated Method 1

Interpolated Method 2

Peat

Interpolated

Peat Peat Peat

Interpolated

Peat

Interpolated

Interpolated Interpolated Peat

late 1 σ cal BC

early 1 σ early 2 σ cal BC cal BC

GU-1987 1600

1660

1880

1900

GU-1988 2200

2290

2470

2560

22.5

2290

2445

2605

2865

3ai-3aii

22.5

2355

2455

2655

2770

3a CN3

GU-1989 2300

2460

2620

2900

2510

2620

2840

2985

GU-1990 2580

2670

2910

3010

GU-1991 3020

3100

3360

3490

2c CN3

GU-1992 3510

3530

3710

3780

2c cereal

3605

3650

3830

3875

2c CN3

GU-1993 3700

3770

3950

3970

2b-2c

3830

3905

4095

4120

2a-2b

5525

5660

5980

6075

Sharp birch drop

5655

5795

6215

6225

GU-1234 5850

6000

6340

6450

3b-3c CN3

2d-3a CN3

2d CN3

2c-2d CN3

53.5

Charcoal appears

Projected 1-2a

Projected

Projected start 1

late 2 σ cal BC

6045

6205

6555

6675

6305

6475

6845

6975

6695

6880

7280

7425

Radio carbon ating \ 943

Calanais Survey and Excavation, 1979-88

23.4 Dating Calanais 23.4.1 Introduction The date probability distributions in Illus 23.1 are normalised by area; those with a small error are shorter and wider than those with a large one. What follows here is based on the individual dates within the framework of an interpretative phasing of the site. After a brief introduction to how the ages grouped together into clusters, they are discussed in the order of the phases in Table 23.2 rather than either purely chronological or stratigraphic succession. Many of the dates from the Calanais samples can be grouped into two clusters (Illus 23.1). One cluster of 13 dates fell in the last third of the 4th millennium, their probability distributions spilling over into adjacent centuries. All of these were from charcoal old at the time it was deposited at Calanais. Another of 9 dates occupied a period from about 2900 to 2500 cal BC. In addition a group of 4 dates occupied the period between about 2000 and 1800 cal BC, spilling over into the preceding and succeeding centuries. 11 other charcoal dates ranged from the 6th to the mid 2nd millennia BC, with a basal peat date belonged in the 1st millennium BC. Only two of the dates from pieces of charcoal at Calanais fell in the part of pollen sub-zone 2c (starting at some date between 4000 and 3750 cal BC and ending at some date between 3490 and 3020 cal BC) when cereals were grown somewhere near Calanais Leobag. They came from Areas A and F by the Avenue to the north of the Ring. The group of 13 dates between about 3350 and roughly 2900 cal BC from Areas S, D, H and B in and next to the Ring demonstrated pre-Ring activity, perhaps at some distance from where the Ring was to be built. The dated charcoal belonged to the latter part of sub-zone CaN-2c when agriculture was practised around Leobag, or to sub-zone CaN-2d when there was no evidence for cultivation in the pollen columns at Leobag (Illus 23.1). Only at the beginning of CaN-3a was there again evidence for cereal growing around Calanais itself. The on-site pollen sampling showed

that pollen sub-zones 3a and 3b certainly included archaeological phases 5 to 10 and very probably most or all of phases 11 to 12. Three dates from cereal grains found in the capping of the cairn fell in pollen zone CaN-3c, when agriculture was practiced continuously in the area around Leobag; in fact the cereal dates from Calanais agree (of course approximately) with those estimated for the cereal pollen maximum in Leobag peat column CN3. 23.4.2 Phasing summary The interpreted phasing at Calanais is as follows. Table 23.2Â Archaeological phases 0 1

2a 2b 3 3 4 5a 5b

Pollen evidence for woodland clearance in CaN-2a and 2b at Calanais Leobag. Forager activity in Avenue area?

Foraging or early farming in Avenue area? Early farming in Avenue area and near Calanais Leobag.? Early farming in Avenue area?

Deposition of charcoal and early to middle Neolithic pottery in settlements, later imported with soil found in the area of the later Ring. Early ditch

Cultivation beds and subsequent flatter cultivation Turf formation

Monolith erection

Ring erection

Clay and soil imported with Phase 3 artefacts and charcoal. Green clay platform and light timber structures

Interrupted slot

Chambered cairn

Enclosure Stage 1 building Enclosure Stage 2 building Radiocarbon dating \ 944

Calanais Survey and Excavation, 1979-88

10e Ground working and burials

TPQ for the context in which it was found. The date fell in zone 2b. The birch pollen percentages at Calanais Leobag dropped suddenly in the latter part of zone 2a, perhaps reflecting human interference and recovered in 2b, although still subject to fluctuations. The birch charcoal from Area A at Calanais may reflect a human hand in those fluctuations even though the posthole may have been much later.

11b Ground working including ploughing

23.5.2 Phase 2a Early farming in Avenue area?

Pollen evidence for woodland clearance in CaN-2a and 2b at Calanais Leobag.

10a Ground working and burials 10b Rows start now?

10c Ground working and burials 10d Avenue starts now?

11a Enclosure Stage 3 building 12a Enclosure Stage 4 building

12b Ground working including cultivation 13

Kerb erection

Encroachment of peat

14 16 17 18

Cultivation

Disturbance

Peat growth and post-medieval peat clearance 19th excavations and presentation

23.5 Phases 1 to 3: Foragers and Early farmers 23.5.1 Phase 1 Forager activity in Avenue area? SUERC-11588; 6295±35 and SUERC-11989 6245±35 came from a piece of birch charcoal (sample 269) from black greasy clay 018 at the bottom of possible posthole 015 on Area A. The posthole underlay beach pebbles 006 immediately to the east of the glacial erratic Q. The ages were measured on the same piece of birch charcoal; the laboratory re-measured it because the first result was unexpectedly early. The ages can be combined to produce a calibrated date between 5305 and 5215 cal BC. The conclusion of discussion of Area A (Chapter 6) was that the birch charcoal was somewhat more likely to come from human activities than not, but that it might not provide a close

SUERC-11589; 4880±35 came from a piece of birch charcoal (sample 4277) from possible posthole (093) on Area A. This possible posthole underlay beach pebbles 006 immediately to the east of the glacial erratic Q. The age suggests the birch (if short-lived, or if the charcoal came from tree rings which grew shortly before death) was cut between 3720 and 3530 cal BC. It belonged to the early period of farming attested by cereal pollen in pollen sub-zone CaN-2c at Leobag. The charcoal may reflect farming activities in or near the Avenue area, even though the posthole from which it came could have been much later. 23.5.3 Phase 2b Early farming in Avenue area? SUERC-11601; 4760±35 came from a piece of hazel charcoal (sample 256) from a dark greyish brown, slightly gritty clay (648) which was cut by a stake hole and overlay a ring of material filling a depression on Area F near the western side of the avenue. It was dated to between 3640 and 3380 cal BC. This fell mainly in the later part of Leobag pollen zone 2c in which cereal farming was still practised. An attempt to combine SUERC-11589 (Phase 2a) and SUERC-11601 (Phase 2b) failed (Chisquared 5.9 against an expected maximum of 3.8 if the ages were the same). Thus the activities reflected by the occurrence of charred wood on Areas A and F were very probably not of the same date as each other. It is of course possible that the charcoal in Area A (and indeed that in Area F) was residual so none of the dates need provide a close terminus quem for the contexts in which they were found.

Radio carbon ating \ 945

Calanais Survey and Excavation, 1979-88

Illus 23.1 The calibrated ages (one pair of early dates from Area A has been omitted) 23.5.4 Phase 3 Early farming in Avenue area? There is no particular reason to suppose that a piece of hazel charcoal form Area F (sample 257/80) was as residual as the Group 3 dates considered below. It was found in a charcoal-rich layer 649 forming part of a turf line on Area F alongside the Avenue. It produced an age (AA-24965; 4385±50) which calibrated to between 3330 and 2890 cal BC. It could belong around the date of the transition of Leobag Pollen zone 2c to 2d, in the former of which cereal pollen occurred. It hints at the possibility that some of the clays with charcoal producing Group 3 dates inside the Ring were brought from the Avenue area. 23.5.6 Phase 3 Charcoal dated c. 3300-3000 BC inside the Ring No constructional activities could be directly associated with a group of 13 dates which attest to the period of growth, in the last third of the 4th mil-

lennium, of those tree-rings which were charred. The charcoal all belonged in a period from about 3300 to 2900 cal BC, between a half and one millennium earlier than the contexts from which the samples came. An attempt to prove a null hypothesis that all of the pieces of charcoal could have had exactly the same age as each other failed, although only marginally. The ages fell on the notorious calibration plateau in the last few centuries of the 4th millennium. In this period charcoal of very different true ages would have produced similar radiocarbon dates. Thus although the charcoal might have been produced over a fairly short period it could equally have come from the wood of plants which died centuries apart. The context details in what follows are organised by stratigraphy rather than by age. Group 1: charcoal from contexts under the chambered cairn. SUERC-11600; 4515±35 came from a piece of alder charcoal (sample 592) from a layer of green clay 398 onto which rested the Radiocarbon dating \ 946

Calanais Survey and Excavation, 1979-88

base layer of the cairn on Area D. This layer also contained a significantly later age SUERC-11598 (4390±35) which calibrated to between 2900 and 2620 cal BC. Several ages came from contexts associated with a slot cut into the green clay under the cairn. SUERC-11607; 4490±35 and SUERC-11606; 4455±35 came from a piece of hazel charcoal (sample 2020) and from a charred hazel nut shell (sample 2011) from a dark grey/green gritty charcoal rich material 883 with rotted stone and a few tiny fragments of burnt bone under the bottom course of wall-stones in the passage in Area BIVWX. SUERC-11606 had a surprisingly low d13C value of -28 ppm more characteristic of peat than of wood charcoal. SUERC-11616; 4430±35 came from a charred hazel nut shell (sample 2038) from dark grey organic material 772 spilling out of the top of slot 773 under infill 770 of the chamber wall on Area H, while SUERC-11617 4425±35 came from a charred hazel nut shell (sample 2051) from the underlying slot fill, a slightly gritty, very humic clay 778. SUERC-11596; 4495±35 from a piece of alder charcoal (sample 2045-2) and SUERC-11592; 4465±35 from a piece of pine charcoal (sample 2045-1), both came from a greasy orange/brown clay 360 under the chamber wall on Area D. The layer appeared to have been deliberately laid for the overlying course of cairn boulders. The green clay 398 which they overlay produced a much younger date as well as that described above. The cairn itself also produced two much younger ages, and it is clear that the charcoal in the greasy clay 360 was residual. Group 2, charcoal from pre-plough-soil contexts south of the cairn. SUERC-11611; 4450±35 came from a piece of hazel charcoal (sample 2084) from a black, slightly sandy clay 352 on Area D, which developed into a series of small features outside the limit of cairn in the SW quadrant of the stone ring. It underlay the plough soil 315 and it overlay a roughly rectangular area of mottled black/brown clay 389. SUERC-11597; 4495±35 from a piece of pine charcoal (sample 688/2567-1), and SUERC-11599; 4475±35 from a piece of alder charcoal (sample 688/2567-3), came from soil 369 also at the base of the plough soil 315 on Area D.

All these dates came from samples in contexts overlying a turf line which had continued to grow until pollen zone 3b, which probably started between 2560 and 2200 cal BC and ended some date between 1900 and 1500 cal BC. This zonation matched a considerable amount of other evidence. Thus the clays containing the samples must have been moved from elsewhere along with the charcoal found in them. Almost all of the potsherds from these contexts were early or mid Neolithic but they included probable and definite beaker sherds interpreted as introduced during subsequent ground-working (Technical Note 24.4.4). Group 3, charcoal from plough soil. Two samples from Area S; SUERC-11608; 4510±35 came from a charred hazel nut shell (sample Bag 4) and SUERC-11618; 4450±35 from a piece of hazel charcoal (sample Top Block). Although there was no stratigraphic connection between the soil on Area S and the plough soil 315 on Area D, which contained much later charcoal, they seem to have represented at least part of the same activities. Group 4, charcoal from a depression. SUERC-11612; 4475±35 came from a piece of hazel charcoal (sample 2352) from sandy clay 746 above green clay 1005 filling a depression above a dip in turf line 1009 on Area H. It immediately underlay the locally uppermost turf line, although technically unrelated to more widely spread layers during excavation it was thought to be contemporaneous with turf line 771, possibly equivalent to a turf line on the green clay platform under the cairn. 23.5.7 Discussion of Phase 3 charcoal The Phase 3 charcoal on Area F came from a coherent spread, making it likely that it was produced at much the same time as the context. Area F produced several pieces of Hebridean incised pottery and broadly speaking their date could have been much the same as that of the charcoal. The mineral soil there does seem to have been created partly by cultivation and a few ard marks survived. There is enough evidence to allow a suggestion that farming or some similar activity was undertaken in the Avenue area by people using Hebridean wares at some date in the centuries around 3000 BC, but not enough to prove the point. Radio carbon ating \ 947

Calanais Survey and Excavation, 1979-88

If the activities which created abundant charcoal some date around the last third of the 4th millennium had taken place inside the Ring one might have expected some charcoal to have found its way into basal ditch fills, but they contained none. On the other hand if the dirty clays containing that charcoal were imported to the place just before the cairn was built, as seems most likely, they have no bearing on the absence of charcoal in the ditch fills. Maybe the dirty clays in the Ring came from near Area F, given the similarity of the date from Area F to the later ones in the dirty clays. There was not enough evidence from Area S to absolutely preclude the possibility that the ploughing there had brought up pieces of charcoal from an in-situ 4th millennium context but it seems unlikely given the evidence from the other dates considered here. The same is true of the date from depression 746 on Area H. Five of six pollen samples from dirty clay 360 under the cairn contained spectra characteristic of sub-zones 3ai, and the other of 3aii. Subzone 3ai started some date between 2980 and 2500 cal BC while subzone 3aii ended some date between 2560 and 2200 cal BC. The transition from 3ai to 3aii, although not directly dated, probably took some date between 2750 and 2300 cal BC. Perhaps the clay was moved some date close to that period. The preferred explanation is that the charcoal came from reworking of old soils, perhaps in the Avenue area The other possibility is that the soil and the charcoal in it was taken from caches of â&#x20AC;&#x2DC;ancestralâ&#x20AC;&#x2122; soils in chambered cairns, cultivation soils or domestic contexts, and brought to Calanais to connect outlying communities with the cairn. 23.6 Phases 4 to 5: Pre-Monolith activities 23.6.1 Phase 4 Early ditch No radiocarbon dates were associated with the ditch, which was only superficially explored. There was no evidence that it was related in date to the Phase 3 samples. It underlay cultivation beds ascribed to the end of CaN-2d or more probably early CaN-3i so was definitely earlier than about 2500 cal BC (probably considerably earlier) and

may for instance even have belonged in CaN-2c, which started at some date between 4115 and 3750 cal BC and ended some date between 3490 and 3020 cal BC. 23.6.2 Phase 5a Cultivation beds The cultivation beds were not dated directly. Their soils contained pollen of CaN-3ai and they immediately overlay soils of CaN-2d. Overlying soils contained pollen ascribed to CaN-3aii and later. Their construction thus started some date between 2980 and 2500 cal BC and they had been abandoned before some date between 2770 and 2350 BC. There was some evidence for a period of levelling following their abandonment. 23.6.3 Phase 5b Turf formation The turf which formed over the cultivation beds appears from the north section on Area D to have belonged near the start of pollen sub-zone CaN3ai, at some date between 2980 and 2500. However, the turf lines in the south part of Area D were highly complex because some had formed on cultivation beds, subsequently truncated, and other in slowly filling troughs between the cultivation beds. The same may have been true of those in Area H. The soil above early turf line 766 under the cairn may have been spread from pits dug for the Ring stones and it surprisingly early pollen characteristics (most similar to CaN-2d) suggest that it included mostly residual pollen from the soils and clay through which the pits had been dug. A date for the underlying turf line of initial CaN-3ai (at some date between 2980 and 2500 cal BC) is credible. 23.7 Phase 6 Monolith, Ring and other early 3rd millennium constructions 23.7.1 Phase 6 Introduction Phase 6 includes a slot, central monolith erection and Ring erection. Charcoal from these layers was dated to after 2910 cal BC (earliest 2 sigma limits). The youngest possible age for these activities, given the possibility that all of the charcoal was Radiocarbon dating \ 948

Calanais Survey and Excavation, 1979-88

significantly residual, is provided by the chambered cairn which must date to after 2500 cal BC given the securely stratified early International Beaker sherd in the cairn makeup. 23.7.2 Phase 6a Monolith erection There was no charcoal which could be shown to go with erection of the central monolith. It was later than the cultivation beds so it was erected after c. 2980 cal BC. Its stratigraphic relationship to the other elements of Phase 6 is ambiguous. It was earlier than the chambered cairn. 23.7.3 Phase 6b Ring erection In previous publications a piece of willow charcoal (sample 230/81; AA-24969; 4095±45) was ascribed to the generally clean green clay fill 767 of the pit of Ring stone 42 in Area H (Ashmore 1999a, 2000). A thorough check of primary and early post-excavation records in November 2011 showed that this sample came from secondary capping of the cairn. Thus the only constraint on the earliest possible date for Ring stone erection is that it post-dated the cultivation beds and formation of the bottom turf line 766 on them. That means the Ring could date a century earlier than the previously suggested constraint, perhaps as early as 2980 cal BC, or in round terms 3000 BC. 23.7.4 Phase 6c Early slot on Area H (along with discussion of charcoal from a fire on Area B) The interrupted truncated slot 795 on Area H may represent a light timber structure. It was later than abandonment of the cultivation beds because it cut turf line 766 above the earliest soil 792. But although not the preferred interpretation the slot may have been cut from higher up if its top had been destroyed by ground-working of soil 777, which is interpreted as forming in spoil above turf line 766 from the digging of the Ring stone pits. The slot produced a date of between 2910 and 2630 cal BC from a piece of hazel charcoal (sample 2256/81; AA-24970; 4205±45). The date

provides a terminus post quem for filling of the slot. Because an original upper part of the slot may have been destroyed (although that is not the preferred interpretation) it does not provide a secure terminus post quem for overlying soil 777. Its chronological relationship to the central monolith and Ring is unclear, again because it could have been cut through soil 777, but it must have been earlier than turf line 758, the green clay platform and the light timber structure built on top of it. On Area B a piece of birch charcoal from part of charcoal concentration 871 was dated to between 2880 and 2580 cal BC (sample 2368/81; AA-24959; 4140±45). A fire had been lit on a cultivation surface then dug over at least once. The cultivation soil may have been the same as soil 792 on Area H which was cut by the slot of the early light timber structure, but there was some ambiguity in the relationship. The perils of collating turf lines in different parts of the site have been emphasised already, and the stratigraphy of Areas H and B around the base of Ring stone 42 was very complex. The concentration of many small pieces of charcoal 871 does suggest a higher than normal likelihood that the charcoal was not residual. It could represent a casual camp fire lit shortly before cessation of ground-working. But it may for instance represent the burning of the decayed remnants of a light timber structure such as that postulated to explain slot 795 and may thus be significantly older than the last use of the plough soil. Charcoal (AA-24958 4065±45) from a shallow feature 815 under ‘paving’ inside the enclosure produced a date of 2860 to 2470 cal BC. However, the charcoal was residual, because the feature also contained glass fragments. The feature was not all that far away from the fire in the grey soil B871 which contained charcoal of an age which was not significantly different. It seems quite possible that that was where the charcoal originated. 23.7.5 Discussion of Phase 6 features Technically there was no direct and unambiguous stratigraphic distinction between these features. The preferred interpretation is that the slot and the Ring pits were cut from the ground surface repRadio carbon ating \ 949

Calanais Survey and Excavation, 1979-88

resented by early turf line 766. Some period must be allowed for creation and abandonment of the cultivation beds, which had the same dating limits. Overall a date for Ring building somewhere in the time span 2950 to 2650 BC seems most likely. The monolith had the same stratigraphic relationship to the turf lines as did the Ring so it seems unlikely to have stood alone for a long period. Thinking more generally it would have been more difficult (but not impossible) to erect the monolith once the Ring had been set up. Given the preferred interpretation on rather general grounds that the monolith was erected before the Ring the inference (and I use that in a slightly pejorative sense) is that it was erected at most a few generations earlier than the Ring; but strictly speaking they could have been put up in a single extended operation. 23.8 Phase 7 clay spreading and light timber structures Phase 7 includes the spreading of clay within the Ring and erection of a light timber structure or structures. 23.8.1 Phase 7 Clay-spreading The several different materials spread inside the Ring included green clay. The preferred interpretation is that it was imported because the ground on which it was spread had become water-logged. The best preserved area of pre-cairn spread clay survived on Area H where its lower part contained pollen of sub-zone 2d overlying soil pollen-zoned to the later 3aii. Less well preserved, and somewhat confused with the green clay of the mound round the base of the central monolith, was a layer of green clay 398 on which rested the base layer of the cairn in Area D. A piece of birch charcoal (sample 592/81) from this layer produced an age SUERC-11598 (4390±35) which calibrated to between 2900 and 2620 cal BC. Unfortunately, given how many pieces of charcoal from Calanais proved to be old when buried, and the fact that this age was significantly older than that from the stratigraphically earlier Ring pit, it seems all too possible that this

too was residual. If it in fact had come from clay mound originally round the base of the central monolith, spread out subsequently, the same applies. It provides a TPQ for the cairn, but one probably 300 to 500 years earlier. Clay 369 a layer of soil at the base of the plough soil 315 in Area D has already been discussed. It produced residual charcoal with Phase 3 dates. It contained 11 potsherds, mostly of early to mid Neolithic type and the exceptions may well have been intrusive given the number of ard marks at the same level. It also contained quartz artefacts. This contrasts with the more mixed assemblage including beaker pottery in the plough soil above. It also produced a significantly later age (AA24964; 4185±45) from a piece of hazel charcoal (sample 688/2567-2) which calibrated to between 3100 and 2910 cal BC. This age is not in a technical sense significantly different from that of a piece of willow charcoal in secondary levels of the chamber wall, albeit older. 23.8.2 Phase 7 Light timber structures At the risk of becoming monotonous, I have to say that contexts 772, 778 and 883 forming the structured fill of the light timber structure which preceded the chambered cairn contained highly residual charcoal, ages from which were discussed along with others assigned to Phase 3. 23.9 Phase 8 Chambered cairn construction The body of the cairn in Area H produced two dates. An age from a piece of birch charcoal (sample 116/81; AA-24966; 4210±50) from the infill 732 of the cairn provided a fairly unhelpful TPQ of between 2910 and 2630 cal BC for construction. A piece of hazel charcoal (sample 2243/81; AA-24967; 4050±45) came from a closely related context 736. Its field interpretation (not now the preferred interpretation) was a possible token burial in the cairn and the preferred interpretation is as an area scorched by a fire lit during cairn construction. It provided a TPQ between 2860 and 2470 cal BC. Although these two contexts should each date to the within a few days of each Radiocarbon dating \ 950

Calanais Survey and Excavation, 1979-88

other the ages are significantly different from one another (chi-squared = 5.2 against an upper limit of 3.8 if the two ages were the same). The older date was thus from residual charcoal. Even the younger of the two dates seems to have been from a piece of residual charcoal, for context 736 also contained a well-stratified all-over-corded beaker sherd which cannot be earlier than 2500 cal BC. It also contained a small body sherd of â&#x20AC;&#x2DC;early to mid neolithicâ&#x20AC;&#x2122; type (PC18) emphasising that residual material had been included in this context. 23.10 The Phase 9 Stage 1 and Stage 2 enclosures and Phase 10 pit digging, ground working, and possibly the start of Row and Avenue building Ground-working and ploughing seem to have occurred at several times after construction of the chambered cairn. A major episode of ploughing occurred after the Phase 9 Stage 1 enclosure had gone out of use. If it be accepted that the various soils labelled 112 on Area B were related to that ploughing of the Stage 1 enclosure bank then the East Row post-dated the ploughing. 23.10.1 Phase 9 Stage 1 and Stage 2 enclosures There was no direct radiocarbon-dating or pollen-zonation evidence for dating this phase. The Stage 1 and Stage 2 enclosures were interpreted as later than the Ring and very probably, albeit not certainly, later than the chambered cairn. Dark brown plough soil 315 in Area DI, labelled 344 in and near Area DV overlay abundant ard marks commonly filled with charcoal-rich material. It seems to have incorporated artefacts and perhaps also charcoal associated with dilapidation of the chambered cairn. It is possible that some deposition of soil and associated occupation material from a nearby settlement of settlements took place when the Phase 7 light timber structures were in use, because two of five dates from the soil attested to incorporation of material earlier than the date of the cairn. One from 315 came from a piece of hazel charcoal (sample 687/81; AA-24960, 4205Âą50) dated

to between 2910 and 2630 cal BC. A piece of birch charcoal (sample 685/81 AA-24961, 4055, 50) dated to between 2860 and 2470 cal BC. It had an anomalously negative d13C value of -30.6. It may have derived from birch preserved in peat. On the other hand these samples could represent material imported in soil offerings from earlier settlements, in which case they would belong to Phase 10a. 23.10.2 Phase 10a Pit digging and ritual deposition Few of the pits in Area D south of the cairn were pollen-sampled or radiocarbon-dated, but some of them were first noted after removal of the Phase 6b turf line. It seems fairly likely that their tops were missed during excavation because the turf over them healed over fairly rapidly; although it remains a formal possibility that some were created even before turf formed over the clay spread during Ring-building that is not the preferred interpretation. One of the pits cutting both of the old ground surfaces under spread clays on Area D contained pollen assigned to sub-zone CaN-3a (Pit 386 sample 2028). A piece of pine charcoal (sample 2006; SUERC-11590, 3965+/-35) dated to between 2580 and 2340 cal BC. From plough soil 344 came another piece of pine charcoal (sample 95; SUERC-11591, 3915+/-35) dated to between 2490 and 2290 cal BC, and a piece of birch charcoal (sample 95/81; AA-24962, 3555+/-50) dated to between 2030 and 1750 cal BC. The piece of pine charcoal (sample 95) dated to between 2490 and 2290 cal BC and the piece of birch charcoal (sample 95/81) dated to between 2030 and 1750 cal BC in the plough soil 315 could possibly belong with this phase. But this must remain speculation because ground working and pit digging continued sporadically for such a long time. 23.10.3 Phase 10a Ground working and burials At least one small stone setting and other pits were found at the base of the plough soil. A pit was also Radio carbon ating \ 951

Calanais Survey and Excavation, 1979-88

found in the body of the plough soil. Digging of small pits and ground-working continued south of the cairn in Area D to after 1800 cal BC, perhaps much later, judging by pollen characteristic of sub-zone CaN-3c in plough-truncated Pits 385 and 392. More generally the fancier artefacts in the plough soil are best explained as coming from burial deposits. The origin of the many small worn pieces of earlier styles of pottery is more open to speculation. The preferred interpretation is that they represent ancestral soils brought to Calanais from nearby settlements for ritual reasons, rather than deriving from pragmatic enrichment of the soil with baskets of local plough-soil which just happened to contain old pottery (see the discussion of Area E). Pit digging and ploughing must have lasted until some date, at earliest 1800 cal BC and possibly centuries later. A piece of Pomoideae sp charcoal (sample 54/80; AA-24956 3580Âą45) dating to between 2120 and 1770 cal BC was found in an ard mark cut into green clay 123 on Area B. The clay was at the base of the local sequence and the charcoal appears to have dropped into the ard mark during ploughing. Although it may represent burning of undergrowth immediately prior to ploughing it was abraded and it seems more likely that it was lying around in the soil prior to ploughing. 23.10.4 Phase 10b Rows now? There was no direct evidence for the dating of Stones 30, 31 and 33a of the East Row, but the pit for Stone 30 cut clay 812. Pollen in a sample (2054) from 812 suggested a date in sub-zone CaN-3a which ended some date between 2560 and 2200 cal BC, providing a terminus post quem of 2560 for erection of the stone. A piece of well preserved heather charcoal from overlying layer 139 produced a radiocarbon date (AA-24957) between 1940 and 1690 cal BC. Because the heather twig was very fragile it seems very likely to have been contemporary within its context and the date thus suggests that Stone 30 had been set up before, perhaps well before, 1690 BC. Two samples (2021, 2030) from this overlying layer 139 produced pollen characteristic of subzone CaN-3c which started between 1900 and

1500 cal BC. Overall the evidence suggests that Stone 30 was erected some date between 2560 and 1690 cal BC. At first glance this suggests a theoretical possibility that it could have been set up before the Ring, because Stone 42 had a terminus post quem of between 2880 and 2490 cal BC. However the stratigraphy on Area B demonstrated that clay 182 was later than turf line 162 which post-dated the Ring. The dating does leave open the possibility that Stone 30 was erected before the chambered cairn, which must date to after 2500 cal BC, but the weight of the radiocarbon determination suggests a greater probability that it was erected after the cairn was built. There was very little evidence to suggest when Stone 33a at the east end of the East Row in Area C was erected. It appears to have occurred after, perhaps well after, the transition of zone CaN-3a to CaN-3b at some date between 2560 and 2200 cal BC, because its pit was cut into mineral soil 205. 23.10.5 Phase 10c Ground working and burials No dating evidence 23.10.6 Phase 10d Avenue starts? Excavation of the area southwest of Stone 34, which has variously been regarded as part of the east side of the Avenue and as a stone erected with Stone 9 to mark an astronomical alignment, produced no useful radiocarbon or pollen dating evidence. 23.11 Phase 11 The Stage 3 enclosure and ground working 23.11.1 Phase 11a The Stage 3 enclosure There was no direct radiocarbon-dating or pollen-zonation evidence for dating this phase. Interpretation of an age (AA-24968 3575Âą 45) from a piece of willow charcoal in greasy clay 738 outside the cairn on Area H, together with the inverted pollen zonation of an old turf line 751 and Radiocarbon dating \ 952

Calanais Survey and Excavation, 1979-88

overlying soils led to a hypothesis that the material of the cairn body was pulled down onto area H and the stones removed for building elsewhere. Perhaps the stones were taken for the wall-base of the second enclosure. That would allow the age, which calibrated to between 2040 and 1770 cal BC, to provide a terminus post quem for the Phase 11a (second) enclosure. 23.11.2 Phase 11b Ground working The state of the remains of the third stage enclosure strongly suggested robbing and ploughing before the third stage enclosure wall-base was laid out, but there was no direct radiocarbon-dating or pollen-zonation evidence for dating this phase. 23.12 Phase 12 The Stage 4 enclosure and ground working 23.12.1 Phase 12a Fourth Stage enclosure No radiocarbon or pollen evidence was discovered for the date of this stage of enclosure. Stratigraphically it could have been part of Victorian or later site management, except that the lack of a southern side suggests that it was damaged by ploughing and robbing. If that be accepted its partial destruction may be related to the Phase 12b ground-working described below or to Phase 14 cultivation 23.12.2 Phase 12b Ground working A piece of heather twig charcoal (sample 132/80; AA-24957 3495±45) dated to between 1940 and 1690 cal BC was in a charcoal-rich spread in context 139 on Area B alongside the east alignment of standing stones. It is unlikely to have been residual because of its fragility. Ploughing may been renewed north of the east alignment, with first burning of ground cover and then field clearance leading to creation of the line of dumped stones 134 which overlay charcoal spread 139. It must be emphasised that although I have interpreted the event which led to deposition as belonging after Phase 12, to explain destruction of the southern part of the Phase 12 enclosure, it could have gone with damage to the Phase 11a Stage 3 enclosure.

23.13 Phase 13 Secondary cairn and kerb slab erection A piece of willow charcoal (81.230) from secondary chamber wall capping 768 was dated (AA24969 4095±45) to between 2880 and 2490 cal BC. In earlier publications (Ashmore 1999a, Ashmore 2000) it had been wrongly ascribed to context 767 (Technical Note 12.8.5). Along with it were various pieces of pottery including Early to Middle Neolithic corky wares and three pieces probably from fine Beakers. The sherds suggest that the corky wares and probably also the charcoal were residual. As described above, clay 738, laid down at the north edge of the cairn, contained a piece of willow charcoal (sample 203/81; AA-24968 3575±45 BP) dated to between 2040 and 1770 cal BC. The clay overlay olive-green clay 750, probably equivalent to the green clay under the cairn, and a turf line 758 into which were pressed the shallow impressions of the first kerb of the chambered cairn on Area H. It was suggested above in the context of the enclosure (Phases 10 to 12) that the northern part of the cairn may have been used as a quarry for stone. The clays from the body of the cairn were left behind in the area north of the cairn, judging by the pollen content of the soils overlying greasy soil 738. Clay 738 seems to have been deposited after removal of the early laid kerb. The later, large kerb slab on Area H was set into clay 738 and the date thus provides a terminus post quem for its erection of 2040 cal BC. The cereal grains in the capping of the cairn (Phase 14 cultivation) may have been put there in turf added during the Victorian period but another possibility is that they were part of the fills put on the cairn when the kerb slab was erected. That would indicate a terminus post quem of about 1530 cal BC. 23.14 Phase 14 Cultivation and Phase 15 encroachment of peat 23.14.1 Phase 14 Cultivation Three cereal grains dated at Calanais came from the capping 708 of the cairn 704 in Area H (Technical Radio carbon ating \ 953

Calanais Survey and Excavation, 1979-88

Note 12.2.7). They all produced dates very similar to one another. A charred barley grain (sample 764-2 SUERC-11610 at 3220±35), a charred wheat grain (sample 764-1 SUERC-11609 at 3215±35 and a charred barley grain (sample 374 SUERC-11602 at 3195±35 BP) from 10/1 F15, dated to 1610 to 1410 cal BC, 1610 to 1410 cal BC and 1530 to 1400 cal BC respectively. The context may represent filling up behind the massive kerb slab in the second millennium BC, or possibly Victorian capping of the cairn. The former explanation is preferred because of the lack of peat fragments below it. The grains probably came originally from a plough soil not far from the Ring. Given the widespread evidence for ploughing in various areas at Calanais it seems very possible that this cereal-growing and charring was local. Thus the grain dates hint at the period of some of the destructive ploughing in and around the Ring after the cairn had become dilapidated, although this has no direct support in the evidence. The dates from the grains, all between 1525 and 1420, coincide (approximately, of course) with the cereal pollen maximum at Calanais Leobag in pollen zone CaN-3c, when agriculture was practiced continuously in the general area around Leobag. 23.14.2 Phase 15 Encroachment of peat An age GU-1403 (sample Call80/54/244) from humic acid of basal peat on Area C suggests, after the error has been increased from ±65 to ±110 C14 years (Ashmore et al 2000) that peat was forming over the site at some date between 1050 and 400 BC with a two out of three chance that it was forming at some date between 930 and 550 cal BC. The basal peat on Area C was attributed to Zone 3e through pollen analysis of spot samples. In the Kubiena box column the transition from soil to peat was assigned to CaN-3d. Two peat samples from sub-zone 3e in Leobag column CN1 were dated. After correction of the errors the ages were GU-1289 (2440±112 BP) and GU-1170 (2355±110 BP). Perhaps only because of the large errors attached to the dates the Leobag dates are not significantly different from GU-1403 when they are considered as a group.

The interpolated date for the start of CaN-3e in CN1, at between 920 and 260 cal BC, was not very different from that provided by the direct radiocarbon date for basal peat at Calanais, and most of the first few centimetres of peat at Calanais may indeed have formed in CaN-3e. Because the pollen zone boundary reflects a change in the local vegetation the true date for the start of peat growth at Calanais probably did lie between 920 and 400 cal BC (the earlier limit from the pollen zone 3d to 3e transition date and the later from the direct radiocarbon date GU-1403). 23.14.3 Discussion of the Phase 15 dates Johnson and Flitcroft’s excavations of a buried field system near Calanais farm in 1999 and 2000 to the south-west of the main setting at NB21253265 ((Flitcroft et al 2000, 102-3; Johnson et al forthcoming) produced peat initiation dates which were in a statistical sense significantly different from that for Calanais. The basal peat column dates at Calanais Farm suggest, in round terms, that peat started growing at around 500 BC (perhaps around the start of Leobag pollen zone transition CaN-3d to 3e much as at Calanais itself ) and at some date after 250 BC. It is possible that peat growth at Calanais itself was diachronic, depending on land-use. Growth might, for instance, have started earlier on pasture than it did on cultivated areas. The evidence for 1st millennium AD activities at the Ring hints that steady growth of peat started later there than in the surrounding areas. 23.16 Phases 16 to 18: 1st Millennium AD disturbances, Peat growth and postmedieval peat clearance, and Post-peat excavation and presentation There was no direct or indirect scientific dating for these phases at Calanais itself. Phase 16 included possible disturbances of the cairn, their date suggested by finds of pottery ascribed to the 1st millennium AD in the passage and in disturbed levels of the chambered cairn. Pollen column CN2 at Calanais Leobag which covered part of these periods of disturbance and Phase 17 peat growth was not radiocarbon-dated. Radiocarbon dating \ 954

Calanais Survey and Excavation, 1979-88

For Phase 18, historical information suggests that the setting was steadily cleared of peat from the late 18th century onward, culminating in final clearance of the central area in 1857. But there was no scientific evidence for more precise dating of the earlier stages of this process. Nor was there archaeological evidence for the date of the various display and conservation activities which affected the site after its clearance. 23.17 Conclusions I believed until well into the post-excavation process that several well stratified samples from Calanais allowed Bayesian analysis (Technical Note 23.2.3). However the realisation that the cairn must date to after 2500 BC reinforced the idea that even the well stratified charcoal samples were significantly older than the contexts in which they occurred, the most likely exception being the fragile heather twig from near the East Row. This re-emphasises the point, alas still not respected in some modern archaeological publications, that it is absolutely crucial to treat the date of a sample even in a well-sealed context as providing nothing better than a terminus post quem for the context (Ashmore 1999, 2000).

Nevertheless the combination of direct radiocarbon dating, indirect radiocarbon dating via pollen sample zoning and, in the very important case of the cairn, the artefact dating has allowed a better chronological scheme for the central monolith, Ring and chambered cairn than existed before excavation. It has allowed indirect dating of the cultivation beds and turf lines revealed by excavation. It has narrowed down the period during which the East Row was set up. There was no datable charcoal directly related to the various stages of enclosure, or the Avenue, nor were spot samples useful in producing pollen zone information for those structures. However, the barbed and tanged arrowheads in the entrance silts of the Stage 2 enclosure did reduce the range of likely dates for its use to somewhere within a few centuries of 2000 BC while a date from cairn-demolition activities which may have been associated with enclosure building provided a tentative terminus post quem of between 2040 and 1770 cal BC for the Stage 3 (first stone-based) enclosure. Its robbing may have coincided with cairn rebuilding for which it can be argued cereal grain in the cairn capping provides a terminus post quem of 1530 cal BC. Comparisons with dates ascribed to other sites can be found in Chapter 24: Discussion.

Radio carbon ating \ 955

Calanais Survey and Excavation, 1979-88

23.18 Table of the radiocarbon ages, contexts and calibrations Table 23.3Â The samples and their ages and calibrated dates Description

Code

Age

Error

d13c

1 sigma

2 sigma

A piece of birch charcoal (sample 269) from black greasy clay 018 at the bottom of possible posthole 015, the upper fills of which included a root mass 016 over dark brown fibrous peaty clay 017 SUERC-11588 which also contained charcoal (sample 370/380). This posthole underlay beach pebbles 006 immediately to the east of the glacial erratic Q on Area A of the 1980 excavations.

6295

-27.1

5315 to 5225 5350 to cal BC 5210 cal BC

A piece of birch charcoal (sample 269) from black greasy clay 018 at the bottom of possible posthole 015, the upper fills of which included a root mass 016 over dark brown fibrous peaty clay 017 SUERC-11989 which also contained charcoal (sample 370/380). This possible posthole underlay beach pebbles 006 immediately to the east of the glacial erratic Q on Area A of the 1980 excavations.

6245

-27.2

5300 to 5210 5310 to cal BC 5070 cal BC

A piece of birch charcoal (sample 4277) from possible posthole 093 on Area A. It measured 0.18 x 0.12 m and contained charcoal impregnated, loose humic clay 094. It cut a small patch of SUERC-11589 angular stones 095.This possible posthole underlay beach pebbles 006 immediately to the east of the glacial erratic Q on Area A of the 1980 excavations.

4880

-26.4

3695 to 3640 3760 to cal BC 3530 cal BC

A piece of hazel charcoal (sample 256) from a very dark greyish brown, slightly gritty clay 648, with a little charcoal, on Area F which was cut by stake hole SUERC-11601 644 and overlay a ring of material 646 filling a depression. The clay 648 abuts a stone 617 otherwise entirely surrounded by peat.

4760

-25.5

3640 to 3520 3640 to cal BC 3380 cal BC

A piece of alder charcoal (sample 592) from a layer of green clay 398 onto which rested the base layer of the cairn on Area D. It overlay upper layer 095, one of several OGS under the cairn and SUERC-11600 chamber. The green clay 398 is interpreted as possibly redeposited parts of the capping over paving 381 around Monolith 29.

4515

-26.7

3350 to 3110 3360 to cal BC 3090 cal BC

Radiocarbon dating \ 956

Calanais Survey and Excavation, 1979-88

Description

Code

Age

Error

d13c

1 sigma

2 sigma

A charred hazel nut shell (sample Bag 4) from Area S 13, a dense dark grey and charcoal rich material at SE corner SUERC-11608 of the 1988 trench, appearing to extend under the baulks to S and E, which is to say towards the chambered cairn.

4510

-25.8

3340 to 3100 3360 to cal BC 3090 cal BC

A piece of alder charcoal (sample 20452) from greasy orange/brown clay 360 under the core cairn forming the side wall on Area D in the SW quadrant of the stone ring. The layer here was generally about 0.06 to 0.1 m thick, with quartz chips, which appeared to have been deliberately laid for the overlying course of cairn boulders. However, it was very variable in composition and SUERC-11596 contained a piece of wood (257/81), which is presumably intrusive. Soil 360 also contained two sherds of pottery, one of “late prehistoric” type (although this identification is based on fabric alone), a flint ‘microlith’ (478/81), six pieces of quartz and a piece of chert (584/81). See also a very similar date SUERC-11592 (4465±35) from a piece of pine charcoal from the same layer

4495

-26.1

3340 to 3100 3360 to cal BC 3030 cal BC

A piece of pine charcoal (sample 688/2567-1) from 369 at the base of the plough soil 315 on Area D. It contained 11 pottery finds, mostly of indeterminate affinities but including one plain early Neolithic sherd (Pottery CataSUERC-11597 logue 13) and three with a corky fabric. It also contained quartz artefacts. This contrasts with the more mixed assemblage including beaker pottery in the plough soil above.

4495

-25.7

3340 to 3100 3360 to cal BC 3030 cal BC

A piece of hazel charcoal (sample 2020) from a dark grey/green gritty charcoal rich material 883 with burnt bone and rotted stone. Under 824 = bottom course of wall-stones in passage in Area B4Wx. It overlay a dark feature 881 SUERC-11607 under the north passage wall. It may possibly be the same as fine green clay 887 which more generally underlay the passage wall. See also SUERC-11606 (4455±35 BP) from the same context.

4490

-25.3

3340 to 3090 3350 to cal BC 3030 cal BC

Radio carbon ating \ 957

Calanais Survey and Excavation, 1979-88

Description

Code

Age

Error

d13c

1 sigma

2 sigma

A piece of alder charcoal (sample 688/2567-3) from clayey soil 369 at the base of the plough soil 315 on Area D. It contained 11 pottery finds, mostly of indeterminate affinities but including one plain early Neolithic sherd (Pottery SUERC-11599 Catalogue 13) and three with a corky fabric. It also contained quartz artefacts. This contrasts with the more mixed assemblage including beaker pottery in the plough soil above.

4475

-23.5

3330 to 3090 3350 to 3020 cal BC cal BC

A piece of hazel charcoal (sample 2352) from sandy clay 746, the top fill of a SUERC-11612 depression sealed by the uppermost turf line 771 on Area H.

4475

-24.8

3330 to 3090 3350 to cal BC 3020 cal BC

A piece of pine charcoal (sample 20451) from greasy orange/brown clay 360 under the core cairn forming the side wall on Area D in the SW quadrant of the stone ring. The layer here was generally about 0.06 to 0.1 m thick, with quartz chips, which appeared to have been deliberately laid for the overlying course of cairn boulders. However, it was very variable in composition and SUERC-11592 contained a piece of wood (257/81), which is presumably intrusive. Soil 360 also contained two sherds of pottery, one of “late prehistoric” type (although this identification is based on fabric alone), a flint ‘microlith’ (478/81), six pieces of quartz and a piece of chert (584/81). See also a very similar date SUERC-11596 (4495±35) from a piece of alder charcoal from the same layer

4465

-25.5

3330 to 3030 3340 to cal BC 3020 cal BC

A charred hazel nut shell (sample 2011) from a dark grey/green gritty charcoal rich material 883 with burnt bone and rotted stone, under the bottom course 824 of wall-stones in passage in Area B4Wx. It overlay a dark feature 881 SUERC-11606 under the north passage wall. It may possibly be the same as fine green clay 887 which more generally underlay the passage wall. See also SUERC-11607 (4490±35 BP) from the same context.

4455

-28

3330 to 3020 3340 to cal BC 3010 cal BC

A piece of hazel charcoal (sample 2084) from a black, slightly sandy clay 352 on Area D, which developed into a series of small features outside the limit of SUERC-11611 cairn in the SW quadrant of the stone ring. It underlay the plough soil 315 and it overlay a roughly rectangular area of mottled black/brown clay 389.

4450

-26

3330 to 3020 3340 to cal BC 2930 cal BC

Radiocarbon dating \ 958

Calanais Survey and Excavation, 1979-88

Description

Code

Age

Error

d13c

1 sigma

A piece of hazel charcoal (sample Top Block) from Area S 13, a dense dark grey and charcoal rich material at SE corner of the 1988 trench, appearing to extend under the baulks to S and E, which is to say towards the chambered cairn.

SUERC-11618

4450

-25.5

3330 to 3020 3340 to cal BC 2930 cal BC

A charred hazel nut shell (sample 2038) from dark grey peaty material 772 under SUERC-11616 infill 770 of the chamber wall on Area H.

4430

-25.4

3310 to 2930 3330 to cal BC 2920 cal BC

A charred hazel nut shell (sample 2051) from slightly gritty, humic rich clay with SUERC-11617 ochre 778 filling a linear feature 773 under the chamber wall on Area H.

4425

-24.9

3270 to 2930 3330 to cal BC 2920 cal BC

A piece of hazel charcoal (sample 688/2567-2) from D369 at the base of the plough soil 315. It contained 11 pottery finds, mostly of indeterminate affinities but including one plain early Neolithic sherd (Pottery Catalogue 13) and three with a corky fabric. It also contained quartz artefacts. This contrasts with the more mixed assemblage including beaker pottery in the plough soil above.

4390

-27.9

3090 to 2920 3100 to cal BC 2910 cal BC

A single piece of hazel charcoal (sample 257/80) related to a charcoal rich layer 649 forming part of a turf line 630) AA-24965 which preceded the soils of F615, F622 and 643, which may be late in development on Area F.

4385

-27.2

3090 to 2910 3330 to cal BC 2890 cal BC

A single piece of birch charcoal (sample 116/81) from the infill H732 of the cairn under H 728 and above H 741; it is under H736 from which a date AA- AA-24966 24967 (4050Âą45 BP) has been obtained. It provides a TPQ for construction of the core cairn.

4210

-26.1

2900 to 2690 2910 to cal BC 2630 cal BC

A single piece of hazel charcoal (sample 687/81) from a layer of dark brown soil D315 with abundant ard marks commonly filled with charcoal rich AA-24960 material, which seems to be associated with dilapidation of the chambered cairn. See also AA-24961, AA-24962 and AA-24963

4205

-25.4

2900 to 2690 2910 to cal BC 2630 cal BC

SUERC-11598

2 sigma

Radio carbon ating \ 959

Calanais Survey and Excavation, 1979-88

Description

Code

Age

Error

d13c

1 sigma

2 sigma

A single piece of hazel charcoal (sample 2256/81) from an interrupted truncated slot H795 under the clays and soils under the cairn. It cuts the earliest soil 792 AA-24970 and although it may be residual it need not be. Provides a TPQ for the slot, and for overlying soil 777.

4205

-27.4

2900 to 2690 2910 to cal BC 2630 cal BC

A single piece of birch charcoal (sample 592/81) from the layer of green clay D398 on which rested the base layer of the cairn. Presumably contemporary AA-24964 with erection of cairn, though maybe residual. Should provide a TPQ for the cairn.

4185

-26.1

2890 to 2670 2900 to cal BC 2620 cal BC

A single piece of birch charcoal (sample 2368/81) from an area of charcoal concentration BI 871 which gave the impression that a fire had been lit on a cultivation surface then dug over once. The cultivation soil lay beneath the two main upper OGS of which one, very AA-24959 thin and hardly more than a litter layer, ran under the green clay under the cairn and over the green clay mound round the base of Ring Stone 43. It was cut by the (probably much later) socket of a late kerb stone.

4140

-26.3

2870 to 2630 2880 to cal BC 2580 cal BC

A single piece of birch charcoal (sample 678/81) from the bottom part D369, at the base of the plough soil 315 on Area D. It contained 11 pottery finds, mostly of indeterminate affinities but including one plain early Neolithic sherd (Pottery AA-24963 Catalogue 13) and three with a corky fabric. It also contained quartz artefacts. This contrasts with the more mixed assemblage including beaker pottery in the plough soil above.

4115

-25.2

2860 to 2580 2880 to cal BC 2570 cal BC

A single piece of willow charcoal (sample 230/81) from the secondary capping 768 of the chamber wall on Area HII. This charcoal was previously ascribed to fill H767 of an early pit H775, the primary pit of Ring stone 42; thorough analysis of records in November 2011 AA-24969 showed a very high likelihood that the error in labelling occurred when a find of potsherds and charcoal (81.230) was split up and the charcoal put in a new bag shortly before the charcoal was identified.

4095

-25.4

2860 to 2570 2880 to cal BC 2490 cal BC

Radiocarbon dating \ 960

Calanais Survey and Excavation, 1979-88

Description

Code

Age

Error

d13c

1 sigma

2 sigma

A single piece of birch charcoal (sample 143/81) from a shallow feature B815 in B120 (inner area of structure to east of AA-24958 stone ring). Under paving for which it provides a TPQ.

4065

-26.1

2840 to 2490 2860 to cal BC 2470 cal BC

A single piece of birch charcoal (sample 685/81) from a layer of dark brown soil D315 with abundant ard marks commonly filled with charcoal rich material, which seems to be associated with dilapidation of the chambered cairn. See also AA-24960, AA-24962 and AA-24963

AA-24961

4055

-30.6

2840 to 2480 2860 to cal BC 2470 cal BC

A single piece of hazel charcoal (sample 2243/81) interpreted as a possible token burial H736 in the cairn. It overlay AA-24967 H732 from which a date AA-24966 (4210Âą50) was obtained.

4050

-25.8

2840 to 2480 2860 to cal BC 2470 cal BC

A piece of pine charcoal (sample 2006) from 315, a plough soil on Area D which was ard marked, had many beaker sherds in it and seems to be associated with dilapidation of the SUERC-11590 chambered cairn. The layer contained a residual Neolithic potsherd. See also AA-24963 (4115Âą 45) which is significantly later.

3965

-25.7

2570 to 2460 2580 to cal BC 2340 cal BC

A piece of pine charcoal (sample 95) from a layer of dark brown/black clay 344 with charcoal and Beaker sherds found in the northern part of Trench DV inside the SW quadrant of the stone ring. One ard mark was noticed in it and there were abundant ard marks SUERC-11591 underneath it cutting into the tops of clay layers and the brown clay mound 340 round the base of Ring Stone 47. The layer 344 seems to be associated with dilapidation of the chambered cairn.

3915

-24.3

2470 to 2340 2490 to cal BC 2290 cal BC

A single piece of Pomoideae sp charcoal (sample 54/80) from an ard mark cut into green clay B123; seemingly a clay early in the sequence, although it overlies at least one OGS. Appears to have AA-24956 dropped into an ard mark, rather than having been ploughed up from below; but it was abraded and this cannot be proved.

3580

-26.6

2020 to 1880 2120 to cal BC 1770 cal BC

Radio carbon ating \ 961

Calanais Survey and Excavation, 1979-88

Description

Code

Age

Error

d13c

1 sigma

2 sigma

A single piece of willow charcoal (sample 203/81) from a green dark brown sandy clay H738 with charcoal and rotted stones. It came from under AA-24968 the cultivation soil with ard marks and beaker sherds 707 and it lies above green clay. It is cut by the trench for the later kerb.

3575

-25.8

2020 to 1880 2040 to cal BC 1770 cal BC

A single piece of birch charcoal (sample 95/81) from a layer of dark brown clay D344 thought to be the equivalent of layer D315 which contained abundant beaker sherds and seems to be associated with dilapidation of the chambered cairn. Ard marks became progressively AA-24962 more visible as this layer was removed. The charcoal may have been in the underlying layer and have been brought up by ard ploughing, or may have been introduced at the time of ploughing. See also AA-24960, AA-24961 and AA-24963

3555

-27.2

1980 to 1770 2030 to cal BC 1750 cal BC

A single piece of heather charcoal (sample 132/80) from a charcoal spread B139 under B134, the line of stones running from stone 44 (at B117, natural soil build-up). Its ultimate origin is AA-24957 uncertain. It could have come from clearing out of the chambered cairn. It gives a TPQ for the line of stone B134. Since it is a single twig it probably has not been moved around a lot.

3495

-25

1890 to 1750 1940 to cal BC 1690 cal BC

A charred barley grain (sample 7642) from the late capping 708 of the cairn. Given that the material probably came from somewhere nearby its original context was obscure. This sample was dated for the interest of the barley itself. The three cereal grains all produced dates very similar to one another: SUERC-11610 at 3220±35, SUERC-11609 at 3215±35 and SUERC-11602 at 3195±35 BP.

SUERC-11610

3220

-25.7

1520 to 1445 1610 to cal BC 1410 cal BC

A charred wheat grain (sample 7641) from the late capping 708 of the cairn. Given that the material probably came from somewhere nearby its original context was obscure. This sample was dated for the interest of the wheat itself. The three cereal grains all produced dates very similar to one another: SUERC-11610 at 3220±35, SUERC-11609 at 3215±35 and SUERC-11602 at 3195±35 BP.

SUERC-11609

3215

-25.1

1515 to 1440 1610 to cal BC 1410 cal BC

Radiocarbon dating \ 962

Calanais Survey and Excavation, 1979-88

Description

Code

Age

Error

d13c

1 sigma

2 sigma

A charred barley grain (sample 374) from 10/1 F15, from the late capping 708 of the cairn. Given that the material probably came from somewhere nearby its original context was obscure. This barley grain was dated in its own right. SUERC-11602 The three cereal grains dated at Calanais all produced dates very similar to one another: SUERC-11610 at 3220±35, SUERC-11609 at 3215±35 and SUERC-11602 at 3195±35 BP.

3195

-23.7

1495 to 1435 1530 to cal BC 1400 cal BC

Humic acid fraction of basal peat sample Call80/54/244 overlying soil in Area C, collected in May 1980. The error of GU-1403 ±65 quoted by the laboratory has to be increased to ±110, on the advice of the laboratory (Stenhouse M pers comm).

2640

110

-29.3

930 to 550 cal BC

1050 to 400 cal BC

Radio carbon ating \ 963

Calanais Survey and Excavation, 1979-88

24. Discussion and Conclusions 24.1 Introduction The excavations reported here produced a lot of new information and raised fresh questions. A few existing uncertainties were resolved. For instance, it was confirmed that the chambered cairn was later than the Ring; the Ring was probably set up in the first quarter of the 4th millennium and the cairn was built in the second half of the 3rd because it has a (typologically early) Beaker sherd incorporated in it and probable Beaker sherds under it. The original form of the chamber was different from that recorded after peat was cleared from the Ring in AD 1857. The central area of the setting probably ceased to be used for ritual deposition not long after 1800 BC, judging by evidence from pollen. Other questions, like the chronological relationship of the Central Monolith to the Ring, and the Avenue and Rows to the central constructions, received only partial answers. Erection of the stone at the west end of the East Row probably dated to somewhere between 2560 and 1690 cal BC; the early part of this range is preferred and the stone may have been put up before the chambered cairn was built. Stone 35a to the southeast of the Ring was probably not part of the pre-peat setting and thus the idea of a southern avenue received no support. We did not prove whether the Ring was built piecemeal or largely in a single operation. That said, the balance of probability is that the Ring was a single short-period construction apart possibly from one stone (Ring stone 52). But the discovery of two pits interpreted as having held Avenue stones close to one another on the line of the west side of the Avenue suggested that it had a more complex history than supposed a few decades ago. 24.1.1 Ancient meanings and modern ideas It is difficult to assess whether what archaeologists notice and record about prehistoric structures is what was important to those who built them. It is nearly as tricky to gauge what underlies mod-

ern ideas about ancient meanings. During the last decade and a half there have been several attempts to understand why a considerable effort was spent on creating large monuments in Britain during the 4th and 3rd millennia BC. Richards has suggested that banks and ditches embodied a microcosm of surrounding landscapes and more abstractly Brophy has followed others in proposing that the re-ordering of a set of materials into a relationship with the landscape gave the building of large structures a special meaning (Richards 1996, 331; see also Richards 1993; Brophy 2000, 68; Thomas 2007, 262-3). I must admit to some unease with these and similar ideas even though on the whole they are as good as many of the notions adduced in the 1970s and 1980s to explain why monuments were built. They have properly been developed from ethnographic studies showing that many non-western societies meld geography and culture (for a summary see Ruggles 1999, 120 and associated notes). Yet looked at from a different perspective they are reminiscent of the 18th and 19th century speculations of Stukeley, Toland, Pinkerton and others. Where those authors used the monuments to make religious or political points related to the controversies of their times, many modern interpretations seem to relate to current debates about the environment; the subtext is that people in the deep past linked their environment and their culture intimately, providing precedents for us to do the same. Even if the interpretations of many modern authors are influenced more by their own social environment than by those of the monument-builders, that does not mean that their arguments are unfounded or unappealing. But the intentions behind building will probably not have been related solely to a desire to create a link between the natural and social worlds. Nor need the aims of all of the individuals involved have been identical. Monument-building may have played a subtle or an overt part in establishing or maintaining social rank, social alliances and interpersonal dominance. Another reason for their construction may have Discussion and conclusions \ 964

Calanais Survey and Excavation, 1979-88

been to provide a focus for group identity. Yet another may have been to establish rights to the land or at least to its resources. And there are many other motives which might have played some part, including religion, a desire to realise cosmologies, bragging, or conspicuous consumption. Or they may have been built to foster a sense of belonging in an area or to a social group. Other ideas tentatively ascribed to ancient people show less relationship to modern fashions. Bradley (2000) thought that the creation of the Balnuaran of Clava cairns near Inverness in the centuries around or soon after 2000 BC might have been a reaction to a lack of pre-existing visible ancestral monuments. He suggested that people dealt with this by making analogues to the by then ancient cairns further north, some of which were re-used in this period (Bradley 2000, 230-1). Where meaning could not be laid on earlier structures it was always possible to create new ‘ancestral places’. This is an ingenious and appealing idea, but it poses the question of what people were doing around Inverness before 2000 BC, and why they changed only then. Had they previously rejected or ignored the idea of imbuing landscapes with cultural meanings, or had they been content to use natural features or structures which are invisible in the current archaeological record for those purposes? It is obviously wrong, however, to deny ideas to our ancestors because they are similar to modern ones. Bradley suggested other aspects of stone circles which do seem to reflect both modern and ancient concepts. Writing of the Recumbent Stone Circles of Aberdeenshire he proposed that they were architectural devices embodying particular beliefs. The grading of the stones might refer to the importance of the south-western sky, and the use of red and white stones might even have stood for the moon and the bonfires lit within the rings. More generally, architectural traditions were statements about the concerns of particular communities. They also provided the physical framework within which those ideas could be expressed in public ceremony (Bradley 2005, 114). Similarly Thomas, in discussing Holm and Holywood timber alignments in south-west Scotland, which seem to have been burnt down and replaced more than once, emphasised the likely

importance of both creation and destruction of monuments along with public acts associated with them (Thomas 2007, 262-3). Jones and Thomas (2010) have suggested that some small chambered cairns were rebuilt or built anew as a result of cultural stimuli imported by the people who introduced Beakers along with other exotic objects after about 2500 BC. The Clava cairns may be one aspect of this. But an example particularly relevant to Calanais is the southern chamber of the by then ancient chambered cairn at Embo in east coast Sutherland, considered below in the discussion of the Calanais chambered cairn. Richard Bradley’s and Julian Thomas’s ideas are speculative. So in a somewhat different way are those of Jones & Thomas. The latter in my judgement currently have too high a proportion of speculation to pertinent fact, and although they cannot be dismissed they should be treated with caution. But the ideas put forward by Bradley and by Julian Thomas strike a good balance between descriptions of what was found during excavation in the context of local and distant comparanda with an attempt to understand how the monuments were used within a local social context. It is a balance between ancient facts and modern ideas which I hope to emulate in discussing Calanais. For instance, in considering the landscape around Calanais, I shall discuss whether the building of stone rings and rows may have been the result of a ritualising or codifying of previously more fluid beliefs. 24.1.2 Comparing the structures at Calanais with those elsewhere In seeking comparanda for Calanais the issue of distinguishing related developments from coincidentally similar ones arises (Tilley 1998). To what extent can findings at a distant monument be used to increase understanding of why Calanais was built? Were Britain and Ireland mosaics of areas where people felt such strong local identities that social connections between those areas were severely impeded? Did similar-looking structures and other artefacts in different areas have very different meanings to the locals? Ian Kinnes suggested 27 Discussion and conclusions \ 965

Calanais Survey and Excavation, 1979-88

years ago that the â&#x20AC;&#x2DC;Neolithicâ&#x20AC;&#x2122; in Scotland should be treated as diverse only within an overall common tradition (Kinnes 1985). Quite what that means in practise is hard to define. Suppose (leaving aside the factor of earlier forager populations) we could demonstrate that around 4200 to 3800 cal BC there were two different networks of contacts between groups of people for whom farming was an important part of subsistence, one up the west coast and the other up the east Sheridan (2009) has suggested there may have been even more. Then the common ancestor of the traditions of east and west would probably lie further back in prehistory than if there were a single network, and the common elements in the various regional traditions in Scotland would be weaker, perhaps considerably so. There is another problem in describing what those overall traditions might have been; the degree of similarity between different sets of structures in different parts of Scotland may have been exaggerated. Barclay, in particular, has argued that there is a considerable bias in our understanding of the period in Scotland because descriptions of sites in Highland Scotland have been applied to those in the rich agricultural lands (Barclay 1992, 3-14; 1996, 61; Barclay and Maxwell 1998, 1; Barclay 2000). His underlying proposition was that throughout 4th millennium BC Scotland the lowlands were overall richer than the highland and island areas where stone monuments survive as prominent features. Lowland monuments should not be interpreted mainly through a highland lens. Yet another complication is that it would be wrong to regard lowland and highland Scotland as coherent units; both were patchworks of local environments and there are for instance some apparently favoured areas in the Highlands, such as that round Loch of Yarrows in Caithness (Davidson & Henshall 1991), and unfavoured areas in the lowlands, such as the stiff sandy clays east of Haddington in East Lothian (Lelong and MacGregor 2007, 6). Recent work in the Kilmartin Valley has shown that it contained timber-built features of types generally regarded as more typical of the lowlands (Ellis C 2000, 16; 2002, 145; Ellis and Crone pers. comm.; Cook et al 2010). There is still another difficulty in assessing how similar or different lowland and highland com-

munities were; it is not clear whether, in general, stone-built cairns, rings and linear settings should be seen as equivalent to timber and earth ones. It has been argued instead that stone was thought more suitable for monuments associated with ancestors and timber for monuments used by the living (Parker-Pearson & Ramilsonina 1998). But Thomas (2007, 262-3) citing ideas from Tilley (1996) has suggested that in the 4th and 3rd millennia BC stone may not have been regarded as an inert substance; the differences we perceive between stone and timber monuments may be greater than those recognised by their builders. So use of stone or timber may after all mainly reflect the ease with which they could be obtained. Or, given that several of the large early timber structures so far excavated in Scotland seem to have been purposefully burnt down, stone structures may have been a consequence of a change in beliefs from a concentration on the acts of creation and destruction towards building structures for the descendants, perhaps reflecting a change towards a greater sense of ownership of territories. 24.1.3 Testing ideas about regionality It is important to be clear what problem (Popper 1976, 132-5) we wish to solve here by exploring evidence for regionality. That problem lies in judging whether comparisons between structures at Calanais and apparently similar structures in other regions are valid. What criteria do we have for judging that? The best criterion is probably the quantity of (apparent) similarities found between the material cultures of two areas, the degree to which different strands of evidence match up. There are immediately two problems. The first is the circularity involved in using apparent similarities and differences to judge whether those similarities and differences are valid. The second is that the database of well excavated sites, of distinctive artefacts and ecofacts and of highly distinctive unexcavated sites is too small to attach satisfactory measures of probability to statistical analyses. Nevertheless, at this stage of enquiry, and accepting that errors will take a lot more data to iron out, the best approach is to allow the quantity of Discussion and conclusions \ 966

Calanais Survey and Excavation, 1979-88

similarities and differences to serve as a proxy for the degree to which communities at various distances from each other exchanged people, goods or ideas. If there was a lot of contact at a given time (as measured by the finding of similar structures, artefacts and ecofacts in them) then comparisons between the Calanais area and other places are more likely to be valid than if there were very few dealings. We can start by noting that it has long been accepted by most of those studying Scottish prehistory that there were regional variations at many different scales in Scotland during the ‘Neolithic’ (Scott 1951; Piggott 1954; Henshall 1963; Megaw and Simpson 1979; Kinnes 1985; Armit & Finlayson 1992; Sharples 1992, Barclay 1992, 1997, 2000, 2003; Ashmore 1996, 2003; Brophy 2006; Bishop et al 2009). Theoretical approaches to defining that regionality at least in neighbouring countries have taken many forms. For example, J G D Clarke framed his 1975 study of early prehistoric Scandinavia within four different types of territory: sites, the area covered by a group in its annual movements, larger socially linked areas and areas within which people had similar technologies (Ballin 1999, 101; Clarke 1975). Madden (1983) explored three different models in her study of social networks in Southern Norway between about 8500 and 2000 BC. She looked at undifferentiated networks, networks differentiated by distance and networks differentiated by social boundaries (Ballin 1999, 102; Madden 1983, 191-200). But pragmatically the interpretations of the nature and extent of the social phenomena which artefact distributions represent often cannot be formally tested by the data. For instance the scales of the geographical distributions of stone axes vary considerably in Scotland (Ritchie and Scott 1988, 87), as they do in England and Wales (Cummins 1979, 8-11). That cannot be explained completely by fieldwork biases, by differences in the quality of the stone and by geography; complex social factors no doubt played a role. The latter are suggested, for instance, by the different natures of the distributions of axe-hammers made of picrite and those found in the same area but made from other stones, the one falling off with distance from source and the other evenly spread over the same

area ((Hodder and Orton 1976, 107). Less theoretically based studies with a high descriptive component can be illuminating. The evidence provided by plant remains found on Scottish sites dating from the first signs of farming to the middle of the 3rd millennium BC suggests that people in different parts of Scotland used different proportions of domestic and wild plant foods (Bishop et al 2009, 56). Food production is an activity which must have affected the time available for others and the experiences undergone during different kinds of food procurement will have been dissimilar to one another. Major differences in food procurement strategies will often have spilled over into other cultural practices. For instance cereal production allowed longer term food storage - although that does not mean that people always took advantage of that as well as they might - and its concentration in fewer hands; although see Stevens & Fuller 2012 for an argument that cereal storage techniques may not have been adequate in parts of the British Isles during the 4th and 3rd millennia. Those and other approaches called into question simple ideas of regional differences and showed that it is unclear whether artefacts and structures indicate a significant proportion of the cultural differences between different areas. David Clarke has argued that pragmatically prehistorians should develop a set of narratives based on differences rather than similarities: firstly narratives about individual sites such as Skara Brae, secondly small areas rich in artefacts such as Culbin Sands, Glenluce Sands and Little Ferry with their many flint arrowheads, contrasting them with superficially similar areas with far fewer artefacts, and lastly regional and national narratives (Clarke 2004. 46-48). None of these, he suggested, should have priority over the others. Underlying this was his belief that it is essential to study the material from many angles before embarking on grand theories. Brophy’s review in 2006 of Scotland’s ‘Neolithic’, while tending to focus on structures rather than artefacts, showed the advances in understanding since Ian Kinnes reviewed it 21 years earlier (Brophy 2006). Brophy concluded that rough regional traditions can be detected, that regionality is best defined by the everyday and mundane and Discussion and conclusions \ 967

Calanais Survey and Excavation, 1979-88

that the concept of regionality works both at the level of local and wide scale patterns. That said, he did not think that regions should be assigned fixed boundaries. He thought that networks involving the movements of ideas, objects and people, and connection-sets of exchange and obligation, were the mechanisms which drove ‘Neolithic’ society (Brophy 2006, 38). As should be apparent from some of my own publications I am broadly in accord with this position (Ashmore 1996, 2003). Thus regionality need not imply a neat little-changing partition of Scotland into inward looking areas. For that matter much the same may be true of (chest of drawer) chronological divisions. Barnett might have been right in supposing that architectural variations between small ritual/ceremonial/funerary monuments could mask a countrywide and multi-period similarity in the ideas which gave rise to them (Barnett 1998, 8-13). That said, a recent study of the late prehistoric period (1500 BC onward) in the Eastern Netherlands using a richer database of excavated sites and artefacts than is available for Scotland has demonstrated the need for regionally specific models (Beek 2011, 25, 43-5). Its main conclusion was, much as Barclay argued for Scotland, that to lean too heavily on research results from one region to explain those of another carries a risk that the specific characteristics of the latter will be overlooked. Beek also argued that some important changes occurred almost simultaneously across large parts of north-west Europe, most likely reflecting radical ideological changes (Beek 2011, 45). Until many such studies have been completed the practical answer to the problem of assessing whether comparisons between structures in the Calanais area and those in other parts of Scotland is that one should be very wary. But despite the existence of very real regional differences, too much emphasis on restricting comparisons to local phenomena, ignoring links to distant places, may force one to seek local explanations for change when its true motor was change elsewhere. Even if ‘local’ or ‘regional’ approaches could be argued to have produced the major successes in Neolithic studies there is much to be said for the leavening provided by broader studies (Thomas 1998, 37-8).

24.1.4 Regionality and Connexions Taking ‘success’ here in the limited sense of creation and retention of a rich material culture, the success of people at Calanais in the 3rd and early 2nd millennia BC was probably dependent on the existence of low-cost two-way links between productive groups of people at geographically various locations . Judging by its architectural features and the artefacts found there the people of Calanais seem to have been well-connected to distant places. Conversely failure to maintain a rich material culture at Calanais during the rest of the second millennium BC probably reflected loss of those links. Paradoxically the pollen evidence suggests more local success in cereal cultivation after the loss of cheap long-distance links. Maybe changes in ideas amongst the other peoples of the lands bordering the Irish Sea, rather than local transformations, led to impoverishment of the Calanais people's material culture. Perhaps people in this wider area stopped thinking it worth their while to travel long distances to visit religious centres, for instance. Severing even one or two long distance links might drastically change connectivity; only local links might survive (Appendix 8). Bradley has described what may be a Scottish example of such behaviour. He saw similarities between the Clava Cairns and some western Scottish and Irish sites, and suggested that the area of the Clava cairns was in a pivotal position in relation to (land-based or coast hugging) communications in the north (Bradley 2000, 22830). He drew an analogy with the same area in the early medieval period, noting the proximity of the important fortified sites at Burghead and Kineddar and the abundance of stray finds from the nearby Culbin Sands. He proposed that then and also in the late 3rd and early 2nd millennia BC Culbin Sands was a place where specialised artefacts were made and exchanged. In the terms of network theory the activities at Culbin Sands formed a node which both attracted many twoway low cost-benefit connections fostering long distance links. Bradley pointed out that while the Culbin Sands contain abundant late 3rd and early 2nd Discussion and conclusions \ 968

Calanais Survey and Excavation, 1979-88

millennium material along with similarly rich sets of medieval artefacts, little survives from other prehistoric periods. Changes in the environment may have been a factor but perhaps in both cases the collapse of the activities which produced the finds was accelerated by the innate characteristics shared by all complex networks. Culbin may itself have ‘failed’ or some other place crucial to Culbin’s ‘success’ may have ceased to provide linkage. Whatever the reason for Culbin’s success and failure the growth or decline of its links with distant places may have caused (or been caused by) changes to other parts of the social network in Scotland, possibly including those related to Calanais. So even though regionality was a prominent feature of the 4th and 3rd millennia in Scotland, it is dangerous to interpret local phenomena in isolation from wider changes; distance alone should not be treated as a criterion for rejecting distant comparanda. Despite the dangers of assuming connections when none existed (Tilley 1998) it is quite as dangerous to assume that changes at Calanais took place independent of those in other regions. 24.1.5 Travelling The amount of long term connectivity between Calanais and other areas must have been linked to the real and the perceived cost of travel. It seems unlikely that much non-local travel was by land. Accounts of how quickly people could move along rough tracks carrying a load vary, but in the 18th century a packman with a full load could be expected to travel about 10 miles (16km) a day. There were also limits on how far people could travel without procuring new supplies of food. In the early 20th century Chinese packmen carrying food for famine relief could not travel more than 75 miles (120 km) without consuming all that they carried (Scott 1951, 23). This sort of statistic is important particularly when considering whether a route including Loch Broom and the modern-day A835, or the Great Glen, might have seen frequent use for travelling between the Western Isles and north-east Scotland. It is roughly 50 miles (80 km) along the route

of the A835 from the head of Loch Broom to Inverness; it might have taken 2 to 4 days to walk, depending on conditions. The modern Great Glen Way walkers route is 79 miles (127 kms) long between Fort William and Inverness (www.greatglenway.com) and can fairly easily be traversed (without a heavy load) in 5 to 6 days. Paddling a canoe along the Caledonian Canal including Loch Lochy, Loch Oich, Loch Ness and Loch Dochfour normally takes at least 3 days (www. scottishcanals.co.uk). Some portages will have been required in prehistory, increasing the effort and time for travelling. By way of contrast, travel by sea may have been relatively low-cost. Nobody knows what sorts of boats foragers and subsequently farmers used in Scotland in the 4th millennium Log boats are a possibility and there are more than 3500 European examples of which around 550 are known to be prehistoric. The vast majority come from inland sites. The earliest known is a slightly contentious small log boat from Pesse in the Netherlands, dated to between 8250 and 7750 cal BC (GrN6257 8825+/-100 BP) although they seem to have become common only towards the mid 4th millennium BC (McGrail 1987, 86). The log boat from Brookend township, Lough Neagh, Ireland, the closest known 6th millennium log boat to Calanais, has been dated to between 5490 and 5246 cal BC. The earliest radiocarbon-dated Scottish log boat is from Catherinefield, Locharbriggs, Dumfries and very vaguely dated to between 2900 and 1600 cal BC; SRR-326 3754 +/- 175, corrected to +/- 175 to acknowledge larger than quoted errors). Plank-built boats seem to be a phenomenon of the mid 2nd millennium BC and so far have been found only in the east and south of Britain (Megaw and Simpson 1979, 295; Champion 1999, 104-5). Skin boats are a strong possibility given the use of curraghs in historically recorded periods. Their absence from the prehistoric Scottish record probably reflects only the fact that exceptional conditions are required for survival of recognisable remains. Historically recorded curraghs had wickerwork sides and ribs covered with cow-hide. They could be 7 or 8m long and up to 2m wide. They could Discussion and conclusions \ 969

Calanais Survey and Excavation, 1979-88

carry 9-10 people, and Irish seamen in the 1970s regarded two cows and 21 sheep as a good load to take to sea. Their sailing versions had long lifting bows and could probably accomplish as much as 145 km a day (90 miles; Bowen 1972, 36-7). On Severinâ&#x20AC;&#x2122;s Brendan Voyage 10 oarsmen could make an average of 74 km a day (McGrail 1987, 184). According to Fridjof Nansen (1893) a fully laden

umiak (skin boat) crewed by women with their children and enough equipment to make them self-sufficient could routinely travel 50 English miles (80 km) a day. Given the lack of 4th millennium and earlier evidence from Scotland, it is impossible to know how advanced boat-building technology was. Boats may not have had sails, or if they did the

Illus 24.1Â Travelling to and from Calanais Discussion and conclusions \ 970

Calanais Survey and Excavation, 1979-88

sails may have been markedly less efficient than those of 20th century curraghs. However, if historically recorded paddled or oared skin boats were routinely used to travel 75 to 80 km a day it seems reasonable to suppose that similar boats dating to before 2000 BC could routinely be used to travel at very least 50 km a day. As the examples quoted above suggest, much shorter journey times than this ‘routine’ figure would have been possible on direct long distance journeys (see also Garrow and Sturt 2011, 62, quoting Callaghan and Scarre 2009). 24.1.6 Travelling to and from Calanais Illustration 24.1 is an attempt to provide a basis for assessing how many days it usually took to get from Calanais to other places by sea. The model proposed here is that most journeys involved several short legs, an idea recently discussed afresh by Garrow and Sturt (2011, 62). Therefore the conservative figure of 50km a day has been used. Admittedly, winds and currents will have meant that the time taken had only a loose relationship with distance. Further, the seaways between northeast Ireland and south-west Scotland are particularly perilous because the tidal streams running on either side of Ireland meet there and particularly at spring tides dangerous races could extend well offshore. The Kyles of Skye also present hazardous conditions (Scott 1951, 32, quoting the West Coast of Scotland Pilot 6th edition). Scott considered that the main prehistoric route from southern areas past the Western Isles ran west of Skye to the Uists and thereafter up the Minch, to the east of the islands (ibid 33). Indeed, early modern settlements favour the east rather than the west coast of Harris although in South Lewis the east coast has little settlement. Similarly the west coast of South Lewis, north of Scarp is fairly bleak for 20 km or so. But on the west coast, once the headland at Ard More Mangersta has been navigated the area round Loch Roag supports many townships. Even north of Calanais where the west coast can hardly be described as gentle there are sandy bays such as Dalmore and Barabhas. Given the changes in the coastline described below finding safe havens along a western route

may have presented different difficulties to those of today. But the differences between the western seaways of c. 6000 to 3000 BC and those of today may have been much slighter than those on the east coast of Britain (Garrow and Sturt 2011, 63-65). It seems possible that, in summer at least, the sea-route along the west coast of Lewis might sometimes have been preferred. Burl (2000, 94) suggested that the central monolith at Calanais could have served as a landmark for seamen. His idea that the group of stone rings at Calanais and that at Machrie Moor on Arran ‘may have resulted from the arrival of crews from southern Scotland, Ireland and England taking shelter on the long voyage to the Orkneys.’ has gained additional force from the discovery of a grooved ware pot at Calanais remarkably similar to one at Stones of Stenness in Orkney. On Illus 24.1 a secondary centre in Orkney has been added to allow a better assessment of the time it would take to get to eastern Scotland by sea from Calanais. Any one of the havens on the Caithness coast might have provided a more economical stopover on a long voyage shortening journeys by a day or so; Orkney has been chosen because of demonstrable links between it and Calanais in the early 3rd millennium BC. Using the conservative figure of 50km a day travel times in fair conditions would have been as follows. All of the landing points on the west coast of Lewis and Harris were within a day’s travel. —— All of the landing points in the Western Isles north of Benbecula, North Skye and most parts of the facing mainland were less than two days away from Calanais. —— The rest of Skye and the west coast mainland north of it could easily be reached in three days. —— It would have taken about 5 days to get to Orkney, the south end of the Great Glen or most of the southern Hebrides, which were roughly the same distance away as each other. —— An 8 day voyage could get people to Shetland, Antrim or the Rhinns of Galloway, and via Orkney to the south coast of the Moray Firth. —— A 10 day voyage would get people to the Isle of Man and nearby parts of Ireland and England, and via Orkney most of the coast of Aberdeenshire. Discussion and conclusions \ 971

Calanais Survey and Excavation, 1979-88

—— A 12 day voyage would get them to the Forth, and to north Wales or the east coast of central Ireland. But it would be too simple to conclude that connection would always have been fairly easy. On top of the perennial hazards of sea travel, mutual hostility may have prevented people from passing close by some communities, or making overnight stop on some parts of the coast. The 50km a day routine voyaging rate used here is much less than the ‘long distance’ rate used by Callaghan and Scarre (2009); they suggested that a voyage between Brittany and Orkney might have taken only 12 days (Garrow and Sturt 2011, 62). Urgent journeys, or non-stop journeys designed to avoid trouble with other communities, may have been completed in half the times listed above. A sea and land route including Loch Broom would have included one or two days sea travel and around 4 days overland, a total of 5 or 6 days; it would have required a safe boat-storage spot near the head of Loch Broom. The Great Glen route might have taken a leisurely 5 days by sea and 3 or 4 along the Great Glen waterways, 8 or 9 days in all. Even if done at haste the trip would have taken 4 or 5 days. My impression from the figures discussed above is that links beyond the Western Isles and Skye, even by sea, cannot be described as low-cost in a modern every-day sense. However, time may have been a less important factor in judging cost than it is for industrial and post-industrial societies today; cost is measured against aspirations; religious or acquisitive motives could have made an extended effort seem good value. As suggested above, one reason why network links failed after the early 2nd millennium may have been that belief systems changed.

24. 2.1 Warming During the last glaciation the Western Isles were probably largely or completely covered by ice (Lambeck 1995a, 85; Lambeck 1995b Figs 3a, 3b; Gilbertson et al 1996, 10 Fig 2.4). At some date around 17,000 cal BC ice may still have covered the Minch, but the long island itself may have been largely ice-free, with a western coast line considerably to the west of the present day one (16000 BP Lambeck 1995b, Fig 3c). But the relationship between land and sea levels will have been complicated because sea rise during the main ice-melting will have been opposed by land rebound for mainland Scotland and the Inner Hebrides (Ballantyne 2004, 35).

24.2 Before Calanais was built: the end of the last glacial period in the Western Isles

Illus 24.2 The Great Western Island at some date between 12,500 and 11,000 cal BC (after Lambeck 1995a, Figure 2d and Gilbertson et al 1996)

It seems proper to provide a sketch of the major climatic and geographical changes in the area around Calanais after the peak of the last glaciation, for their effects rumbled on throughout the period in which Calanais was built and used.

By about 12,000 cal BC the mainland of Scotland was also ice-free (Bell and Walker 1992; Lambeck 1995b, Fig 3e). On Illus 24.2 the present Western Isles are laid against a background of land approximately as it was at some date in the warm Discussion and conclusions \ 972

Calanais Survey and Excavation, 1979-88

period between about 12,500 BC and 11,000 cal BC when a human presence is first attested in Scotland (Ballin et al 2010, 344, 357). The modern submarine contours are shown at 20m, 50m and 100m. I shall refer to this as the Great Western Island. 12.2.2 A cold snap Another very cold period (the Younger Dryas) started around 11,000 and ended about 9600 BC; an ice sheet reappeared in Scotland during this period and land and sea falls and rises kept pace with one another along much of the western seaboard of Scotland; the Main Late-glacial Shoreline formed. After the end of the Younger Dryas earthquakes of magnitude 6.5 to 7.0 may have been fairly common as the land adjusted to removal of the weight of ice above it, persisting with a gradual reduction in intensity until the first millennium BC (Ballantyne 2004, 34), a factor not often considered in studies of the possible mindsets and beliefs of those living then in western Scotland. 24.2.3 New warming and new vegetation By about 9250 cal BC ice-fields and permafrost had disappeared from Scotland, except in mountainous areas, and the land bore tundra vegetation. Temperatures regained their present levels around 8500 cal BC. By c. 7500 cal BC Lewis

was not grossly different in shape from what it is now (Lambeck 1995 Figs 3g, 3h) although some areas which are now shallow sea were still dry land. Absolute sea level rise culminated in a period of very rapid world-wide sea rise around 6000 cal BC (7000 BP; Buchanan 2004, 35). Around this time Britain was severed from the continent (ibid 35). Relative sea rise along the coasts of the Great Western Island continued locally for several millennia after 6000 BC judging by the inundation of field walls at Leobag near Calanais (Illus 24.3; Bohncke 1988; Cowie forthcoming). More generally in the Western Isles there are many archaeological sites between low and high tide mark (Ritchie 1979). After a date somewhere between 6000 and 5500 BC birch and willow seem to have varied in a complementary fashion for about 500 years. Grass and heather became more abundant although the scarcity of some other species suggests that there was little increase in open ground. At some date between 5375 and 4825 cal BC, birch suffered a set-back and willow almost disappeared; woodland around Calanais Leobag. was reduced. Hazel increased modestly. No other tree type expanded significantly to take advantage of the ground surrendered by birch and willow. Dry grassland declined and heath expanded greatly (Bohncke 1988). After a wetter phase, the recovery of Betula from a date between 4700 and 4150 cal BC coincides with an increase in grasses and bracken, a fall in sedges and rushes and the local absence of

Illus 24.3Â Modern sea depths around Calanais, illustrating the effects of lower sea levels Discussion and conclusions \ 973

Calanais Survey and Excavation, 1979-88

Sphagnum as the ground became drier (Chapter 21 Palaeoenvironment 21.3.4). During this period the woodland structure may have had no exact modern equivalent (Tipping 2004, 46-48) 24.2.4 Early peoples in the Western Isles The nearest place with evidence for human activity before the cold millennium and a half of the Younger Dryas, the very cold period between about 11,000 and 9600 cal BC, is at Kilmelfort, Argyll (Saville & Ballin 2009, 37). From a slightly earlier period, stone tools found at Howburn Farm, Elsrick near Biggar in southern Scotland have Later Hamburgian affinities probably dating to around 12,000 cal BC (c. 12,000 BP) (Ballin et al 2010, 344, 357). The initial extremely cold period of the Younger Dryas may have driven even the hardiest prey animals and descendants of the Later Hamburgian hunters from Scotland. But the occasional discovery of flints of Ahrensburgian style, such as the small tanged point from Shieldaig in Wester Ross, suggests a human presence not far from the Great Western Island during the main part of the Younger Dryas. Perhaps they hunted reindeer, like those using similar tanged points on the continent (Ballin & Saville 2003; Edwards and Mithen 1995). It is not beyond credence that their successors maintained a breeding population in N Britain (Technical Note 24.2.4) although Woodman (2012, 6) has recently emphasised how brutally cold the Younger Dryas period was, expressing doubt about the survival of large mammals in Ireland, and has cast doubt on claims for Ahrensburgian points in Scotland. But marine resources such as seals would still have been available and, conceivably, some of the ancestors of some of those who built Calanais could have been moving around the area at least 7000 years before the Ring was built. There is not enough archaeological information to allow discrimination between the cultures of forager communities within Scotland, although differences can be seen between it, Ireland, and England (Saville 2003). Current evidence for settlement in Eastern Scotland is considerably richer and up to a millennium earlier than evidence for settlement in Western Scotland (Saville et al 2012,

81). That difference may however reflect only the small number of sites excavated. Settlement of hunter-gatherers at Kinloch on Rum, which is only a short sea voyage away, began at latest about 7500 cal BC (Wickham-Jones 1990). A bevelled tool from Druimvargie rockshelter, Oban, is dated to c 7580â&#x20AC;&#x201C;7180 cal BC (7890 Âą 80 bp; Bonsall et al 1995). Pollen studies by Edwards (1990), Bennett et al (1990) and others document evidence for what seems to be human activity in the long island at dates before 7000 cal BC. Bohncke (1988, 450) suggested human activity around that time at Calanais Leobag but as discussed below reconsideration of the radiocarbon dates available to him suggests a slightly later date for substantial evidence of woodland clearance and charcoal creation (see also Chapter 21: Palaeoenvironment 21.2.1). At Northton, Harris, two individual hazel nut shells in a midden have been dated to between 7060 and 6690 and between 7040 and 6650 cal BC (AA-50336; AA-50335; Gregory et al 2005, 945; Simpson et al 2006, 23). It could be argued that the charred nut shells were the result of a natural fire incorporated in a later midden almost by chance. They were not however surrounded by an abundance of charred hazel wood. That reduces the likelihood of their originating in a natural fire and favours the explanation that they were a result of nearby human activities related to hazel-nut processing, an activity well attested elsewhere in Scotland on sites of much the same period (e.g. Mithen 2000, 435-6). A recent report of an eroding early land surface at Northton with hearth deposits and food waste (Bishop et al 2010, 178) broadly fits the interpretation of food-related charring. Three bevel-ended tools from a midden at Sand, Lochalsh have been dated to between 7050 and 6500 and between 7050 and 6450 cal BC (OxA10384 7855+/-60; OxA-10175 7825+-55 BP) while several other tools and pieces of charcoal dated from then until some date between 5540 and 5320 cal BC (Ashmore and Wickham-Jones 2010 Table 173). A bevel-ended tool from a midden at Loch A Sguirr, Raasay has been dated to between 6220 and 6000 cal BC (OxA-9255 7245+/-55) and a single piece of birch charcoal from the midden to between 6640 and 6250 cal Discussion and conclusions \ 974

Calanais Survey and Excavation, 1979-88

24.3 The sequence of activities at and near Calanais from c. 5000 BC

Illus 24.4Â Activities at Calanais and zoning of pollen column CN3 at Calanais Leobag Discussion and conclusions \ 975

Calanais Survey and Excavation, 1979-88

BC (OxA-9305 7620+/-75) (Ashmore and Wickham-Jones 2010 Table 174). A bevelled tool from the rock shelter at An Corran, Skye has been dated to between 6610 and 6245 cal BC (OxA-4994) along with a cluster of three similar dates around 6500 to 6300 cal BC from aurochs bones (Saville et al 2012, 73-6). It is clear that foragers had long settled the shores of western Scotland and its islands by the time, between 6600 and 6200 cal BC, when woodland clearance first took place near Calanais. There is no reason to suppose that the people who burnt the woods there were permanent residents. They will have ranged widely over the seasons. On present evidence they were probably part of a set of communities exploiting the coastal resources further to the south and east. 24.3.1 The overall sequence Illus 24.4 has been constructed using radiocarbon dates from Calanais and Calanais Leobag (see Chapter 23 Radiocarbon and Chapter 21 Palaeoenvironment). The grey boundaries between the pollen zones named to the right are meant to con-

vey how imprecise the pollen zone dating was. The headings and diagrams at the top show on which of the areas A to H the phenomena in the diagram below occurred. Area S has been omitted. The excavation did not reveal conclusive evidence for human activities at the stone setting itself before about 3000 BC, although soils possibly indicating earlier cultivation were found in Area F (in the Avenue, Illus 24.4 top). Earlier activity in the general area was inferred from charcoal and pottery imported to Calanais and from pollen evidence at Calanais Leobag. In the 3rd and early 2nd millennia BC various structures were built; human remains were deposited, soil was imported and ground-working, including ploughing, took place several times (Illus 24.4). The area where the structures had been built was largely abandoned except for outfield grazing by the middle of the first millennium BC; mosses and other peat-land plants had started to cover the ground at some time between 920 and 400 cal BC. But there may have been some activity at the pre-peat ground level in the Ring around the middle of the 1st millennium BC.

24.3.2 Calanais Phasing Table 24.1Â Phasing of early features north of the Ring from about 6500 to about 3000 BC Phase Date BC Date source

Events

Dating and comments

Pollen

Woodland clearance

Started at some date between c. 6600 and c. 6200 cal BC. in CaN-2a and 2b at Calanais Leobag.

C14

Foragers in Avenue area?

Date between 5305 and 5215 cal BC

6500

0 6000 5500 1 5000 4500 4000

Discussion and conclusions \ 976

Calanais Survey and Excavation, 1979-88

Phase Date BC Date source 2a

C14

Events

Foragers or farmers in Avenue area?

Date between 3720 and 3530 cal BC

C14 and pollen

Farmers in Avenue area and near Calanais Leobag.?

Date between 3640 and 3380 cal BC. Soil disturbances. Early cereal cultivation started near Calanais Leobag at some date between 3875 and 3605 cal BC

Pollen

Cultivation finished

Date between 3490 and 3020 cal BC.

C14

Early farming in Avenue area?

C14

Charcoal dated to between 3330 and 2890 cal BC, in Original deposition of charcoal and early to middle round terms between c. 3300 and 3000 BC. Imported with Neolithic pottery in settlements nearby. soil to Calanais many generations later in Phase 7.

3500

Dating and comments

Charcoal date between 3300 and 2890 cal BC. Soil disturbances.

Table 24.2aÂ Phasing of features in and near the Ring between about 3000 and about 2250 BC Phase 4

Date BC 3000

Date source

Events/processes

Comments

Pollen

Early ditch

Before cultivation beds.

Pollen

Cultivation beds and subsequent flatter cultivation

At beginning of CaN-3a, start between 2983 and 2510 cal BC.

Pollen

Turf formation

Formed over partly flattened cultivation beds (including, judging by pollen samples, formation at various subsequent dates over silts in cultivation troughs). Continues after Monolith erection until Ring erection

Stratigraphy

Monolith erection?

Near beginning of CaN-3a. 3000/2850 BC After 5b turf growth, and interpreted (but not demonstrated) as before Ring.

2950 6a?

Discussion and conclusions \ 977

Calanais Survey and Excavation, 1979-88

Phase

Date BC

Date source

Events/processes

Stratigraphy

Monolith erection? Ring erection?

Stratigraphy, analogies

A piece of birch charcoal from a charcoal spread in pre-Ring soil 871 proRing erection? duced a radiocarbon age (AA-24959 Turf formation? 4140+/-45 BP) corresponding to a Use of Ring area? date between 2880 and 2580 cal BC. Interrupted slot? That suggests that the Ring was set up after 2880 cal BC.

Stratigraphy, pollen

Ring erection? Continued turf formation? Use of Ring area? Interrupted slot?

The new turf is that above green clay spoil from Ring pits but where not covered by clay the old turf continued to grow

C14

Interrupted slot? Other use of Ring?

TPQ some date between 2890 and 2620 cal BC from charcoal.

6c?

C14

Interrupted slot? Other use of Ring? Continued turf formation?

Artefact

Some deposition The Grooved Ware pot may have been outside the Ring? made considerably earlier

Stratigraphy

Clay platform and light timber struc- Possibly 2450/2350 BC? tures built.

2900

6a, 6b? 2850

6b?

2800 6b? 2750 6c? 2700 2650

2600 7a?

7b?

Speculation 2550

Comments

Clay and soil imported with Phase Possibly as early as this, or even part of 3 artefacts and a long term process starting earlier. charcoal.

Discussion and conclusions \ 978

Calanais Survey and Excavation, 1979-88

Phase

Date BC

Date source

Events/processes

Comments

Interpretation

Clay and soil imported with Phase 3 artefacts and charcoal.

Layers on Areas D and H, with Phase 3 charcoal and also potsherds. Preferred date shortly before clay platform built.

Indirect C14 and interpretation

Rows start now?

Artefacts

Closure of green clay platform structure? Dumping of clays 810 and 812 in BV?

In general undated but East Row Stone 30 probably erected at some date between 2560 and 1690 cal BC; earlier than green clays with very early Beakers.

2500 7b? 2450 2400 7c

7d 2350

Slot fills include probable fine and â&#x20AC;&#x2DC;domesticâ&#x20AC;&#x2122; Beaker sherds. Green clays 810 and 812 with very early Beakers and no obvious later ones. After 2500 cal BC because the primary cairn and immediately pre-cairn contexts included an International Beaker sherd and at least one fine Beaker sherd.

Artefacts

Chambered cairn building

Interpretation / stratigraphy

Enclosure Stage 1, stake-hole defined after disuse of Starts around 2400/ 2300 BC? timber structure and after cairn construction

Interpretation / stratigraphy

Embanked enclosure Stage 2 earthen bank building,?

Starts around 2400/ 2300 BC? Barbed and tanged arrowheads in entrance silts may reflect use or post-date end of formal use.

C14

Ground working and burials

Pine charcoal dated between 2490 and 2290 cal BC could belong here, or Pomoideae charcoal dated between 2120 and 1770 cal BC

2300 10a

Discussion and conclusions \ 979

Calanais Survey and Excavation, 1979-88

Table 24.2bÂ Phasing of features at Calanais from about 2250 BC Phase

Date

2250 BC

10c

Date source

Events/processes

Comments

Starts around 2400/2300 BC?

Ground working and burials

See 10a

10d

Avenue starts now?

No specific dating evidence. Perhaps building on from a North Row? Stones set up over several generations?

Ground working including ploughing

Birch charcoal dated to between 2030 and 1750 cal BC and Pomoideae charcoal dated between 2120 and 1770 cal BC, from an ard mark north of the enclosure, could belong here.

2000 BC 10e

C14

11a

11b

12a 12b

Artefacts; Indirect C14 and interpretation.

Wall-based enclosure Stage 3 building

Interpretation

Ground working including ploughing

Analogy

Wall-based enclosure Stage 4 building

Interpretation

1000 BC

Later period of plough soil 141 belongs in pollen zone CaN-3c which started between 1880 and 1520; and was earlier than 139, so dates to between 1880 and 1690 cal BC.

Undated; possibly in use at some date between 2150 and 1750 BC, judging by an analogy with Ardnave, Islay. Includes layer 139, by hypothesis, dated to between 1940 and 1690 cal BC by a piece of heather charcoal.

Interpretation

After 2040 and perhaps after 1525 Kerb erection and cal BC; may post-date ruin of fourth secondary cairn stage enclosure?

Stratigraphy

Cultivation

1500 BC 14

Ground working including cultivation

Between 2100 and 1750 BC? After deposition of barbed and tanged arrowheads; also a possible terminus post quem between 2040 and 1770 cal BC.

Undated except near Area C it cannot be later than steady local peat growth.

Discussion and conclusions \ 980

Calanais Survey and Excavation, 1979-88

Phase

Date

C14 and pollen 500 BC

15 BC/AD 15 AD 500 16

17a

Date source

Events/processes Cultivation and encroachment of peat

Comments

Peat starts to grow steadily near the east end of the East Row between 920 and 400 cal BC.

Depends on Phase Growth of peat 16 being real. except by Ring?

Continued use of area of the stone setting, dissuading local peat growth?

Depends on Phase Growth of peat 16 being real. except by Ring?

Continued use of area of the stone setting, dissuading local peat growth?

See comments AD 1000

Disturbance and erection of slab in passage? Growth of peat

Possibly in the second half of the 1st millennium AD judging by occurrence of late Iron Age pottery. But the activity itself may be post-1857.

Peat growth

AD 1500 Documentary

Cutting and post-medieval peat clearance

North end of setting cleared before mid 19th century. MacCulloch surves before 1819. Worsaae visits 1846.

Documentary

1857 clearance; 19th excavations and disturbances

Pit digging at earliest in AD 1846 but probably after 1857

Documentary

Modern conservation

From 1882 to 1980

18c

Documentary

Archaeological excavation

18d

Documentary

Post-excavation conservation

Resisitivity survey 1979, excavation 1980 and 1981; intrusive sampling etc 1982; excavation 1986.

17b AD 1850 18a AD 1880 18b AD 1980

Improvement of drainage and returfing and path-building.

Discussion and conclusions \ 981

Calanais Survey and Excavation, 1979-88

24.4 Phases 0 to 3: Early activities at and near Calanais 24.4.1 Phase 0: Burning and clearing of woodland near Calanais There are no signs in the pollen record from Calanais Leobag of a human presence before some date between about 6675 and 6045 cal BC (the end of Pollen Zone CaN-1). Before that birch trees seem to have been abundant along with some willow and limited areas of grassy heather-rich moorland. The environment was probably fairly harsh given the overall lack of diversity in plant species. Sphagnum spores were abundant at Leobag, reflecting the damp conditions which led to formation of the peat (Chapter 21: Palaeoenvironment 21.3.1). The abundance of trees does not preclude the presence of coastal foragers, for the shore was further from Calanais Leobag than it is today, but it does not indicate any human impact on coastal woodlands. There was an increase in charcoal in peat column CN3, at the 55 to 54 cm level, in Bohnckeâ&#x20AC;&#x2122;s Pollen Zone CaN-2a, so foragers may have been burning woodland near Leobag, about 600 metres SE of Calanais, at some date in the second half of the 7th millennium BC Microscopic charcoal continued to appear in the peat column thereafter. At 52 cm (dating between 6160 and 5770 cal BC) in pollen zone CaN-2a the percentages of birch pollen dropped suddenly; willow flourished. That may reflect use by foragers or a change in local conditions. Towards the end of this pollen subzone (between 6015 and 5635 cal BC) the grass, heather, herb, dwarf shrub and fern pollen percentages increased firmly although that may have been due more to decreased density of the canopy cover than to a true increase in open ground. That said an increase in potentilla, if really tormentil, suggests the establishing of grassland. Sub-zone CaN-2b lasted from some date between 6075 and 5525 to some date between 4120 and 3830 cal BC. Birch and willow pollen fluctuated inversely to one another. There was little increase in open ground. But at a date between 5375 and 4825 cal BC, birch was adversely affected and willow almost disappeared. Towards the end of the

sub-zone at some time between 4120 and 3830 cal BC heather increased markedly, suggesting dry heath. Birch recovered while remaining subject to fluctuations. Alder began to form a continuous curve in the pollen record. 24.4.2 Phase 1: Sporadic visitors to Calanais The pollen evidence from sub-zone 2b provides a general background for a piece of birch charcoal found on Area A at Calanais, dated to between 5350 and 5210 cal BC (SUERC-11588) and check-dated to between 5310 and 5070 cal BC (SUERC-11989). The date came from a piece of charcoal in a possible post-hole cutting a patch of angular stones and underlying an area of beach pebbles, presumably introduced at a later date during manuring with sea weed. It lay immediately east of a glacial erratic which probably broke the ground surface of the time. The mere presence of charcoal does not demonstrate the presence of people. However in western Lewis, with its generally damp climate, wild fires seem a less likely explanation for charred wood than human activities. The survival of pieces of charcoal in these postholes does not date the postholes themselves, except for providing somewhat unhelpful termini post-quem. They seem to have been chance inclusions and suggest no more than fairly local fires long before the postholes were dug. Perhaps some hunter-gatherers stopped beside the glacial erratic, warming themselves at a fire built in the windbreak provided by the stone. It is as good an explanation as any. 24.4.3 Phase 2a: Later foragers or farmers at Calanais Another piece of birch charcoal from Area A from another possible posthole also suggests a human presence, at some date between about 3760 and 3530 BC (SUERC-11589 4880+/-35 BP), though whether this should be ascribed to foragers or to farmers, or the common descendants of both, is a moot point because there were probably farmers nearby; the first cereal pollen in column CN-3 at Calanais Leobag is dated to some period between Discussion and conclusions \ 982

Calanais Survey and Excavation, 1979-88

3875 and 3605 cal BC (see Chapter 21: Palaeoenvironment 21.3.5). Apart from that, the earliest radiocarbon dates from the Western Isles which can be associated with the activities of people who farmed cereals come from Eilean Domhnuill, N Uist. They were measured on charred barley found in a hearth. One date lies somewhere between 3640 and 3350 cal BC (OxA-9157) and the other between 3710 and 3520 cal BC (OxA-9079). They almost certainly demonstrate local farming from about 3600 BC (Armit 2003, 93) though it is in the abstract conceivable that the grain was brought in from elsewhere. The question whether Calanais was built by indigenous foragers who had adopted farming or by incoming farmers seems to me of considerable interest, so I shall explore it in some depth. Those impatient to find out what happened at Calanais should skip to 24.4.4. Many models have been developed for the interaction between farmers and foragers (Zvelebil 1998). At one extreme lies demic expansion (population spread of farmers) and at another adoption-diffusion (ideas spread from farming communities to hunter-gatherers) (Gkiasta et al 2003, 45-7; Bentley et al 2003, 63, 65). These broad models probably over-simplify a complex reality, changing in time (Zvelebil 1998, 7-8). Sheridan has put forward a case for the model of multiple movements of farmers from various points of origin on the Continent with subsequent complex interactions with foragers in which ‘indigenous acculturation is regarded as a subsequent development from, rather than a prime mover of, the introduction of novel traditions and practises’ (Sheridan 2009, 1). I have to admit to an a instinctive liking for these ideas, despite the demonstration by Whittle et al (2011), described below, that the current radiocarbon date-set suggests only one important point of entry to Britain, in SE England. Of possible relevance to Calanais Sheridan has discussed an early ‘diaspora-like spread’ of farmers from Brittany up the Atlantic coasts (Sheridan 2003, 5; 2009). In a 2005 paper Thomas appeared to believe that he had falsified the model within which Sheridan presented her evidence, though he freely acknowledged that narratives other than his own

were possible. His has for some years been one of the more coherent voices arguing against diaspora-like spreads of farmers and similar ideas to explain the arrival of farming in Britain (Thomas 2003, 2005, 125-6). Chapter 25: Technical Note 24.4.3 contains a review of his arguments; here a summary is presented. The background was his belief that although in the southern and eastern parts of Europe farming was introduced by incomers, further north and west the process involved both incomers and adoption of farming by local foragers; and in the lands facing Britain and Ireland farming was largely adopted by indigenous peoples from other such groups. He argued that the presence of domesticated cattle in at least one late Mesolithic community at Ferriter’s Cove in Ireland demonstrates contacts between indigenes and farmers long before farming became prevalent in Britain and Ireland. Woodman and McCarthy have pointed to a lack of lithic evidence for contact between the continent and Ireland in this period (2003, 36) but as Thomas (amongst many others) has pointed out elsewhere people can adopt single traits of a culture while ignoring the rest. He asserted that the evidence for rapid changes in British and Irish material culture around 4000 BC cannot be explained by the arrival of a few boat loads of farmers. Instead it required either a massive influx or the sudden and widespread adoption of new practises by indigenous groups. This claim is difficult to understand; why should contacts with distant farmers have an effect not allowed with local ones? Even more difficult to understand is his argument that, given the homogeneity of indigenous forager groups before about 4000 BC and the regionalised nature of the first farming groups in Britain, the lack of a single donor population on the continent means that indigenous people must have adopted farming. But if indigenous cultures were homogenous around 4000 BC, the subsequent regionalisation of Britain can best be explained by differences between incoming farmers from different parts of the continent, as argued by Pailler and Sheridan (2009, 47-8). Others have suggested that indigenous groups may have adopted aspects of farming practices. Discussion and conclusions \ 983

Calanais Survey and Excavation, 1979-88

Bishop and her colleagues perceived continuity in subsistence practises between forager and farming communities in some parts of Scotland (Bishop et al 2009, 84). However it is difficult to have much confidence in assigning domestic sites to the descendants of indigenes purely on the basis that they made considerable use of wild resources. That evidence could equally well be used to support the idea that cereal farming was introduced by farmers and failed (Stevens & Fuller (2012). Farming is labour-intensive, and Ester Boserup in her ‘The conditions of agricultural growth’ showed that the introduction of farming and its intensification are normally dependent on increasing population densities (1995, 11-14). Some experts have perceived an ‘unusually high density of [Mesolithic] human populations’ on the western coasts of Scotland (Mellars 2004, 172). But there is no evidence from radiocarbon-dated Scottish sites for a major population increase in the centuries leading up to 4000 BC. Instead the number of dated sites is reasonably constant through the 5th millennium and continues at much the same level until about 3800 BC; after a short period fall-off it continues until about 3500 BC (Ashmore 2004, 88). Admittedly mine was a minimal interpretation of a small amount of radiocarbon-dated evidence, much blurred by variations in the amount of C14 in the then atmosphere. Although use of the OxCal Sum Function to produce a diagram of dates spread over millennia has been criticised (Bayliss et al 2007, 10-11, that criticism seems very odd (provided that OxCal’s horizontal bars showing estimates of standard deviations are deleted). The technique is merely a way of plotting probability distributions of dates and continues to be used as such (see Stevens & Fuller 2012, 711 for a list of recent uses of this technique). The radiocarbon dates certainly do not demonstrate a rise in population in the centuries preceding the introduction of farming. Thus indigenous adoption of farming in Scotland purely through long-distance contacts without immigration currently seems unlikely. In the foregoing my argument has not been that all insular foragers failed to or decided not to adopt farming. It is very likely that some of them did,

given the existence of long-lasting and evolving relationships between historically recorded foragers and farmers (Zvelebil 1998, 9-10). Instead the proposition that farming was introduced to Britain largely by incoming groups of continental farmers (themselves perhaps as Thomas argued predominantly the descendants of foragers) has not been falsified. That proposition satisfies current evidence better than the idea that farming was adopted amongst insular foragers only by a desire to bolster group identities. Indeed, as has long been suggested, farming was probably introduced to Britain mostly by small incoming groups of farmers and the main interaction between farmers and foragers happened thereafter (Piggott 1954; Sheridan 2009; Pailler and Sheridan 2009). That interpretation has been given more detail by the results of an extensive radiocarbon-dating programme focussing on the introduction of innovations to Britain and Ireland (Whittle et al 2011). Its conclusion is simpler than that put forward by Pailler and Sheridan (2009). The authors interpret the results to suggest smallscale colonisation of SE England starting slightly before 4000 cal BC. There is the usual problem of distinguishing between adoptions of continental practises by insular foragers, and immigration, so the latter has not been proven. Only isotopic or DNA analyses of much prior and subsequent skeletal material can settle that. The study shows that around 3800 cal BC there was a surge in building activities (Whittle et al 2011, 19-20); By then activities of types usually associated with farmers were abundant enough for their results to feature strongly in the archaeological record in Scotland and Ireland, although their manifestations were different from those in south-east England and there is as yet no clear case for suggesting that the earliest known structures in the north and west were built by the very earliest farming groups there (Ashmore 2004a, 134). However, as Stevens & Fuller (2012) have argued cereal growing may have been introduced early in the 4th millennium BC but that does not mean that it remained a continuous success story throughout that and the next millennium, a point I shall take up again in discussing Phases 2b and 3 at Calanais.

Discussion and conclusions \ 984

Calanais Survey and Excavation, 1979-88

24.4.4 Phases 2b and 3: First farmers near Calanais The currently available evidence is scanty but it seems that by the second quarter of the 4th millennium the main source of plant food in Shetland, Orkney and the Western Isles was barley. In most of mainland Scotland cereals, hazelnuts and wild fruit were all used. In Atlantic Scotland domestic species dominated plant assemblages of c.4000 [perhaps better c 3800 cal BC] to 3300 BC (Bishop et al 2009, 56, 72). On the west coast of mainland Scotland and on the Inner Hebrides, however, people predominantly used wild resources (Bishop et al 2009, 90). The latter conclusion is however based on assemblages of plant remains from a very small number of sites (Bishop et al 2009, 50-57, 73). Even so, farming in Scotland probably had different histories in different areas (ibid 84). Bishop and her colleagueâ&#x20AC;&#x2122;s review could be taken to suggest a sharp distinction between local foragers and farmers in the Western Isles; but the evidence is limited: only 3 sites in N Uist and one in Barra were available for their appraisal (Bishop et al 2009, 57). Also the remains of farming communities so far identified in the Western Isles need not represent pioneer settlement of farmers or adoption of farming; the typologically earliest pottery so far discovered (carinated bowls) was used for several centuries. A less regionally precise study by Stevens and Fuller (2012) of dated cereal remains in Britain has led to the suggestion that cereal farming diminished greatly in importance around 3350 cal BC and remained unimportant, compared to keeping of domestic animals and harvesting of wild resources, until around 1500 cal BC. The authors suggest that the main cause was a climatic deterioration around 3350 cal BC (Stevens & Fuller 2012, 718-9). Although the authors suggest that islanders may have continued to plant cereals because of a scarcity of wild foods their conclusion for mainland populations of the late 4th millennium fit the pollen evidence from Calanais Leobag better; for it suggests that cereal cultivation ceased locally at some date between 3490 and 3020 cal BC, with the reservation that the

evidence may instead have reflected changes in pollen capture rather than changes in pollen rain (see Chapter 21 Palaeoenvironment 21.3.6). But their accompanying suggestion that populations may have crashed at the same time does not match the (very limited) evidence from Calanais itself. There, as described below, there was fairly abundant evidence from dated charcoal accompanied by artefacts that there was at very least one settlement in the neighbourhood during at least a part of the last third of the third millennium (Phase 3). At Calanais a piece of hazel charcoal found on Area F produced a date between 3640 and 3380 BC (SUERC-11601 4760+/-35 BP). It came from a dark greyish-brown slightly gritty clay soil overlying a depression; an underlying layer produced a piece of charcoal of significantly later date (SUERC-11612; 4475+/-35 BP, calibrated to between 3350 and 3020 cal BC) and the soil is interpreted as the result of a combination of levelling, shallow prehistoric cultivation and natural processes. The earlier date falls within the latter part of Pollen Zone CaN-2c, which started between 4120 and 3830 cal BC and ended between 3490 and 3020 cal BC. Evidence for grazing of the local grassland appeared. At some date between 3875 and 3605 cal BC, ribwort appeared for the first time, along with cereal-type pollen. Cereal was grown near Calanais Leobag until the end of CaN-2c when its pollen disappeared temporarily from the record. The earlier charcoal date thus overlaps with a period when farming was being practised in the area. The later date approximates to the transition between CaN2c and 2d, so the charcoal could belong to either. Farming in the vicinity of Calanais Leobag ceased during zone 2d and it is tempting to suppose that the charcoal preceded abandonment of farming at Calanais itself. No samples from Area F were analysed for pollen. It is frustrating that the evidence provided by the earlier date in its (secondary) context is so ambiguous. It merely provides a hint that further excavation in this part of the site may produce significant information about early farming activities at Calanais. Thirteen radiocarbon dates from charcoal found inside the Ring belonged somewhere in a period between about 3350 and 3000 BC (Chapter 23: Discussion and conclusions \ 985

Calanais Survey and Excavation, 1979-88

Radiocarbon Illus 23.1). That corresponds roughly to the transition between CaN-2c and 2d when farming may have failed locally at some date between 3490 and 3020 cal BC and may not have been reintroduced until some date between 2980 and 2500 cal BC. But it must be stressed that the charcoal could be a few centuries older or younger than the transition and that the apparent similarity of the charcoal dates to one another may be illusory because of the plateau in the calibration curve at this period. Also the apparent cessation of cereal farming near Leobag, based on a lack of cereal pollen in the peat at Leobag may instead reflect only a thickening of the birch woodland canopy there. The Phase 3 dates came from 9 contexts inside the Ring. Several of those contexts produced pottery (see Technical Note 24.4.4 for descriptions, catalogue numbers and other details). Areas S and H in the northern half of the Ring produced pottery from two contexts with Phase 3 dates. Nearly all were from E/MN Corky pots although there was also a piece of a Hebridean Incised pot and two sherds probably from Beakers. The other three contexts (352, 360 and 369) all came from Area D in the southern part of the Ring, and were probably imported soils and clays. Again most of the sherds came from E/MN Corky pots. But there was a definite fine Beaker sherd from basal cairn layer 360 while layer 369 in addition to many E/ MN sherds produced a few E/MN Non-corky sherds, an E/MN Hebridean Incised sherd and a probable fine Beaker sherd. Most of the pottery could belong to the same period as the Phase 3 charcoal. It seems likely that these contexts included both artefacts from the source of the soils and clays and artefacts contemporary with their movement into the Ring. Indeed in addition to Phase 3 dates layer 369 included charcoal of the first half of the 3rd millennium BC. However three of the possible Beaker sherds from Context 369 (Cat 759_761) may have been from a much later type of pot. Those sherds, if they were indeed much later in date, may have been intrusive, for there were many ard marks at the same level and Victorian and possibly earlier disturbances nearby.

24.4.5 Further thoughts about Phases 2b and 3 The indigenous groups in the Calanais area before farming started were small and mobile, moving from resource to resource throughout the year; there is no particular reason to see them as having very complex societies (Zvelebil 1998, 12; Hardy & Wickham-Jones 2010). The best prospect for finding relevant evidence locally, given the general peat cover and changes in sea level, is exploration of the shallow waters of inner East Loch Roag. The simplest interpretation of the pollen evidence from Calanais Leobag is that groups of farmers settled near Calanais from neighbouring areas at some date between 3875 and 3605 (the interpolated date for appearance of cereal pollen in pollen column CN3 at Calanais Leobag). If farming had thriven in mainland Scotland from 3800 cal BC it is credible that resource pressures had grown and that people sought new land rather than improving the productivity of their areas of origin (Boserup 1995). The evidence from Calanais Leobag would fit well with the idea that small groups of farmers moved from time to time to previously unexploited areas. Of course it cannot exclude the possibility that local foragers started farming. Even though the calculation of the durations of the farming period and the subsequent period without evidence for farming is beset with uncertainties both seem to have lasted for many generations (Chapter 25 Technical Note 24.4.5a). The early local farming period lasted for between 285 and 585 years (say between 12 and 30 generations), and part of the area now occupied by the Avenue may have been used for cultivation (see Chapter 11: Area F). The subsequent gap in the record of cereal pollen at Calanais Leobag corresponded to some 330 to 520 years (between 14 and 25 generations). It is interpreted here as a period without local farming. However the possibility that cereal pollen ceased to be deposited at Calanais Leobag in pollen zone 2d because and only because of a local increase in tree canopy cover cannot be ignored (see Chapter 21.3.6). The pollen evidence from Calanais Leobag suggests that whether or not after a period of non-cultivation cereal was grown locally from a Discussion and conclusions \ 986

Calanais Survey and Excavation, 1979-88

date between c. 3000 and 2500 cal BC and the evidence from Calanais itself suggests that it was grown from around 3000 BC. Despite the gap in the record of cereal pollen at Calanais Leobag there were almost certainly settlements in the general area of Calanais at some times between about 3300 and 3000 BC. The preferred interpretation of the presence of late 4th millennium charcoal and pottery inside and near the Ring is that the vast majority of it was brought there during Phase 7 (at some time between 2600 and 2350 cal BC) in clay and soil from by then old settlements or activity areas. That ‘ancestral material’ may have been intended to be the soils and clays themselves rather than the potsherds included in them. One can speculate that it in some way reflected claims to the monument from more than one local community. A settlement from somewhere in the early to mid Neolithic may lie near the pre-modern grain-drying kiln just to the west of the stone setting where stonework and (mainly E/MN noncorky) potsherds were found in a test pit (Chapter 13: Test Pits G and J; Chapter 18: The Pottery Assemblage). In favouring the idea that farming was introduced by incomers, I have to follow Thomas (2005, 113) in stressing that this is only ‘one attempt to reconcile recent discoveries’. A range of other possibilities is discussed in Chapter 25 Technical Note 24.4.5b. More cogently there is a clear interpretative problem in, on the one hand, the presence of charcoal probably derived from settlements of the period 3300 to 3000 cal BC, and on the other hand the gap in the cereal record in the pollen at Calanais Leobag in the later 4th millennium cal BC (if that gap was not caused by closing of the birch woodland canopy over the sampling spots). More fieldwork including a search for settlements and a better-dated pollen record is required. 24.5 Phase 4: The early ditch 24.5.1 Description of the ditch In Illus 24.5 shows Area D south of the chambered cairn. The early ditch is outlined by very pale green streaks where primary silts have been

truncated by later ploughing. They are most readily seen to the right in the bottom third of the picture. The nearer green streak runs in a curve as far as the stones near the central monolith at top left. The ditch is partially obscured by the remains of cultivation bed soils and it was clear that the stratigraphic sequence was first the ditch, then the cultivation beds and then the central monolith. The fills of the ditch in the foreground of Parts of the ditch had been excavated to the right of the mini-baulk. There was no evidence for the ditch having been cut through a soil or a turf line. Although a fragment of a turf line survived on Section DC (Illus 24.6 that seems to have been associated with a pit. Probably the uppermost fills of the ditch and the surrounding contemporaneous ground surface had been destroyed during creation of the cultivation beds. In some places the top several centimetres of the natural green clay by the ditch had been modified into dirty orange clay. Near the east baulk of Area DI a small dump of dirty orange clay overlapped both the fills of the ditch and the clay through which it had been cut. Another orange clay dump was recorded near Ring stone 47. They were reminiscent of a basal orange clay hump at the east end of Area DII which also lay on natural clay without an intervening soil or turf line. It seems that there too any pre-existing turf line had been removed during the succeeding cultivation phase. It was not clear what the humps were because we did had too little time to excavate the lowest levels of Area D, but because one of them overlay truncated fills of the ditch and natural clay they may have been related to cultivation bed building. The surviving top level of the ditch was about 0.6m wide and 0.2 deep. There were two bands of greasy clay in it which may represent short duration stabilisations. Where best preserved its main lower fills consisted of gritty green clay probably derived directly from the natural clays. No artefacts were found in it. In the northwest it may have stopped about 2m from the position of the subsequent central monolith (Illus 24.5-24.6), but at this point excavation did not reach below cultivation soils. Near where it met the east baulk of Area D its Discussion and conclusions \ 987

Calanais Survey and Excavation, 1979-88

Illus 24.5 The early ditch c. 5 June 1981 (Film 1981.20.11]

Illus 24.6 The early ditch 921 and extracts from Sections 95a and 95b [NMRS DC38194, (part) DC38226 and DC38227-8] Discussion and conclusions \ 988

Calanais Survey and Excavation, 1979-88

north side was cut by a cultivation trough and the plan made of it therefore has a kink northward which probably does not reflect its original line (Illus 24.6). With one possible exception no traces of any related features were identified within its circuit, but that may have been partly because the overlying cultivation beds were not fully excavated and partly because of destruction by cultivation. The possible exception relates to a non-preferred hypothesis that slots formed an egg-shaped enclosure inside it (see Phase 8, Chapter 24.7). 24.5.2 Interpretation as a linear scoop Given that its uppermost levels had been destroyed, it is faintly possible that the ditch was part of a structure like that at Low Clone on the Solway Firth, where a long scooped hollow 20m in length and c. 0.5m deep with a flat bottom was associated with stake holes and linear stone settings

(Wickham-Jones 2004, 231, 234 Figure 12.2). But the surviving ditch at Calanais was too narrow and shallow to form a useful windbreak and no potentially contemporaneous hearths, stake holes or linear stone settings were found in or by it. 24.5.3 Interpretation as a circular enclosure On Illus 24.7 the features shown on Areas H and B were later and they are included because they may have destroyed earlier evidence. Two circles defining possible continuations of the ditch if it were circular are depicted, one 6m in diameter and the other 13m in diameter. In Area H a 6m diameter enclosure ditch would have been partly removed by the slot and other complex features. The ditch would have run across the entrance of the passage where there was a change in the green clay possibly caused by the presence of a man-made feature. However on Area DIII where a large expanse of the natural green clay was

Illus 24.7Â The early ditch and the likely range of its diameter if it were circular Discussion and conclusions \ 989

Calanais Survey and Excavation, 1979-88

revealed it included at least two distinct varieties of clay; a similar variation on Area B provides a preferred explanation. If the enclosure diameter was about 13m - as if it might be a functional predecessor to the Ring in a more easterly position - it is conceivable that its traces on Area B were confused by the later earthen bank just inside the line of the 13m circle. There was a shallow broad depression outside the bank there. But to the north it would have hit a large prone slab (Illus 24.7). There was no trace of a ditch in the south-eastern part of Area B. It is unlikely that it would have been missed even though the earliest strata here were not excavated. In Area S there was a slight depression just inside the 13m diameter circle but that was at a very superficial level (Level 2) and levels below 0.15m (the depth of the turf which was to be replaced here) were not excavated. Other hypothetical circuits with diameters between 6m and 13m run up against similar problems. Altogether there was no convincing continuation of the ditch in the excavated areas if it had indeed been circular; and only the c. 6m diameter one cannot be dismissed entirely. I am not aware of any well-dated analogies for a roughly circular enclosure of about 6m diameter as early as the first appearance of cereal pollen at Calanais Leobag (in Pollen zone CaN-2c at some date between 3875 and 3605 cal BC). The nearest parallel of perhaps that period in Scotland is Structure H at Chapelfield, Cowie, Stirling (Atkinson 2002, 145-7). There a â&#x20AC;&#x2DC;pressure trenchâ&#x20AC;&#x2122; about a metre wide, its edges demarcated in places by stake holes, enclosed an oval area 6.6 by 5.7m (measured from Atkinson 2002 Illus 6 on page 147) or 6m across (according to the text). There was a ragged pit near its centre. Its longer axis pointed NNW-SSE and its entrance was near the south end of the east side and about a metre wide. At the north side of the entrance was an L-shaped slot. A radiocarbon date (AA-26225; 6840+/-85) was obtained for charcoal in a posthole in this slot, but it was from multiple fragments of pine charcoal which were probably residual, for the related Pit II produced a similar date (GU-7201 6710+/70 BP) from pine charcoal but a hazel nut shell from the same layer 694d in the pit produced a date calibrating to 3980 to 3780 cal BC (OxA-

9234; 5085+/-45). The lumps of pine charcoal may have come from fossil trees in the surrounding bogs as Aldritt suggested in her study of the plant remains (ibid 178). So Chapelfield Structure H probably dated to the first quarter of the 4th millennium and that corresponds roughly to the likely dates for the earlier part of Pollen zone CaN-2c at Calanais, overlapping with the first appearance of cereal pollen. The analogy does raise the possibility that the structure at Calanais was a domestic structure of some kind, the interpretation preferred at Chapelfield (ibid 188). But the comparison is a very loose one based on suppositions about the date, shape and size of the Calanais ditch. Also, given the regional variations in Scotlandâ&#x20AC;&#x2122;s archaeology (as discussed in Chapter 24.1.2 - 21.1.5) there is no reason to suppose that people were doing the same things at Calanais and near Stirling. So the Chapelfield evidence is not a good guide to the nature and date of the Calanais ditch. If the ditch had instead been dug at some date around 3000 cal BC it might have had a similar function to small hengiform enclosures. But only two radiocarbon-dated Scottish examples of small hengiform enclosures, Pullyhour in Caithness and the henge at Broomend of Crichie, have been published in modern times and both date much later. Pullyhour had a terminus post quem of the middle of the 2nd millennium BC (Bradley and Lamdin-Whymark 2009, 3). Broomend dated to after 2000 BC (Sheridan 2007, 221). Neither can usefully be compared to the Calanais enclosure except at a very abstract level. The ditch need not have outlines a (roughly) circular area. An extremely loose analogy might be drawn with the very much larger enclosure ditch of Structure 2 of the ceremonial complex at Balfarg Riding School in Fife. Only partly preserved, the enclosure was of the order of 50m in diameter. The ditch itself was between 0.5 and 1.1m deep and its width varied between 2.2 and 4.5m (Barclay and Russell-White 1993, 90). It was dated to around 3000 BC (ibid, 47). At a similarly very broad conceptual level the Calanais ditch could be seen as a miniature version of the early enclosure at Stonehenge, which probably dates not much earlier than 2910-2670 Discussion and conclusions \ 990

Calanais Survey and Excavation, 1979-88

cal BC, the date for piglet bones in the lowest secondary fills (Parker-Pearson et 2009, 29). But that too was very much larger. Both of these enclosures were very different in scale from that implied by the ditch at Calanais. All came from different regions with possibly quite different traditions. They should not be used to suggest its function. 24.5.4 Interpretation as a sub-rectangular enclosure Perhaps the ditch was not part of a round enclosure. Could it have been straight-sided with rounded ends (Illus 24.8)? Two linear features, one in Area H and the other in Area BV, might be called into aid to reconstruct such an enclosure despite the southern one being interpreted as a cultivation trough and the northern one (as excavated) being at a slightly ‘wrong’ angle.

Illus 24.8 Reconstructed long enclosure based on the early ditch and two supposed cultivation troughs Could these have originally been other parts of a sub-rectangular enclosure ditch? If so, and if the ditch dated to around 3000 BC, it is conceivable that the enclosure was a structure similar in shape to Balfarg Riding School Structures 1 and 2 (Barclay and Russell-White 1993 173 Illus 71). The latter were post-built not ditched, but their walls were intended to define an open area. They were between 9 and 9.5m wide, with gently rounded ends. Structure 1 was slightly under

19m long - almost exactly twice as long as it was wide. The possible enclosure at Calanais would have been slightly less than 7.5m wide with gently rounded ends (Illus 24.8). If similar in proportions to the Balfarg Riding School structures it would have been around 15m long. Its purpose would have been to define an area in which human bodies were exposed after death (Barclay and Russell-White 1993 182). The broad similarity of Grooved Ware pot 61 from Calanais and some of the grooved ware from Balfarg Riding School could be called in aid of the suggestion that similar things might have been happening at both places, albeit with the usual proviso about using distant comparanda and the perhaps more decisive criticism that the grooved ware pot at Calanais was found in a stratigraphically much later context than the ditch. For this hypothetical reconstruction to succeed we would have to have missed remains of a ditch up to 0.6m wide and 0.2 deep forming the north side of the structure on Area B. Such a feature might well have been destroyed within the later oval enclosure immediately east of the Ring but even if partially ploughed out it should have been visible in the eastern part of the trench. Near the position of the projected northern ditch of the hypothetical Balfarg-style enclosure there were indeed enigmatic linear features (Illus 24.8). They were interpreted as possibly the remains of part of a cultivation bed system. But their fills had no characteristics in common with those of the ditch. The fact that we did not find any postholes similar to those inside the enclosures at Balfarg Riding School is also a deterrent to embracing the idea. Overall, the idea that the ditch was part of a similar enclosure for exposing bodies on raised timber platforms is almost certainly wrong. 24.5.5 Discussion It is far from clear what shape and size of area the ditch enclosed. It is even less obvious why it was dug. Although it could have been part of a domestic structure or a funerary enclosure, it could have been created for a more sophisticated reason, for instance to delineate an area in which the moon appeared to set at the lunar maximum Discussion and conclusions \ 991

Calanais Survey and Excavation, 1979-88

when viewed from some position near the north end of the ridge on which Calanais sits. It is just east of the lunar major standstill moon-set position viewed from the north-east stone of the later Avenue (Curtis and Curtis 1998, 23). Given these uncertainties the preferred interpretation is simply that it enclosed an area of unknown shape and dimensions, dated to c. 3000 cal BC or earlier and was subsequently obscured by cultivation beds. 24.6 Phase 5a: Cultivation beds 24.6.1 Description and reconstruction On Illus 24.9 and 24.10 the troughs and possible troughs between cultivation beds are shown in brown. In the southern part of the Ring on Area D the beds were labelled 1 to 4. The beds averaged about 1.5m between the centres of flanking troughs (note the subsidiary scale set at right-angles to the beds). They seem to have been created by spade or hoe cultivation. No definite ard marks were associated with them, and in places, notably on Area DII, turfs appear to have been cut and turned upside down to help form them. The optimum breadth for ‘no-tread’ cultivation beds worked by modern hoes is about 1.3m. With a trough about 0.3m wide most of the beds at Calanais fitted that prescription fairly well. Silting and turf lines formed in some of the troughs. The whole system seems to have been affected by subsequent level tilling and in some parts of Areas DI, BV and H by manual levelling. Only the troughs in the southern parts of Areas DII and BV and the middle trough in Area H were well formed; and only Area D Bed 4 and Bed 2 were clear. All the rest were to varying degrees represented by fairly minor changes in the composition of the soil or by slight changes in height. Gulley 100 in Area BI was the base of a much later drainage ditch running from the mouth of the chambered cairn passage but it is conceivable that the drain followed an earlier depression. Gulley 121 in BIII (not part of the system but on roughly the same orientation) was a textural change in the lower part of the later plough soil rather than a neatly defined feature. The rubbly ridges and very

vaguely defined troughs in Area BIII may instead have been related to the entranceway of a subsequent enclosure (Phase 9b). The beds in Area C, to the east, were not excavated and at the level where the troughs between them appeared there were traces of other features (see Chapter 8 Area C Illus 8.26). One may have been the base of a circular mound; but it may instead have been the west end of the upper middle cultivation bed on Illus 21.10. So the reconstruction depicted on Illus 24.9 and 24.10 is tentative. Most elements of the system, considered individually, can be interpreted in different ways. Nevertheless, the consistency of the orientation of the putative beds and troughs and of later features which might have followed their line suggests that even if the detail is sometimes wrong there was an early cultivation system similar to that shown. The system on Areas B-D-H and that on Area C need not have been in the same ‘field’ as each other. The best-defined cultivation bed on Area C was significantly broader than those in B-D-H, although the other two were about the same breadth as on Area D (Illus 24.9). Their general orientation was closer to east-west. They remained largely unexcavated after their surfaces were revealed. No pottery was retrieved from their soils. 24.6.2 Dating the cultivation beds Cereals reappeared in the pollen record at Calanais Leobag at the transition from Pollen zone 2d to 3a (Chapter 21: Palaeoenvironment). The turf lines immediately overlying the beds were also zoned to that transition. The cultivation beds were probably used and abandoned somewhere between 2980 and 2510 cal BC. Other dating evidence from Calanais suggests that they belonged around the earliest part of that range, in round terms dating to c. 3000 BC. 24.6.3 Discussion and Comparanda Cultivation beds thicken the soil, improve drainage, and allow the sun to warm the soil better, adding a few days to each growing season. In the centuries around 3000 BC average temperatures Discussion and conclusions \ 992

Calanais Survey and Excavation, 1979-88

Illus 24.9Â The cultivation beds and troughs in the central area (note supplementary scale shown in Area D)

Illus 24.10Â The cultivation beds and troughs (with 10m grid squares) Discussion and conclusions \ 993

Calanais Survey and Excavation, 1979-88

around Calanais were possibly much like those of present day Argyll, and those extra days would have been precious. There was no evidence from the Calanais excavation for the precise crop (the only cereal grains discovered date to the mid 2nd millennium BC); it was presumably either a variant of barley closely related to the bere still grown today, or emmer wheat or both (Bishop et al 2010). Cultivation beds have long been known from a variety of contexts under 4th to 2nd millennium mounds in England and Wales (Barclay 1993, 232). However, the discussion here will be restricted to Scotland and Ireland. Nearby sub-peat cultivation beds at Calanais Fields have not been dated directly. Peat started growing over them at different times in the 1st millennium BC (Chapter 23: Radiocarbon; Flitcroft et al 2000; Johnson et al in prep). But given the likely interval between abandonment of beds and peat growth initiation on Area C at Calanais it is conceivable that those at Calanais Fields were also significantly earlier than the basal peat there. At Machrie North on Arran several areas of cultivation beds were found under peat (Barber (ed) 1997, fig 59). The beds in Trench 24/50a had a wavelength of 0.8 to 1m. They were interpreted as having been shifted each year, the furrow of one year being overlain by the bed of the next (ibid, 108). Beds in other areas were not excavated after they had been identified. On Area 24/50 beds set at much the same angle as each other were identified by sample-trenching in an area at least 60m by 55m. Beds at a different angle were found at the NW edge of the area investigated, the change in angle reflecting a change in topography (Barber (ed.) 1997, 106 Figure 59). Pollen analysis of a nearby column produced cereal pollen at an estimated date of 5375 BP (Robinson and Dickson 1993, 117). A large error of +/-217 years should be assigned to this estimate (Technical Note 24.6.2). The date then calibrates to between c. 4700 and 3700 cal BC. But the earliest basal peat date from typical parts of the Machrie North area varied between the 1st millennia BC and AD. In effect the cultivation beds may have been of almost any date between about 4000 BC and 1000 AD. Beds found under the barrow at North Mains, Strathallan, were spade or hoe dug, up to 0.15m

high and had a wavelength of 1.8-2m (Barclay 1983, 191, 231-2). They were thus 12% to 25% wider than most of those at Calanais. They overlay pits, one of which contained a sherd of a ‘Neolithic’ bowl (ibid, 191, 215). A charcoal deposit at the interface between the fossil soil and the overlying mound produced a radiocarbon age (GU-1134; 3805+/-100 BP) (ibid, 192) which (after its quoted error has been adjusted to 140 to compensate for weaknesses in dating technology at the time it was measured) calibrates to some date between 2900 and 1700 cal BC. The charcoal appears to have been mixed and the precise relationship of the context to the cultivation activity was uncertain (ibid 192). Field systems of roughly the same epoch but without cultivation beds survive in peat-covered areas elsewhere. At Scourd of Brouster, Shetland, Whittle excavated a complex of field systems and domestic structures (Whittle at al 1986). The first phases of two of the houses were dated to between about 3350 and 2900 cal BC. The fields formed an irregular patchwork covering a huge area (Whittle et al 1986). Charcoal from cultivation levels in a bog was dated to between 3050 and 2450 cal BC. The latest dates from the area (from charred barley in House 3 Phase 1) fell in the first half of the 3rd millennium BC. At Machrie North on Arran another early field system (24/03) was dated to the ‘Neolithic period’ by the occurrence of grooved ware in a small pit cut into hill-wash deposits overlying them (Barber (ed) 144). Their absolute date is not clear but they should belong before about 2300 cal BC, despite the fact that technically the sherds provide only a terminus post quem for the pit; the sherds may have been collected during later activities nearby, as were the Food Vessel sherds on Calanais Area C and the Hebridean incised ware sherds on Calanais Area E. The system appeared to consist of field enclosures measuring up to 200 by 50m (Barber 1997, 144-5) and may have been very different from that at Calanais. Early field systems have been found at several places in County Mayo, Ireland; others survive below peat there and in County Donegal. They varied from large regular fields to small irregular ones. Cooney, in reviewing the evidence, cites a Discussion and conclusions \ 994

Calanais Survey and Excavation, 1979-88

suggestion by Woodman and colleagues that some of them may have been related to control of animal movements (Cooney 2003, 50). At CĂŠide Fields, near the north coast of County Mayo, there was a 12 km2 system of large regular enclosures. Pollen analysis suggested a pasture phase starting about 4000 BC that ceased by 3200 cal BC (5200 to 4500 BP; Caulfield et al 1998, 635). They thus seem to have been used somewhat earlier than the cultivation beds at Calanais. However Caulfield did not record the presence of cereal pollen or cultivation beds so the comparison is very loose, particularly because farming in 4th millennium Ireland and Britain had very different emphases (Thomas 2005). At Belderg Beg in County Mayo a large area of shallow cultivation ridges broadly comparable in size to the cultivation beds at Calanais has been studied using micro-morphological and palynological analysis (Verrill & Tipping 2010, 12141225). A phase of ard cultivation immediately preceded the ridges. Midden material was added to the soil of the ridges. A radiocarbon date for basal peat suggested a Mid to Late Bronze Age date for the cultivation. Barber thought that the survival of the field systems at Machrie North may have been aided by a lack of â&#x20AC;&#x2DC;late Bronze Ageâ&#x20AC;&#x2122; (late 2nd to early 1st millennium BC) activity (Barber 1997, 145). This is a cogent point. Remains like those at Calanais are even more fragile than the dilapidated field walls on Machrie North. It seems all too likely that most remains of many 4th and 3rd millennium field systems disappeared well before the start of history. Nevertheless, the brief survey of early cultivation bed systems in Scotland presented here makes it clear that abundant evidence probably still survives in Scotland under peat and 4th to 2nd millennium earthen structures, and possibly under colluvium. Direct evidence from Scotland for early cereal growing exists in the form of radiocarbon dated grain from around 3800 cal BC and from later in the 4th and 3rd millennia (Ashmore 2004, 127, Fig 11.3). A review of plant remains from Scottish excavations (Bishop et al 2009) suggests that cereal cultivation was widespread in Atlantic Scotland during the period from 4000 to 2500 BC, although there were

considerable regional variations in subsistence practices. Stevens and Fuller have argued from a study of radiocarbon dated grain that in many areas of Britain cereal farming failed around 3350 BC when there was a deterioration in climate, although perhaps not in offshore islands (2012, 715-6, 718). The study by Bishop et al (2009) suggests much local variation in subsistence in Scotland but its main impression on me is how patchy and poorly dated much of the evidence for cereal cultivation in Scotland is. Perhaps Barber was too pessimistic (1997, 146-7) in supposing that lowland and highland agricultural systems would have been very different from each other during every past period, rather than being variations on an overall theme. The examples from Calanais, Machrie North and North Mains allow a prediction that 3rd millennium cultivation systems discovered in future in Scotland, and possibly also 4th millennium systems, will consist of beds or ridges up to 0.2m tall separated by shallow troughs or fairly abrupt gullies, with a wave-length between 0.8 and 2m. Ard marks may delineate beds as at Machrie (Barber 1997, 106) but judging by the evidence at all three sites most cultivation earlier than 2500 BC will prove to have been by hoe or spade. The evidence from Machrie North suggests that (some) systems will be extensive and will lack formal boundaries (Barber 1997, 145). That seems to have been the case at Calanais. This is a bold prediction, given the small size of the present sample and the ambiguity of the chronological evidence. Also, a drastic reduction in cereal farming on the mainland between 3350 and about 1500 BC remains a possibility, as Stevens and Fuller suggested. So surviving remains of early cereal cultivation systems may be restricted on mainland Scotland to the half a millennium between 3800 and 3350 cal BC. 24.7 Phase 5b: Turf line formation 24.7.1 Formation processes and distribution Before turning to the detail of the earliest surviving turf lines it seems useful to provide a general description of the occurrence and nature of turf lines at Calanais. Discussion and conclusions \ 995

Calanais Survey and Excavation, 1979-88

Illus 24.11Â Some of the sequences including turf lines, which are shown as black lines, in dashes where weak Apparently coherent and widespread turf lines survived on Area D, Area H and the western part of Area B (Illus 24.11). Most of them had been preserved by overlying clays, the latter often altered by soil processes to light grey or grey-brown. Weaker ones also left traces on those areas. Fragments of turf lines survived elsewhere, particularly in the fills of depressed areas like the early ditch on DI or the troughs between cultivation beds, foe example on BV. On Area DII three turf lines in a cultivation bed may include stacked layers of turf rather than consisting solely of in situ turf growth. Turf lines at the south end of BIII, BV and on Area S are not shown on Illus 24.11. The colour and texture of their litter layers varied. Some turf lines were black and very smooth

textured. They may originally have had a mossy element. Others were grey or brown with an almost greasy texture. The soil horizons below the litter layers were also diverse. Some were a thin near-white band of stone-free clay. Others were blotchy grey and brown. In some cases the litter layer itself was not distinct from its underlying horizon. Distinguishing between such pairs of layers and a worked soil with no litter layer was sometimes possible only during post-excavation analysis of box-samples. The most extensive strong turf lines survived on Area D (contexts 334 and 365) and on Area H (contexts 751 and 758). On D the context numbers were applied to two patches of old ground surface in each of which was a stack of two turf lines with a thin grey layer between them. But their apparent Discussion and conclusions \ 996

Calanais Survey and Excavation, 1979-88

coherence was probably partly an illusion. It seems highly likely that some pits had been cut from the top surface and turf had re-grown over them, for the pollen in the pits suggested a date later than that of the main turf line. Also, in places, the two turf lines merged; elsewhere they were well separated (for instance by clay 389 near the east section on DI on Illus 24.11). In its final form the upper turf line included pollen characteristic of the transition from CaN-3a to CaN-3b which dated to between 2560 and 2200 cal BC at Calanais Leobag. More generally on Areas B, D, H and S a variety of circumstances and successions was recorded and determining the significance of all of the variations and interactions of the turf lines proved impossible using the traditional excavation technique of trowelling and attempting to remove later strata first. The process of turf formation in Area H may have been as complex but less clay and soil seems to have been dumped there before the cairn was partially demolished and its stones taken away for construction elsewhere. The sequence on Area H was therefore somewhat simpler to record. In the western part of Area BI the sequence was similar to that on Area D, with mostly two turf lines a few centimetres from each other, or one turf line. There were traces of a third higher turf line between Ring stones 42 and 43, but the surviving sequence had been considerably complicated by subsequent activities related to the Ring, chambered cairn and enclosures. The crucial question is whether the turf lines in various parts of the Ring and immediately to its east ever formed a single old ground surface, and if so, how often. One widespread pair of turf lines does appear to have consisted of the pre-Ring old ground surface and a turf line which grew over spread spoil from the digging of the Ring stone pits. It is represented by 164 and 162 in Area B and the two layers of 334/365 on Area D (Illus 24.11).But the sequences further west in Area D may well include earlier turf lines and also, possibly, sets of cut and stacked turfs. A superficially similar sequence of turf lines in the top of ditch 921 is not included in Illus 24.11. We found no evidence (apart from the very imprecise hints offered by stratigraphy) to show how long individual turf lines had grown. In Areas

B, D and H where an area of turf had been partially covered by clay it stopped developing while continuing to form elsewhere. Their colour (and thus distinctness) of their litter layers could have had more to do with the amount of moss in them than their longevity. Clearly turf lines would be treacherous chronological indicators. Overall, the answer has to be that from time to time turf seems to have been growing on an old ground surface common to more than one of the excavated areas, but different parts of the ground surface were covered up at different times, the turf lines were cut by intrusions which healed over, and other turf lines and layers of cut turfs grew or were laid at different times. In practise, then they could not be assumed to be of the same date as each other. Each comparison had to be argued separately 24.7.2 The earliest turf lines In outline the way in which the earliest turf lines formed was fairly clear. It is illustrated in Illus 24.12. 1. Thin turf and soil was removed and piled up with some of the turfs left upside down; this created a set of beds and troughs with a wavelength of around 1.6m; 2. hoe cultivation led to accumulation of soil, mud, silts, clay, weeds and a few cleared stones in the troughs; 3. at least some of the troughs were cleared out at least once leaving stones but removing silts and decayed organic material; 4. the troughs again silted up with short lived turf lines (or layers of organic material from decayed weeds) forming and being covered in turn by silts and debris; 5. clay and soil was dumped on the old beds in some areas; some of the cultivation beds were truncated along with the turf lines which had formed on them; this process was probably repeated at least once in some areas. Discussion and conclusions \ 997

Calanais Survey and Excavation, 1979-88

Illus 24.12Â Formation and deterioration of turf in the cultivation bed system Barber (1977, 108) may well have been right to suggest that when early cultivation beds were renewed their position was shifted by half a wavelength. If so, further complications would be expected. For instance the trough between beds 1 and 2 in Area D contained a succession of turf lines (or thin layers of decayed organic material). There and elsewhere new turf lines formed, merging with earlier ones. In several places there were two successive thin turf lines but in some of the troughs there were remains of three or more and elsewhere sometimes only one. Sometimes that was because of truncation and sometimes because a turf line had continued to grow while other parts of it had been covered by soil or clay.

CaN-3aii and to the start of CaN-3b. The earliest surviving turf line on Section 105 may have been a patch on the pre-cultivation surface normally removed elsewhere (or evidence for the shifting of beds sideways between one period of cultivation and the next; more excavation is needed to resolve this). The second turf line probably corresponded to abandonment of the cultivation beds.

24.7.3 Chronology of the earliest turf lines on Area D Two sections were sampled using Kubiena boxes and analysed for their pollen contents (Illus 24.13). On Area DI the earliest surviving turf line on Section 105 belonged in pollen zone CaN-2d. A second one belonged at the transition from CaN-2d to CaN-3ai. Subsequent turf lines, probably post-dating the cairn, belonged to

Illus 24.13Â Location of Sections 105 and 102a The earliest surviving turf line on Section 102a, a metre or so from the central monolith, was asDiscussion and conclusions \ 998

Calanais Survey and Excavation, 1979-88

cribed to 3aii and the underlying silty layer 904, interpreted as its soil, was ascribed to CaN-2d and CaN-3ai. The sequence to the north of the sampling point was difficult to interpret but included at least one higher turf line and a fragment of turf. A patch of black greasy clay 379 in the fill of the southern trough of Bed 4 on Area DII was assigned to the beginning of pollen zone 3ai. It may have been a turf line or the decayed organic-rich remains of weeds pulled from the adjacent cultivation beds and dumped in the trough. The transition from CaN-2d to CaN-3a took place between 2983 and 2510 cal BC at Calanais Leobag but the interpretation of the dates of later features at Calanais itself suggests the true date of the first turf line covering part of a cultivation bed was at the beginning of that range. Thus the cultivation beds probably ceased to be used around 3000 BC. Those interested in detail should refer to Chapter 21: Palaeoenvironment. 24.8 Phase 6a: The central monolith 24.8.1 Excavation The height of the monolith before excavation was 4.8 m (Ponting and Ponting 1984, 28) or 4.75 m (15 feet 7 inches) according to the original RCAHMS survey (NMRS RCD/13/12). The pit dug for the monolith was about 1.7 m across from east to west with the eastern part more than twice as wide as the western. Its east side sloped at between 20 and 30 degrees. Its west side was more nearly vertical (Chapter 9: Area D). The shallow angles of the eastern side of the pit imply either that the side of the pit was slanted for erection of the monolith or that its edge was compressed as the monolith was levered up. The asymmetrical pit solution is proffered in Illus 24.15. If it be assumed that about one fifth of it was buried in the ground then the pit in which it was set would be 1.2 m deep and the stone would have been about 6 m long. At about 6 m long by about 0.3 sq. m in cross section, and a density of 2.52g/cm3 (measured from a large Calanais gneiss pounder in my temporary possession), the monolith would have weighed about 4.5 metric tonnes.

Illus 24.14Â The Central Monolith from the west [David Henry, Historic Scotland 1984]

Illus 24.15Â The pit for the monolith 24.8.2 Raising the monolith Pine could have provided substantial timbers which could have been used for erecting stones. Substantial fossil pine stumps survive under the Discussion and conclusions \ 999

Calanais Survey and Excavation, 1979-88

Illus 24.16Â The monolith and pit peat in several parts of Lewis. The main cause of an early loss of pine woodland cover about 6800 BC appears to have been one or more short periods of violent weather. Radiocarbon dating places many other pine deaths between 3700 and 2700 cal BC (Dickson and Dickson 2000, 37-39, 67). That covers the period when the monolith was erected. Pine pollen appears (admittedly in small quantities) in the earliest levels of the CaN-3 part of pollen diagrams at Calanais and is present in all subsequent levels. Pine charcoal at Calanais was radiocarbon-dated both to between about 3400 and 3000 BC (SUERC-11597; SUERC-11592) and to between about 2500 and 2300 BC (SUERC-11590; SUERC-11591). Given that only four pieces of pine charcoal were dated that is

not very significant, but the earlier charcoal might have come from trees available about 3000 BC. Margaret and Ron Curtis have conducted experiments in moving and erecting stones of about 1.5 tonnes in the Calanais area. They tried both rollers and a wooden frame consisting of a horizontal bar held up between two tripods. In the latter experiment the stone was joined to the bar by two pairs of ropes, and poles were twisted between each of the sets of paired ropes to provide lift. This arrangement allowed a few people to move the stone slowly across the landscape (Curtis and Curtis 2008, 49). Other possible techniques for moving stones with a small number of people have been described. It has, for instance, been suggested that stone or wooden ball-bearings could have been used on wooden rails. The existence of carved stone balls of consistent diameter, with decorative affinities to Grooved Ware, has been called in aid of this theory. The technique allowed 8 people to move stones of 3.3 to 6 tonnes at a rate of a few miles an hour (Young 2011, 44-5). Curtis and Curtis found that levering a stone to raise it required a large amount of timber. Beyond an angle of about 60Â° the stone became difficult to manage. The easiest way to proceed was to raise the stone before the pit was dug, using a bar and two tripods as above, with four rope-pairs used two by two to raise the stone to the vertical (Curtis and Curtis 2008, 49). Their method required only the two of them to erect a 1.5 tonne stone. In discussing the erection of stones at Stonehenge, R J C Atkinson (1960, 129-134, particularly 133) suggested that the foremost of a set of rollers could have been used to pivot the stone through an initial angle, reducing the effort required to raise it. Given the observed dimensions of the pit at Calanais, once the leading roller reached the edge of the ramp the front 1.7 m of the monolith could have been hanging over the pit (Illus 24.17). It could then be raised until it came to rest on the ramp. Atkinson quotes E H Stones (1924) to suggest that shear-legs would greatly diminish the effort required to pull a monolith towards the upright. Thereafter the monolith would tend to jam against the far side of the pit and timbers may have been Discussion and conclusions \ 1000

Calanais Survey and Excavation, 1979-88

used to prevent this. Burl cites an experiment at Bougon near Niort in central France where greasy stakes were used to lubricate the heel of the monolith, and clay in the pit served to help hold the monolith in place (Burl 1993, 70). Once the toe of the monolith reached the bottom of the pit the monolith could be pivoted on its end. In Illus 24.17 the shear-legs method is depicted. In the reconstruction I have chosen to place the shear-legs on the east side of the pit because I think that that might allow better lift overall; and the closer the top of the shear-legs was to the top of the stone the better lateral control it would provide. The timber supports are not supposed to be the only timbers in use. The figures, in modern dress, are intended to provide an intuitive scale: the taller ones stand to about 1.8m (a bit under 6ft). Once the stone was upright the socket and the ramp were filled with stiff green clay and stones. A flat-topped mound of large stones and green clay was built round the base, wider on the east side where the pit was wider, presumably to provide additional support for the monolith. Ponting and Ponting (1984, 23) illustrate a variant of this method in which the top of the stone is raised on a latticework of logs (the ‘leverage platform’) to an angle of about 40 degrees to the horizontal before being pulled up the rest of the way. Many other methods have been suggested, including earthen ramps 24.8.3 The workforce required One problem with wooden shear-legs is that they would be weighty. The stone may have been erected without one. Leaving that thought on one side, Atkinson reported that erection of a stone weighing 26 tons would require a pull of 4.5 tons which could be produced by 180 men each exerting a pull of 56 lb. (c. 25 kg.). The monolith at Calanais weighed about 4.5 metric tonnes. Converting Atkinson’s figures accordingly, the monolith would have required a pull of around 800 kg. Ignoring the slight difference between a metric tonne and an imperial ton, Atkinson’s figures imply that the central monolith at Calanais may have required 32 people to pull it

Illus 24.17 Monolith erection upright. But it probably required far fewer. Farmers would be used to hard labour and could probably pull with considerably greater force for considerably longer than Atkinson’s figure suggests. They could probably have exerted a short term pull of about 50 kg (110 lbs) each. Burl suggests a comparable figure of 100 lb (c. 45 kilos) per man Discussion and conclusions \ 1001

Calanais Survey and Excavation, 1979-88

(1993, 70). Thus 16 strong people might have been enough, particularly if the stone was propped up between surges of effort. A few more people might have been needed to ensure that the stone did not topple sideways. This estimate is similar to that of 18 men for a 4 ton stone which Burl based on the erection of a stone at Down Tor on Dartmoor (Burl 1993, 70, 83). Thus around 20 workers along with their dependents could easily have erected Monolith 29. But social factors probably determined how many people actually did the work. Perhaps local beliefs required the efforts of as few people as possible. Perhaps they demanded the opposite and a whole community swarmed round the stone at an event designed to display the social prestige of some powerful or wishful person. 24.8.4 The area supplying the workforce Given the variety of social models which could be hypothesised it is not possible to translate this into a good estimate of the minimum size of the area from which people came to set up the stone. There is no direct evidence for population concentrations near Calanais like those discovered on Orkney at Barnhouse and Ness of Brodgar. Nevertheless, it seems quite likely that communities with 25 strong fit people between them could be found between say Garynahine in the south and Breasclete in the north, particularly if more farmland (including grazing) was available then for exploitation because of the lack of peat on higher ground and the somewhat lower sea levels of the time. Of course, in principle, the labour force could have come from a wider area on the west coast of Lewis, from all of the Western Isles or from a larger part of Scotland. Looked at another way, it might have taken only a few boatloads of people from somewhere else, perhaps people travelling along the western seaways. 24.8.5 The place of the monolith in the Calanais sequence The 1980 and 1981 excavations did not prove that the monolith was earlier than the Ring but the stratigraphy allowed that interpretation and

it is the preferred one because it would have been easier to erect the stone before the Ring had been built. The monolith post-dated the cultivation beds (covered by turf at some date between c. 3000 and 2500 cal BC but probably in the earliest part of that range) and probably pre-dated the Ring, which while not absolutely dated is interpreted as having been erected between about 2900 and 2750 BC. No pottery was found in layers associated with its erection and none was found in underlying layers. No charcoal samples were available for dating. Aubrey Burl thought that the monolith would have been the earliest stone at Calanais: â&#x20AC;&#x153;The orientation of the Callanish stone does not respect the long axis of the ring â&#x20AC;Ś its broader face looking out to sea as a landmarkâ&#x20AC;? (Burl 2000, 203). In this statement he seems to be treating the Ring as elliptical (see Illus 24.31 below). In fact some other reconstructions of the geometry of the Ring suggest that differences between the ring axis and the central monolith are trivial (Illus 24.28, 24.29). Burl compared the monolith to the apparently unaccompanied tall seaside stones of Clach Mhic Leoid and Borvemore in Harris and the 5.8 m tall Clack an Trustal (Clach an Trushal) stone 16 miles up the coast to the north-east (Burl 2000, 203). This argument, if it was that tall isolated monoliths are found not far from Calanais, has recently been weakened by the briefly reported discovery of remains of stone settings, including a stone ring and possibly an avenue or another ring on a flat platform next to Clach an Trushal (Richards and Wright 2006, 171). Nevertheless, that monolith may be earlier than the adjacent stone settings, and given the difficulty of raising the monolith at Calanais once the Ring was in place Burl was probably right that the monolith was the earliest orthostat (with the proviso that there is no direct evidence for the date of the orthostats composing the south row and the southern part of the Avenue). His idea that it was a marker for seamen is attractive, although it is worth bearing in mind that it is tucked well into the deep inlet of East Loch Roag. It may also have been further from the sea than today; a better understanding of the shape Discussion and conclusions \ 1002

Calanais Survey and Excavation, 1979-88

of the contemporaneous coastline would help assessment of this notion. In a letter to Margaret Ponting in 1982, Alexander Thom described how in 1933, seeking a quite anchorage, he navigated his yacht with care between rocky islets and promontories, going as far up the loch as his chart allowed him to go in safety. As he stowed sail he looked up, and there was Calanais â&#x20AC;Ś he had not realised how close he was (Ponting 1988, 423). Perhaps rather than aiding navigation from afar it marked journeyâ&#x20AC;&#x2122;s end for one stage of a voyage. Contrariwise the monolith may have presented a warning rather than a welcome. The preferred interpretation here is that it was primarily intended to mark the place of a planned greater structure including at least the monolith and Ring combined; and what led to its erection in this precise location was a formalisation of the placeâ&#x20AC;&#x2122;s prior significance within the landscape. 24.8.6 Comparanda The central stone in the nearby ring at Garynahine is tiny, only 0.6m tall, and thus comparable only at a broad conceptual level. Nevertheless some have seen it as analogous (Burl, 1976, 153), which leaves open the question whether the Garynahine example was a pale imitation of the Calanais monolith or a feature which was of greater symbolic importance than implied by its diminutive stature. Surviving central monoliths are fairly rare. There is one ring with a central monolith on the mainland of Scotland almost due east of Calanais, a group of four in SW Scotland, one in Berwickshire, two in inland N Ireland, at least twelve in SW Ireland, one in Anglesey, at least two in the middle Marches of Wales, four to six in Cornwall and three to five in Wessex (Burl 2000, 253 fig 30). Burl considered the stone settings in southwest Scotland as the nearest secure comparanda for the Calanais ring and central monolith (Burl, 1976, 205-8). Central monoliths may have been raised for a wide variety of reasons, and there need not have been any direct connection between the people who set up the monolith at Calanais and those who set up central monoliths in other parts of Britain and Ireland. Some surviving monoliths

may be the remains of compound structures. Avebury includes one of the most impressive British central monoliths in its south-south-east ring but it may be a relic of a more complex setting; three even larger central stones remain in the adjacent north-north-west ring (Smith, 1965, 223). Indeed the distribution of rings with central monoliths displays all the interpretational problems posed by the existence of regionality, as discussed in Chapter 24.1. 24.9 Phase 6b: The Ring 24.9.1 The archaeology of the stones The mound and pit at the base of Ring stone 42 were partially excavated in Areas B and H. The layers round the base of Ring stone 43 were examined in Area B. Some of those around Ring stone 47 on Area DV were uncovered and its pit was partially emptied. Because of an error in labelling when a find of pottery and charcoal was split up a piece of willow charcoal was long supposed to have come from fill 767 of the primary pit of Ring stone 42 (Ashmore 1999b; 230/81 AA-24969 4095+/-45); it was understood to provide a terminus post quem for erection of Stone 42. However recent re-examination of primary records shows that the charcoal came from upper fills of the chamber wall (Technical Note 12.8.5). The preferred date for erection of the Ring is now 2950 to 2850 BC because associated strata belong in pollen zone CaN-3ai (starting just after 3000 BC) and because it is similar to Stones of Stenness, where a Grooved Ware pot like the one from Calanais (ASH 61) was found. The strata round Ring stone 42 revealed far more information than those round the others. Illus 24.20 to 24.22 show the preferred interpretation of the sequence. The pit was cut into a cultivated soil on which turf line 766 had formed (Illus 24.20).The orthostat was erected and stones and green clay were used to fill the pit, perhaps with a higher ratio of clay to stone than in the illustrations. There may have been another episode of levelling of soil before the leftover clay spoil was spread out (Illus 24.21). A turf line 751 /758 formed. At Discussion and conclusions \ 1003

Calanais Survey and Excavation, 1979-88

Illus 24.18Â The Ring from the southwest photographed by D Henry of Historic Scotland in 1984, with stones numbered after Somerville

Illus 24.19Â The Ring from the east-north-east, photographed by M Brooks of Historic Scotland in 1980, with stones numbered after Somerville Discussion and conclusions \ 1004

Calanais Survey and Excavation, 1979-88

Illus 24.20 Erection of Stone 42 and emplacement of stones round it Illus 24.21 Turf growth

Illus 24.22 Latest phase of emplacement of stones round Stone 42 some later date boulders were placed around the Ring stone. They sank into the soft soil. The turf line continued to grow elsewhere (Illus 24.11). It looked as if some green clay had been brought in from elsewhere because clay and more stones were dumped on top of the basal stones (Illus 24.22). During our excavations the pit for Ring Stone 43 (on the north side of the entrance to the chambered cairn) was not excavated below the point where it cut the lowest turf line. Because of later disturbance any mound round its base did not survive. The original presence of a mound and subsequent disturbances when the chambered cairn was built, when it was robbed and when the secondary cairn was built may explain why the precise relationship of the pit to the turf line was slightly ambiguous - sufficiently so that it cannot

be regarded as providing independent information about the relationship of the second turf line to erection of the stone. The top of the pit dug for Ring stone 47 in DV at the south end of Area DI had been at around the level of the composite turf line 365 / 334, the lower part of which was elsewhere interpreted as pre-dating the central monolith and the Ring. Because of the truncation near Ring stone 47 it is impossible to say anything more precise. A pit edge was recorded close to Ring stone 47 but the layers immediately to its north appeared to lie in a depression and may have been the fills of a ramp dug for erection of the stone (see Chapter 9.6). If that is right, then the recorded pit edge nearer Ring stone 47 was actually the edge between ramp fill and pit fill. Alternatively, given that the edges of the pit for Ring Stone 42 were 0.5m from the stone and the edge of the depression was c. 0.5m from the edge of Stone 47 we may simply have misinterpreted the fills because we did not explore them sufficiently. Further excavation is required to see which interpretation, if any, is correct. In Area S (Chapter 14) trenches were excavated to a depth of c. 0.15m between Ring stones 50 to 53 and 41 to allow turf replacement. The archaeDiscussion and conclusions \ 1005

Calanais Survey and Excavation, 1979-88

ology was similar to that in Area H and in essence produced no significant evidence about the relative erection dates of the five Ring stones. That however is a neutral result given how restricted excavation was. Despite the complications described above, where clear evidence survived the stone ring was set up through a turf line (Illus 24.11; 164 on Area B, 766 on Area H) into which the central monolith was also set. The chambered cairn was subsequently built inside the Ring with its back rising up over the mound at the base of the monolith. On Area B turf line 164 (the lowest turf line on Area B) formed the old ground surface when the pit for Ring stone 42 was dug. A plain corky Neolithic sherd (Cat 176) was found in soil 871 below 164. The turf line itself contained an abraded non-corky Neolithic sherd (Cat 243), and three abraded sherds from two Hebridean Incised pots (Cat 313-4 and 317). Near Ring stone 42 green clay 872 which overlay a later turf line 162 was overlain by a thicker block of mottled green yellow lumpy clay (870) which formed a low mound southwest of the Ring stone. The same label was used for clay in this area at a lower level of the mound. One of these clays produced 2 conjoining Hebridean Incised sherds (Cat 315). Sheridan sees all the corky, non-corky and Hebridean Incised sherds found at Calanais as parts of a single pottery assemblage (Chapter 18 The Pottery Assemblage 18.5.5). Most reached Calanais in artefact-rich Phase 3 soils and clays imported during Phase 7. But the sherds described above were in earlier contexts than Phase 7 and probably got into the ground during pre-Ring agricultural activities. There is no evidence that the sherds were fresh by the time they reached the positions in which they were found so they are not much help in dating erection of the Ring or defining the material culture of those who did so. 24. 9.2 The Ring and its stones The plan of the Ring is not a true circle. Its east side is flattened slightly. It measures about 12.6m north-south from the centre of Ring stone 53 to that of Stone 47 and, less precisely, 11.6m

east-west from half way between Stones 49 and 50 to half way between Stones 43 and 44 at the mouth of the cairn passage (measurements from plan, Tait 1978).

Illus 24.23Â The Ring and nearby stones after Tait 1978 The stones vary so much in shape that it is impossible to define a typical one. Stone 43 flanking the north side of the later chambered cairn passage is tall and tapers to a point (Illus 24.19, 24.24). Stone 51 has an almost phallic appearance with a blocky top; along with Stones 44 and 48 it has a top larger than its base (Illus 24.19, 24.25). Stone 44 also has a distinctive shape from some angles; its top slopes up to a protuberance (Illus 24.19, 24.26). Stone 48 in the southwest of the Ring stands out as unusually short; and its upper part is notably broader than its bottom (Illus 24.19). Many of the stones have strong surface textures (Illus 24.19, 24.24-26). We did not record any evidence relating to preferential use of weathered or quarried sides for the interior faces of the stones. The stones are not regularly spaced, particularly Stone 52 which looks as if it was intercalated between its neighbours (Illus 24.23). If Stone 52 is regarded as secondary the original stone spacing, centre to centre in straight lines, would have been as follows.

Discussion and conclusions \ 1006

Calanais Survey and Excavation, 1979-88

Illus 24.24 Stone 43 from the west, photographed by D Henry in 1984

Illus 24.25 Stone 51 from the south, photographed by D Henry in 1984

Table 24.3 Distances between centres of Ring stones excluding Stone 52

Western circuit

metres

Western circuit

metres

Stone 44-43

2.4

Stone 41-53

2.6

Average

Stone 42-41 Stone 53-51 Stone 51-50 Stone 50-49 Stone 49-48 Stone 48-47 Stone 47-46 Stone 46-45

3.0 3.6 3.9 3.7 3.8 3.5 4.2 2.6

Stone 45-44 Stone 43-42

2.4 2.9 3.2

The average distance centre to centre is 3.2m. The number of measurements is so small that formal statistical tests cannot sensibly be applied to the idea that all of the average distances between centres were intended to be 3.2m and the observed differences were due to errors; but Table 24.4 suggests that this idea is wrong because a much higher proportion of small errors would be expected.

Discussion and conclusions \ 1007

Calanais Survey and Excavation, 1979-88

If gaps between stones are measured the average is 2.3m.

Illus 24.27 Stone spacing Illus 24.26 Stone 44 from the west, photographed by D Henry in 1984 Table 24.4 Error sizes and numbers of such errors if a regular centre to centre distance of 3.2m was intended Error size (m)

Number of such errors

1-1.9

0-0.09

0.2-0.29 0.3-0.39 0.4-0.49 0.5-0.59 0.6-0.69 0.7-0.79 0.8-0.89 0.9-0.99

0 2 2 1 1 3 0 2 1

Illus 24.28 Stone spacing on the hypothesis that there was an eastern face Discussion and conclusions \ 1008

Calanais Survey and Excavation, 1979-88

In Illus 24.27 the red lines demarcate a difference between stone spacing to east and west based solely on their values. The widest separation in the eastern group is shorter than the shortest separation in the western group. Visually, within the eastern group the four stones 42 to 45 are on a flatter curve and are more closely spaced than the others and the Ring might better be divided into a western circuit and a flattened ‘east ‘face’. Stones 42 and 45 are much longer and thinner in plan than the others, which supports this concept. Table 24.5 Distances between centres of Ring stones excluding Stone 52 and with a short 4-stone east face Western cirEastern ‘face’ cuit (Stones metres (Stones 42 to 42-41 and 45) 53-45

Stone 42-41 3

Stone 41-53 2.6

Stone 53-51 3.6

Stone 51-50 3.9

Stone 50-49 3.7

Stone 45-44

Stone 44-43

Stone 43-42

Average

Richards has suggested on the basis of ethnographic evidence that stones could have been added to a stone setting over many years as local magnates demonstrated their social standing by organising erection ceremonies (Richards 2005, 217, 224). Although nearly all of the Ring stones appear to form a coherent plan Stone 52 in its northern part looks as if it was added as an afterthought in the gap between Stones 51 and 53 (Illus 24.23). This and the existence of Stone 9 to the southwest and Stone 34 to the north-east of the Ring provide support for the idea of ad hoc erection of some stones.

metres 2.4

2.4

2.9

2.6

Stone 49-48 3.8 Stone 48-47 3.5 Stone 47-46 4.2 Stone 46-45 2.6 Average

3.4

The average centre to centre distance between the ‘face stones’ is three-quarters (0.76) that of the western circuit stones. 24.9.3 Geometry or a practised eye There are three main ideas about how that plan was conceived. One is that the stones were erected ad hoc over a fairly long period. Another is that the shape was laid out geometrically and the third is that it was laid out almost entirely in a single operation by eye and pacing-out.

Illus 24.29 Construction of a flattened circle of Type A fitted to the Glasgow plan (Curtis 1980, Fig. 2) The excavation did not entirely settle whether Ring-stones 42 and 43 were erected over a short or a long period. The similarity of their stratigraphy in relationship to the green clay and turf lines running between them suggests that they were probably built at much the same time as each other, but in testing such arguments the devil is in the detail, and that was certainly ambiguous enough to allow the possibility that several years elapsed between the times that Stone 42 and Stone 43 were set up. Nevertheless the preferred interpretation is that Discussion and conclusions \ 1009

Calanais Survey and Excavation, 1979-88

they (and by extension all of the Ring stones apart from Stone 52) were set up over a short period. The doyen of ideas about the geometry of stone rings, the late Professor Alexander Thom, thought that the Ring was set on a flattened circle of his Type A, based on a complex geometric method (Thom 1967, 122). Ron Curtis has confirmed that Thom’s suggested geometry works quite well, although the orientation of the axes in Curtis’ reconstruction is slightly different from that suggested by Thom (Illus 24.29; Curtis 1980, 31; Ponting & Ponting 1984, 49). Apart from the fact that the intentions and intellectual frameworks of the builders of Calanais and other rings are obscure, the main problem with accepting the Thom method is that there are severe doubts whether a megalithic unit of measurement was ever used (Ruggles 1999, 82-3). It is impossible to exclude the possibility that local measurement units were used at a few individual sites and it is impossible to exclude the possibility that individual sites were intended to have a precise shape (particularly given that the execution of most projects fails to match their designer’s aspirations). But studies of groups of various types of settings in various parts of Scotland designed to check Thom’s ideas do not demonstrate consistent use of any unit of measurement or of consistent precise geometrical figures. The subject is covered in great detail by Ruggles in his magisterial ‘Astronomy in Prehistoric Britain and Ireland’ (Ruggles 1999). I devised a simpler method in 1978-9. It involved creating a small preliminary circle of closeset stakes and then stretching a rope round its half diameter before drawing the rope out. That procedure produced a flattened circle with a cusp at the ‘entrance’ (Illus 24.30). As a matter of record, the radius of the small preliminary circle, at very nearly 2m, would have been almost exactly one of Professor Thom’s proposed units of measurement, the ‘megalithic rod’ (2.07m). However, none of the other Type A flattened circles on which I tested this construction produced a similar match between the radius of a preliminary circle and any of Thom’s suggested units of measurement. So it is very likely that the similarity at Calanais was a coincidence in megalithic rods as much as it demonstrably is in me-

tres, no more significant than the alignment of the West Row on modern National Grid east-west.

Illus 24.30 Another possible laying out process Discussion and conclusions \ 1010

Calanais Survey and Excavation, 1979-88

Illus 24.31 The Ring as an ellipse

Illus 24.32 Lack of correspondence between alignments of ellipse, avenue and line between stones 9 and 34

The prior existence of the central monolith would have been a problem in the positioning of Stones 49 and 50, both of which are fairly accurately on the laid out shape; that is an argument against use of my method. Another, shared with the Thom method, is that there are severe doubts whether precisely shaped ground-plans were intended by those who erected stone settings (Ruggles 1999, 82-3). Also, neither of the attempts to derive the Ring’s shape from geometry explains why the Central Monolith was not at the centre of the circle describing the western half of the Ring. Burl (2000, 204) regarded the Ring as a poorly laid out ellipse. The one depicted on Illus 24.31 has its major axis (in blue) at 20 degrees east of north and an eccentricity of 0.88. Visually it matches the stones quite well, indeed quite as well as the flattened circle produced by the Thom or Ashmore methods. Minor elliptical variants would also fit quite well. This solution also suffers from the problems outlined above, and Illus 24.31 shows that an elliptical fitting does not bring out any symmetry in the plan of the Ring, and is therefore a less attractive solution to those people believing in the use of sophisticated geometry for Ring construction; that of course has no demonstrable bearing on what its builders intended. In passing it should be noted that the avenue sides and centre line (main green line on Illus 12.32) were not aligned with the major axis of the ellipse. Nor was the latter parallel to the line between stones 9 and 34, also in green on Illus 24.32, which Somerville believed indicated a lunar alignment (Illus 24.32; Somerville 1912; see Chapter 3.13). During the excavations of 1980 and 1981 we looked without success for evidence which might support or conflict with my method or Professor Thom’s method. The preliminary circle used in my method does not correspond with any of the curving slots found under the cairn. The centre of the circle lay in the southern part of the subsequent chamber of the chambered cairn. I believe that this method was not used and suspect that the same is true of the Thom method. The exercises in Illus 24.27 to 24.32 highlight the perils of trying to deduce the intentions of the builders by imposing geometrically derived plans Discussion and conclusions \ 1011

Calanais Survey and Excavation, 1979-88

on stone settings. Structures like Calanais could easily be modified circles (and ellipses in general could easily be laid out by tying the ends of a rope to two stakes on the main axis. Large true circles like the Ring of Brodgar in Orkney, must have been laid out using a rope. Once it is accepted that the stones themselves are irregular, and laying-out errors must have occurred, any one of several geometrically based reconstructions is usually possible and the fact that some are slightly better than others is not very meaningful (see Ruggles 1999, 82 note 3 quoting Angell 1977 and 1978 for formally argued views on small variations from true circularity). Despite this the systematic variations in stone spacing make it credible that the flattening of the east side was intended rather than a result of errors. The outer edges of the stones of the western part of the Ring define a nearly true circle 13.0m in diameter, which could easily laid out with a rope despite the presence of the central monolith. That would not require sophisticated geometry. Stone 48 fits it least well and Stone 49 does not fit it perfectly (Illus 24.30) but in the preferred interpretation these divergences are errors rather than having any cryptic meaning. That said, the divergence of the east side from that circle is quite large (Illus 24.30). The preferred interpretation, which is neither strongly confirmed nor refuted by the measurements, is that the builders emphasised the eastern side of the Ring, using two unusually long stones to define the ends of an eastern ‘face’; those long stones and the stones forming an ‘entrance’ were purposefully set back from the position they would be if they lay on a true circle. Put another way, the east side was laid out by eye to create a façade, perhaps to define an entrance or exit facing sunrise on the days around the equinoxes. 24.9.4 Sourcing and erection of the stones The Ring stones were all smaller than the central monolith and would have been even easier to erect by the methods described for the latter (Chapter 24.8.2). Nor need they have come from far away. Traditionally local people thought they might have come from Na Dromannan, less than 2km away to the east-north-east ((Ponting and Ponting 1984,

22). More recently Colin Richards suggested that the presence of hornblende ‘eyes’ in the gneiss used for the ring stones showed that they were quarried from a single source, and he reported that gneiss with similar characteristics forms the outcrop to the west of the stone setting at Calanais. Although no specific signs of quarrying were found he had little doubt that the orthostats were quarried very locally (Richards 2006, 182). Assessment of this claim requires formal comparative studies between the Ring stones, the Calanais outcrop and other outcrops. Ron and Margaret Curtis report significant clumps of dark green hornblende crystals in the central monolith and four of the Ring stones, on one stone on the west side of the avenue and on the three southernmost stones of the east side of the avenue, while hollows on the central monolith and a few of the avenue stones may indicate where similar clumps had fallen out (Curtis and Curtis 2009, 30-31). 24.9.5 Orientation If the Ring was planned to face approximately east its orientation might reflect sunrise at the days round the equinoxes. Ruggles has cast doubt on the idea that the equinox was important to prehistoric people, going so far as to suggest that the word ‘equinox’ should be eliminated from the archaeoastronomer’s vocabulary (Ruggles 1999, 148-9). His main reason was that sunrise continues to vary along the horizon in the same direction as before; unlike at the summer and winter solstices there is no obvious ‘event’ to be observed. The ‘façade’ effect at Calanais is not visually strong enough to proclaim an orientation of the Ring on sunrise on the days around the equinoxes but there is some empirical data to support suggestions that sunrise viewed at one or both the equinoxes was important to the builders of Calanais. About 1.5 km to its east at Cnoc Sgeir na h-Uidhe there is a small erect stone (Illus 24.33). It is now under 0.5m high. Packing stones are visible around it. Photography undertaken by Ponting and Ponting from within the Ring at Calanais in 1977 demonstrated that the autumn equinoctial sunrise takes place over the stone and that the following Discussion and conclusions \ 1012

Calanais Survey and Excavation, 1979-88

day its northern edge was more than its own width away. Equinoctial sunrise and sunset always take place in the same positions so the sun would have risen there when the Ring was built (Curtis and Curtis 2000, 37). The stone does not show any signs of having once been taller, so it would have been visible only to those with keen eyesight. Of course it could have been supplemented by a more obvious temporary marker but without excavation there is no way to assess that idea. It seems that if the small stone at Cnoc Sgeir na h-Uidhe was intended as a foresight it would have been easy to use by those who already understood the reason for its positioning but far from obvious to anyone without foreknowledge. Similarly if an indication of the days around the equinoxes was intended by building an eastern façade that may have been understood by everyone knowing the purposes of the monument; but it may also have been designed to be difficult to perceive without prior knowledge. But the difficulties in retrieving the intentions of the builders (even if they were simple, which is a different question) seem insurmountable with present evidence. The preferred interpretation is mainly that the lay-out cannot be shown to involve any sophisticated geometry, but it does suggest an intention to include an eastward ‘facade’. Reasons for its eastward orientation may have included the use of the area east of the setting for rituals and ceremonies, or a preferred direction for approaching or leaving the Ring, or a mixture of those. On top of that the evidence adduced by Ponting and Ponting suggests that observation of the equinoctial sunrise was a matter of interest to at least some of its users. 24.9.6 Rings near Calanais and further afield There are remains of at least 5 stone rings near the main setting at Calanais. 1. Cnoc Ceann a Gharraidh (Calanais 2), about a kilometre to the ESE of Calanais on low lying land is a ring of 5 standing stones, perhaps originally c. 13 stones. Its long axis measures slightly less than 20m and its short axis c. 17m. The tallest stone is c. 3.25m in height and its shortest 1.75m (Ponting and Ponting 2000, 12.

2. Nearby to the north-east is Cnoc Fillibhir Beag (Calanais 3), a ring of 9 surviving standing stones, perhaps originally c. 17 stones, surrounding a setting of which 4 stones survive. It is a flattened circle slightly over 16m by 14m in plan (Ponting and Ponting 2000, 12-18). 3. Ceann Thulabhaig (Calanais 4; Ponting and Ponting 1981), or Ceann Hulavig (Ponting & Ponting 2000, 19) stands 3 km to the SSE of Calanais on a gently sloping hillside. It is an elliptical ring of 5 surviving tall stones (perhaps originally 6). Its long axis measures about 12.6m and its short one about 9.5m. The tallest stone is slightly less than 3m high. The ring surrounds a tiny central monolith which stands in a pile of stones, possibly a small central cairn. 4. About 2 km to the ENE of Calanais on the high ground at Na Dromannan is a ring originally of c. 17 ring-stones and 5 internal monoliths, along with a possible diminutive approach avenue. The western side was flattened with closely spaced stones, while stones were much more widely spaced in the other parts. None of the stones remains erect (RCAHMS 1928; Ponting and Ponting 1981; Richards 2006, 184). 5. Another ring, Cnoc Gearraidh Nighean Choinnich, buried in peat near Breasclete some 2.5 km from Calanais 1, has been reported by Curtis and Curtis (2003, 138; 2005, 149). It was elliptical measuring c. 48 by 41m. 5 fallen stones survive close by their estimated original positions. All of these sites lie less than 0.5 km from the present shoreline of East Loch Roag. Some pairs and triplets of stones in the surrounding area appear through the peat and may be parts of more rings or rows, although others may be part of field systems. Several apparently single monoliths survive, including those shown on Illus 24.33 (Ponting and Ponting 2000, 6). Curtis has shown that Thom’s methods of laying out stone settings work well to fairly well for Cnoc Ceann a Gharaidh (Calanais 2) if it was meant to be elliptical, for the outer ring of stones at Cnoc Fillibhir Beag (Calanais 3) if it was meant to be Discussion and conclusions \ 1013

Calanais Survey and Excavation, 1979-88

a flattened circle of Type A, for Ceann Hulavig (Calanais 4) if it was meant to be an ellipse and possibly for the fallen or unerected stones at Na Drommanan (Druim nam Eun; Calanais 10) if was meant to be a flattened circle of Type A. He has also shown, using the accurate Glasgow map of the area (Tait 1978), supplemented by his own surveys, that the axes of some settings point to others and that some lines between sites point fairly accurately north or east (Curtis 1980). So far as I am aware no statistical study to the standards advocated by Ruggles has been carried out to assess the significance of these results. It is not clear whether the local geographical clustering reflects mainly the original distribution of rings, or mainly survival. There are apparently unfinished or demolished rings like that at Na Drommanan in peat-covered areas elsewhere on

Lewis, including Achmore, and Druim Dubh (Ponting and Ponting 1981; Burl 2000, 427). Still others may await discovery as recent discoveries at Clach an Trushal demonstrates (Richards and Wright 2006, 171). Nor is it known whether there were also timber Rings on Lewis. Substantial pine tree stumps up to 0.5 m in diameter have been radiocarbon-dated to the 4th and early 3rd millennia cal BC (Dickson and Dickson 2000, 37-39, 67) so the possibility cannot be ignored. More broadly, Burlâ&#x20AC;&#x2122;s impressive synthesis of the stone circles of Britain, Ireland and Brittany demonstrates that there are several large areas with abundant stone rings: in Caithness, round Inverness, in Aberdeenshire, Perthshire, Arran, and Wigtownshire, in Fermanagh, Londonderry and Tyrone in N Ireland and Cork in Southern Ireland, and in Cumbria, the Peak District and

Illus 24.33Â Some of the rings, standing stones and cairns near Calanais, based largely on Curtis and Curtis 2000, 6 Discussion and conclusions \ 1014

Calanais Survey and Excavation, 1979-88

on Dartmoor in England (Burl 2000, 2). Timber rings currently have a generally lowland distribution and some of them will be discussed below in the section on ritualised landscapes (Chapter 24.15) Here discussion will be restricted to comparanda for the dominant visual characteristic of the Calanais Ring, the ratio of the maximum height of its stones (4.4m) to its plan dimensions (c 13 by 11.8m). The ratio of height to diameter for Calanais is 0.34 to 0.37 giving a strong subjective impression of â&#x20AC;&#x2DC;tallnessâ&#x20AC;&#x2122;. It shares this feature with several other rings. The nearby Ceann Thulabhaig partially shares it; the highest stone is 2.75m tall and the ratio of height to diameter is between 0.3 and 0.2. The two main comparanda are at Stones of Stenness in Orkney and on Machrie Moor on Arran. These comparisons are likely to be legitimate despite the sites being in regions with generally different material cultures because of the distinctive characteristics of the structures, the occurrence of very similar grooved ware pots at Stenness in Orkney and at Calanais, and generally similar grooved ware pots at Machrie, and the ease of travel by sea between Calanais and Machrie. Two of at least seven stone circles at the large ritual centre on Arran were built of very tall stones (Burl 1976a Machrie Moor 2 and 3, Barnatt 1989 4:17 and 4:18). One of these, Machrie Moor 2, originally had 7 or 8 tall sandstone orthostats. The largest surviving one is 5.3m tall. That ring was about 12.8m across, giving a ratio of height to diameter of c. 0.4. Only one stone of the other ring (Machrie Moor 3) still stands. This ring was probably not circular and it measured about 16.3 by 15.4m in plan (Burl 2000, 90-1); it too would have had an unusually high ratio of height to diameter. The Machrie Moor tall-stone rings are not scientifically dated. The Orkney complex in the heart of Neolithic Orkney included at least two stone rings, both larger in diameter than that at Calanais. Like Calanais they stand close to water. To their east is the freshwater Loch of Harray and to their west the now brackish Loch of Stenness. The stone ring at Stones of Stenness was near-circular. The surviving 6 of its original up to

12 stones with reliably original positions imply that a slightly elliptical shape may have been intended (Curtis in Ritchie 1976, 49). Its diameter is c. 30m, more than twice that of the greatest dimension of the Calanais ring. The tallest surviving stone is 5.7m high (Ritchie 1976, 9). The ratio of height to width is 0.19, considerably less than that of Machrie and Calanais. Correspondingly it is a less crowded site, but it produces a generally similar subjective impression. There may have been a third ring of tall stones with a very high height to diameter ratio where Maes Howe now stands (Richards 2005, 243-5). The Ring of Brodgar at 104m in diameter is not comparable in this way because its large area diminishes the dominance of the stones used for its construction. There are no scientific dates relating directly to the ring at Stones of Stenness. The oldest radiocarbon age OxA-16484 (4346+/-39 BP) is from bones in fills of the ditch, calibrating to between 3090 and 2890 cal BC (Sheridan 2000). OxA16484 does not form a satisfactory group with the ages OxA-16482, OxA-16483 and OxA-16485 for other pieces of bone from the basal fills (Chisquared = 9.62 against a rejection figure of anything greater than 7.8). It is marginally compatible with the next oldest age OxA-16485, but perhaps the sample should be regarded as having a different taphonomy from the other ditch samples. It is similar to an age for a burnt sheep long bone from the central hearth-shaped feature (OxA-18037 4305+/-35), calibrating to between 3020 and 2880 cal BC, (Sheridan & Higham 2007, 225). However, a cattle bone from the organic basal ditch fill 3 B 25, the stratigraphically earliest deposit producing a radiocarbon date, had an age (OxA-17783 4111+/-32) between 2870 and 2570 cal BC. That is significantly different from the dates from the hearth and from OxA-16484. Strictly speaking the hearth may have gone with a clay feature earlier than the Stenness Ring. But one credible interpretation of these results is that the central hearth and the Ring were, as Ritchie suggested, earlier than the henge. The hearth and by implication use of the Ring dated to between 3020 and 2880 while the basal organic fill of the ditch and by implication the ditch itself dated to between 2870 and 2570 cal BC. The earlier animal Discussion and conclusions \ 1015

Calanais Survey and Excavation, 1979-88

bone dates from the ditch, from stratigraphically higher fills in the original west terminal close to the extension which narrowed the entrance to the henge, will thus represent residual bones from the period of use of the Ring prior to digging of the ditch (Ritchie 1976 Fig 3). So the Ring was probably in use sometime between 3020 and 2880 BC. Grooved ware was found at Calanais, Machrie Moor and Stones of Stenness but not in direct association with Ring stones. Pot 16 from the west ditch terminal at Stenness (Ritchie 1976, 23-4 Fig 6) and Pot 17a from Machrie Moor, one of a group associated with the timber circles (Haggarty 1991, 65-6 Illus 6) are similar to Pot 61 at Calanais. The Stenness pot might have been older than the ditch fills in which it was found, like some of the animal bones, but if the dating arguments above are

accepted it was probably deposited at some time between 2870 and 2570 cal BC. That would be a credible date ranger for deposition of the grooved ware pot at Calanais. The latter came from shallow pit or scrape 877 and an adjacent patch of clay 866 in BV which are ascribed to a period after the Ring was set up but before the chambered cairn was built. As so often at Calanais, however, the problems associated with using turf lines and layers of green clay as stratigraphic markers leave open the (non-preferred) possibility that the pit was stratigraphically somewhat earlier or later. The similarity of the three Grooved Ware pots at the three sites strongly suggests some connection between the places at some time; but the absence of direct association of the cited pottery with the standing stones of the rings means that the

Illus 24.34Â Early slot 795 under the cairn on Area H (and the probably unrelated 913 on Area D) Discussion and conclusions \ 1016

Calanais Survey and Excavation, 1979-88

similarity of the three pots should not be used to bolster arguments that the building of the rings was strictly contemporaneous, particularly since the pot found at Calanais may have been left by visitors. However, the use at all three places of rings with very tall stones surrounding fairly small area does suggest that Calanais and Machrie, and probably Stenness, were built within at most a few generations of each other. Leaving to one side the specific pots cited above, Stones of Stenness was in use at some time between 3020 and 2880 cal BC and was probably set up within a century or so of 3000 BC by people who used Grooved Ware. The ramifications of this interpretation will be discussed after consideration of other structures at Calanais. 24.10 Phase 6c: Early slot 795 under the cairn 24.10.1 Description Slot 795 on Area H was later than the locally oldest turf line, 766, into which the stone-holes for the Ring were cut, and was covered by a thin soil below turf line 758 underlying the green clay platform. Although the slot was very shallow the gap between its two parts was real, not simply an artefact of later ground-working or of our excavation (Illus 24.34). Examination of its western part was hampered by cairn stones in the trench section which prevented its being cleaned back to the edge of the area. It seemed to run fairly straight but might have swung north or south. Excavations in Area S to the northwest were confined to superficial levels (to allow laying of 15cm thick new turf ) so the lack of a slot there does not prevent the possibility that slot 795 swung north and formed part of a structure to the north of the later chambered cairn. The absence of a continuation of the slot in Area BIWX is less easy to explain. It is perhaps attributable to the well-developed plough or spade-worked soil at a basal level there, but excavation did not reach clean subsoil and it may be simply that it remains unexcavated at a lower level than the worked soil. The slot may have been of much the same date

as the Ring, because it cut the turf line 766 into which Ring stone 42 was set and was sealed by turf line 751/758 which formed above upcast from the Ring-stone pit. However the slot may have been later than the preferred date for the Ring. It contained a piece of hazel charcoal which was dated to between 2910 and 2630 cal BC (AA-24970; 4205+/-45). However, the date provides only a terminus post quem of 2910 cal BC for the slot. 24.10.2 A feasible but unlikely interpretation Area DI slot 913 was covered by a turf line which may have been equivalent to turf line 758 which covered slot Area H 795 on Area H. It is possible to draw an egg-shaped enclosure including slot 795 on Area H and the lower remains of slot 913 on Area D (Illus 24.35).t The problem with this speculation is that the slots were cut through material interpreted as spread from the Ring stone pits but such a structure would not have fitted inside the Ring unless it had a flat east face. On current evidence it is very unlikely to reflect reality and the idea is included here only because it can be tested during any future excavations. 24.10.3 The preferred interpretation The preferred interpretation is that the two slots were different in date from each other. The turf line covering the slot on DI is interpreted as later in date than that covering the slot on Area H. It is only the fact that the turf lines important to interpretation were treacherous stratigraphic indicators that allows floating of the speculation shown on Illus 24.35. The preferred interpretation therefore is merely that slot 795 was a fragment of a structure not otherwise recorded during excavation, and not far different in date from the central monolith and Ring. 24.10.4 Deposition outside the Ring Along the south side of the ditch running from the passage entrance to the north of East Row Stone 30 were several patches of dark soil, ashy soil and burning. One of the earliest shallow pits or Discussion and conclusions \ 1017

Calanais Survey and Excavation, 1979-88

scrapes (877) in this area contained several sherds of a fine Grooved Ware pot (ASH 61 Illus 24.63). Its stratigraphy suggests that the sherds were deposited in a period after the Ring had been set up but before the cairn was built. Other sherds of this pot were found in a nearby shallow scrape 866, as if the pit had been disturbed by later activities. Altogether about 15% of the pot was found. Sheridan (Chapter 18 The pottery Assemblage 18.6.3) suggests that the pot was made locally because it has the same corky fabric as seen in some of the Neolithic pottery from Calanais. She dates it through comparisons with radiocarbon-dated pots to the first centuries of the 3rd millennium BC so its creation may have been contemporaneous with the erection of the Ring at Calanais. She sees the presence of Grooved Ware at Calanais as of considerable significance, as this type of pottery is absent from the north-west mainland of Scotland the only other Grooved Ware find in the Western Isles is a small, thin-walled fine bowl from the passage tomb of Unival, North Uist. This specific type of Grooved Ware is widely distributed, having been found from Orkney to Fife, in Co. Meath in Ireland and at various sites in southern England, including Woodlands, Wiltshire (see Chapter 18.6.3).

Illus 24.35 Speculative (non-preferred) reconstruction linking slots 795 and 713

24.11 Phase 7: The green clay platform, imported clay soils and the start of ritual epositiont 24.11.1 The green clay platform A light structure was built in the area subsequently occupied by the chambered cairn. It was set through the clays under the cairn, which on Areas H and B, and to a much lesser extent on Area D, had the appearance of a green clay platform. On Area H a layer of green clay 760 forming the platform was of irregular thickness. Similar but more weathered green clay 750 lay to the north of the cairn at the same or a very similar level. On Areas BIWX and BIVWX green clay was found in an identical stratigraphic relationship to the cairn; green clay was also found outside the cairn in a stratigraphically later position than Ring stones 42 and 43 on Area BIWX. Green

Illus 24.36 The features 866 and 867 in which Grooved Ware was found

Illus 24.37 Grooved Ware pot ASH 61 Discussion and conclusions \ 1018

Calanais Survey and Excavation, 1979-88

Illus 24.38Â The green clay platform (for detail of related slots see Illus 24.39) clay was found at the same or very similar stratigraphic position under the south-western part of the cairn, on Area D. Some of it (in dark green on Illus 24.38) may have been clay spread from the mound round the base of the central monolith rather than clay from the pits dug for the Ring stones. Silty green clay 906 underlying the locally uppermost turf line 905 on Area D is shown in pale grey green on Illus 24.38; it is interpreted as having been greatly disturbed by trampling but it may have been remnants of the platform. The differences between the pre-cairn layers on H and those on D were very marked. In particular, the polygonal drying cracks in the basal-stone-impressed green clay on Area H suggested very little trampling between the time the clay was laid down and the start of cairn-building, but the mixed layers on Area D suggested considerable

spreading of green, light green and grey clays and other poorly understood activities. It is hard to attribute these differences entirely to different destruction histories. 24.11.2 The origin of the green clay used for the platform under the cairn When the Ring stones were erected a surplus of clay would have been produced because of the volume of the buried parts of the stones and that of the packing boulders in the pits. Calculations are presented in Chapter 25: Technical note 24.11.2. The preferred interpretation is that this surplus turf, soil and clay from the stone pits provided the c 50mm depth of weathered clay between the lowest turf line and the middle turf line on Area H and, less clearly, similar turf lines on Areas B Discussion and conclusions \ 1019

Calanais Survey and Excavation, 1979-88

Illus 24.39Â Later slots and clay spread 747 under the cairn and D. That fits the stratigraphy adequately, and otherwise there is no immediate explanation for the presence of clay separating the two turf lines. In this preferred interpretation the subsequent green clay 760 of the platform was imported, presumably from nearby. The immediate impetus may have come from the water-logging suggested by sphagnum spores in the underlying soil 758 on Area H. But, given the importation of other soils and clay containing charcoal and potsherds during Phase 7, non-functional reasons may have been more significant to those who made the platform. Perhaps deposition of the clean green clay had some symbolic meaning to them, related to but different from that of the culturally-enriched imported material. As described in detail in 24.11.1 the ground in and at the east side of the Ring at the same stratigraphic level appears to have been built up

with a layer of clay H750, B872 and B161 shown on the stratigraphic diagram of turf and clay layers in Illus 24.11. 24.11.3 The later slots under the cairn On Area BIWX the green clay under the cairn was delimited by slot 858 with a dark charcoal-rich fill (Illus 24.39). On Area D slot 913 may have had the same function, although the stratigraphy was less clear. We may have conflated the plan of its lower part with the remains of an earlier cultivation trough, as discussed below; because of this ambiguity the upper part, although more amorphous and probably too straight, has been emphasised in illustrations 24.31 to 24.39. On Area H a spread of dark clay 738 lay under and outside the later kerb slab and brown loam Discussion and conclusions \ 1020

Calanais Survey and Excavation, 1979-88

747 inside the kerb was in the same stratigraphic position. A radiocarbon-dated piece of willow charcoal from clay 848 suggesting disturbance between 2040 and 1770 cal BC ( (sample 203/81; AA-24968 3575±45 BP). Clay 747 may have been composed of somewhat less disturbed fills of a bounding slot. Both contained E/MN potsherds. It is admittedly equally possible that they were the disturbed remains of a spread of Phase 7b imported clay like those found on Area D, particularly 738 which spread for 0.2m to the north of the kerb slab. Whatever their origin they may have been disturbed more than once, during cairn construction and during cairn robbing including removal of the early kerb. They need not have been disturbed when the later massive kerb slab was erected. Another set of slots may have defined an inner structure. On Area DI inner slot 915 probably originally ran up the side of the clay mound round the central monolith although the evidence was difficult to interpret. Between it and the area removed by erosion in the chamber were the last remains of an even slighter slot 920/941. Slot 883 in BIVWX immediately underlay the north passage wall (Illus 24.39). This slot was not fully excavated; its southern edge was obscured by numerous small clay patches or shallow pits. Table 24.6 Burnt bone from features in green clay platform in the passage Ctxt 878

Descr.

Clay patch abutting wall

Smpl Descr. 699

Burnt Bone Bone including minute fragments

881

Pit or scrape

694

882

Loam in 881

2022 0.11g

883

Dark greygreen slot fill

2011, 497, 623, 697, 698

0.44g; 1/6/81 and 6/6/81: cremated Bone including one 1 cc in volume

Cremated bone was found in four contexts in the passage; most came from dark grey/green gritty charcoal-rich slot fill 883 under the bottom course of stones of the north passage wall (Table 24.6). There were seven sherds in these features, Cat. nos. 245-7 (all E/MN corky), Cat. 283 (E/ MN Hebridean Incised), Cat. 735–736 (one either E/MN Non-corky or Beaker and the other probably fine Beaker) and Cat 841 (either E/MN Non-corky or Beaker). A piece of fuel ash slag 81.677 found in BIVWX 891 may have been from a hearth in a settlement somewhere near Calanais, and brought in along with soil containing sherds. Conceivably however it may have come from a cremation pyre. On Area H a more fully explored slot 773 with a structured fill had been partly cut away by the chamber (Illus 24.39). Its layered fills suggest careful back-filling. Its top fill to the east was dark grey peat-like material 772 spilling onto the green clay 760 under the body of the cairn. By the east baulk it was a brown loose fibrous clay loam. It was mostly absent in the west although it survived in the west baulk. It did not contain any artefacts. A charred hazel nut shell from it (sample 2038) produced a radiocarbon date (SUERC-11616 4430+/-35 BP) of 3330 to 2920 cal BC. The middle layer 778 was mainly humic rich clay containing charred hazel nut shells. A charred hazel nut shell from sample 2051 produced a radiocarbon date (SUERC-11617 4425+/-35) of 3330 to 2920 cal BC. A pollen sample (2051) was zoned to the middle of sub-zone 3a, the transition between CaN-3ai and CaN-3aii, dated at Leobag to a time between 2770 and 2360 cal BC. The lowest fill (730) was grey-brown with a slight green tinge, charcoal-rich sandy clay with a component of small sub-angular stones It contained 3 E/MN potsherds (Cat 162-5) and two possibly heavily burnt sherds probably from a fairly fine domestic Beaker (Cat 671, Cat 672). t also contained burnt bone (Table 24.7) and at least one quartz flake (CAT 146). A pollen sample (2048) was zoned to CaN-3ai, starting between 2980 and 2510 cal BC and ending between 2770 and 2360 cal BC).

Discussion and conclusions \ 1021

Calanais Survey and Excavation, 1979-88

Table 24.7 Burnt bone from fill 730 of slot 773 in the green clay platform in Area H Ctxt

Descr.

Smpl

Descr.

730

Slot fill 332, 352/81, 350, 2048; 32

22/5/81 Fragment Bone, Bone (burnt), 26/5/81 Cremated Bone, 0.02g

Table 24.8 Neolithic pottery from the slots and related features HII

H730

BIVWX

B878

BIVWX

B881

BIVWX

B885

Fill of slot 773

162_165

E/MN corky

Feature below passage north 67 wall

E/MN Corky

Dark green fill clay in feature 245, 246 & 247 below passage north wall Dark green fill below passage north wall, like fill of slot 735 883, the continuation of the slot on Area H

E/MN Non-corky ?E/MN Noncorky or dom Beaker

Table 24.9 Probable Beaker pottery from the filling of the slots and related features Area

Ctxt

BIVWX

885

HII

730

HII

730

Description and interpretation

Cat No Details

Dark green fill below passage north wall, like fill 736 of slot 883, the continuation of the slot on Area H

Grey-brown with slight green tinge, gritty sandy clay with small sub-angular stones; charcoal-rich, uncemented and non-greasy. The bottom fill of slot 773

Fill of slot 773

The bone, hazelnuts and sherds relate to closure rather than creation of the platform structures. Apart perhaps from the probable fine domestic Beaker sherds they probably reflected Phase 3 ancestral material imported from other sites. It is not clear, however, whether the soil and clay containing them were imported specifically to fill the features or came from imported clays and soils containing material of much the same early date spread out in Area D. None of the possible and probable Beaker sherds from slots 773 on Area H and 883 on BIVWX and contemporaneous features was unambiguously

Prob fine Beaker

671

Prob dom Beaker

672

Prob fine Beaker

from a Beaker. However there was also a small fine but worn conjoining pair of Beaker sherd in primary level 736 of the cairn overlying the green clay platform and the presence of Beaker sherds in the slots and related features is contextually credible. Slot 915 on Area D (Illus 24.39) included what seemed to be stone packing for the bases of postholes. It was not on the same line as the face of the chamber wall, nor was slot 913 on the line of the cairn kerb. Therefore the slots were not dug simply to mark out the subsequent cairn structure but represent an earlier, presumably light timber structure or structures.

Discussion and conclusions \ 1022

Calanais Survey and Excavation, 1979-88

24.11.4 Discussion of the later slots Slot 773 on Area H and slot 883 on Area BIVWX were parts of the same slot. Slot 915 on Area D is interpreted as helping them to form a bag-shaped timber structure (Illus 24.39 -24.41). Slot 915 on Area D, slot 858 on Area BIWX and the spread clay 747 on Area H are interpreted as forming another structural element (Illus 24.39-24.41). Illus 24.40 was produced by flipping a copy of slots in Area D, H and B around a horizontal (east-west) axis. The slots were then joined up in lighter colours and the interior of the hypothetical enclosure was coloured green. Slot 773 on H and slot 883 on Area B were in identical stratigraphic positions to one another and of the same date. The inner slot 915 on Area D was at the same stratigraphic level. The problem with it, as mentioned above, is that the southernmost slot on Area D was covered by an old turf line which was cut by the slot to its north. But turf lines are dynamic, not static and the sealing of the slot can be explained by re-growth of turf after the outer part of the light structure decayed or was removed. An alternative interpretation is that the slots defined successive rather than contemporaneous enclosures (Illus 24.41 1 and 2). It depends on believing the most undemanding interpretation of the stratigraphy, which is that the outer slot was abandoned and had been covered by turf by the time the inner enclosure was built. In either interpretation the traces of the slight slot 920/941 inside the bag-shaped enclosure may represent a contemporaneous feature or another phase of activity. There is a problem with both interpretations in that the fill of the deep slot 773 on Area H was very different from that of shallow slot 915 on Area D. Also clay 747 on Area H, interpreted as possibly a spread out fill of the equivalent of slot 913, contained several potsherds but slot 913 had none. However, the layers on Area D were much more disturbed, with only traces of the slots remaining, and the surviving fills are interpreted as corresponding only to the base of the bottom fills of the Area H slot and the slot presumed to be the source of clay 747.

There are no compelling reasons to prefer one interpretation over the other. Whichever of them is more correct the detailed histories of Areas D and H must have been very different once the timber structure or structures went out of use. Area D had very mixed deposits. In sharp contrast, the surface of the green clay platform in Area H was almost pristine, retaining polygonal drying cracks directly beneath the overlying basal cairn boulders. Because of this the filling of the northern slot of the bag-shaped structure must immediately predate the cairn, but the same cannot be shown to be true on Area D. Perhaps certain activities were restricted to the southern part of the green clay platform. Possibly the eastern the deposits beneath the cairn on Area D were heavily disturbed and partly removed during Phase 10, when the cairn was robbed while ritual deposits were made in the area south of it. During that period only the outermost part of the cairn on Area H was removed. This asymmetry may reflect beliefs in a quartering of the cosmos, discussed in 24.13 and 24.14. I strongly suspect that use of the area was considerably more long lasting and complicated than our (incomplete) excavations suggested. The inner bag-shaped structure may have been preceded by several earlier structures. Slot 913 is Area D, interpreted above as bounding the green clay platform, may have two phases and features 920 and 941 may belong to yet another (Illus 24.39). The interrupted slot 795 on Area H discussed in Chapter 24.10 represents yet another. Dating of this possible sequence is difficult. Several centuries passed between setting up of the Ring stones and building of the cairn. If a structure was built soon after the Ring was completed, presumably the typical local pottery would have been Hebridean Incised ware, although I do not wish to suggest that it was those people who conceived the idea of building the Ring without stimulus from others. It is possible that Grooved Ware had already been introduced to the area. In an earlier discussion (Ashmore 1996, 73) I said that the latest structure or structures were associated with Grooved Ware, a point picked up by Bradley (1998, 2). But all of the definite Grooved Ware (ASH 61) at Calanais had been re-deposited in a shallow pit or scrape outside the entrance to Discussion and conclusions \ 1023

Calanais Survey and Excavation, 1979-88

the structure and the association was not direct. At the other end of the sequence, it seems quite likely that Beakers were already being made (elsewhere along the western seaways, perhaps) when use of latest structure was formally ended by filling in the slots with clays and soils containing ancestral and more current potsherds. If the latest structure was an immediate predecessor to the chambered cairn it may have served for deposition of funerary remains in the inner enclosure or perhaps as a cult house, or both. The fragments of cremated bone in the slot on Areas H and B may turn out to support one of these interpretations if they can be dated and if it can be determined whether they came from animals or humans. I do not know of any precise parallels for the suggested light structure or structures under chambered cairns, nor in early stone circles. The central hearth at Stones of Stenness in Orkney may have been part of something similar, but the small size of the area excavated inhibits further discussion. The possibilities will be discussed below in the wider interpretative context of the chambered cairn itself.

Illus 24.40Â An interpretation of structure represented by late slots under the cairn

24.11.5 Imported clay soils In Area D some clay soil layers below the main plough soil and above well-defined turf lines 344 and 365 seem to have been brought from elsewhere. The most artefact-rich was clay soil 369 found near the edge of the cairn in the eastern part of DI. Other clay soil layers probably formed from dumps of exotic material during this phase include DI 352, 360 and 374 and the fills of a depression 746 in the northwest part of Area H. Area H 738, described and discussed above, may be another example. Soil 369 produced many potsherds and lithic pieces (Chapter 9 Area D 9.3.16). Four radiocarbon dates were obtained from single pieces of charcoal. These dates suggest inclusion of material of various ages in the centuries around 3000 BC. Loose black sandy clay 352 underlying 369 in the south-east corner of Area DI produced a similarly early radiocarbon age from a piece of hazel

Illus 24.41Â An alternative interpretation of the late slots as forming successive structures Discussion and conclusions \ 1024

Calanais Survey and Excavation, 1979-88

(SUERC-11611, 4450+/-35) suggesting a date between 3340 and 2930 cal BC, along with 8 potsherds exclusively of Neolithic type (Technical Note 24.4.4). Table 24. The radiocarbon dates from context 369 Code

Sample 2 sigma date

SUERC-11597 pine

SUERC-11599 alder

SUERC-11598 hazel

AA-24963

birch

3360 to 3030 cal BC

3350 to 3020 cal BC

3100 to 2910 cal BC

2880 to 2570 cal BC

Chocolate brown clay 377 may also have been imported. It contained one small thin abraded sherd probably from a fine Beaker. Potsherds from clay soil 374 were mostly from E/MN corky pots but included one from a fine Hebridean or noncorky pot and another most likely to be from a domestic Beaker (Technical Note 24.4.4). A plough soil 1213 on Area S produced numerous small sherds from a corky Neolithic pot and a radiocarbon age from a piece of hazel charcoal (SUERC-11618, 4450+/-35) suggesting a date between 3340 and 2930 cal BC. These clay soils included Neolithic material. But a few fine Beaker sherds were found in clays 369, 374, 377 and 385 (Chapter 18.7.6 and Technical Note 18.7.6). Along with the radiocarbon and stratigraphic data they suggest importation of clay soils from one or more ancestral settlement sites during the period between construction of the Ring and construction of the chambered cairn. Small loads of soil were presumably brought in in baskets or bags. They need not have been hugely different from the deposits of soil with fragmented domestic artefacts which, along with clay soil from destruction of the cairn, formed much of the plough soil in this area. 24.12 Phase 8: The chambered cairn Elements of the chamber wall or core cairn were found in Areas BIVWX, DI and H. Parts of the outer cairn were found in Areas BIWX, BIVWX,

DI and H but substantial parts of the original outer kerb survived only in Area DI, with stone impressions and a single surviving stone on Area H. A superseded radiocarbon dating model described in Chapter 23: Radiocarbon and detailed in Chapter 25: Technical Note 23.2.3 suggested that the cairn was built between c. 2900 and 2600 cal BC, and previous publications have included that date range. But as described above there were three probable Beaker sherds in contexts sealed by the cairn and, as described below, a Beaker sherd was incorporated in the original build of the cairn. It very probably dates to after 2500 BC. 24.12.1 The original shape of the cairn The lower part of the chamber wall and the outer kerb on Area D were built of laid slabs. Where the original parts survived they were well built (Illus 24.43). Impressions of similar kerb stones were found on Area H below the later massive kerbstone. The chamber wall was very well built on Area H (24.42). The south side of the passage had been destroyed and rebuilt in the Victorian period. In a previous publication (Ashmore 1995) I suggested that the chambered cairn might have been built in two stages, first the chamber wall and then the outer cairn. While that remains a possibility further study of the excavation records suggests that it is slight. Really the only evidence for it was a small pile of gritty material lying against the outer face of the chamber wall on Area H. But the green clay platform on Area H was the same under the chamber wall and the cairn. So was the trampling and other disturbance on Area D. Where it survived the outer kerb in Area D was built in a very similar fashion to the basal face of the chamber wall there. All in all, the chamber wall and outer cairn were almost certainly built as one operation. The simplified reconstruction shown in Illus 24.45 was produced by flipping the elements on Illus 24.42 about an east-west axis and filling in the gaps. The fallen north-east orthostat has been ‘reinstated’ somewhat north of the position of the false orthostat recorded by RCAHMS, making the chamber fairly symmetrical (Illus 24.45). Discussion and conclusions \ 1025

Calanais Survey and Excavation, 1979-88

Illus 24.42Â The plan of the chambered cairn The earliest version of the chamber had a round back end which rose over the mound at the base of the central monolith. Its wall was well built of large stones. The cairn material between the kerb and the chamber wall was much less carefully laid, but on Area H the surviving primary cairn was nearly as neat as the chamber wall. 24.12.2 The orthostats The orthostats need not have been part of the first build of the chamber. The southwest upright on Area D had glass fragments amongst the small stones at its base and more generally post-peat damage and repairs to the wall left the question moot. Sharbeauâ&#x20AC;&#x2122;s sketch (Chapter 3 Illus 3.35 and Chapter 4 Illus 4.7) does seem to show it was embedded slightly in the wall but less so than in the RCAHMS field drawing

of 1923 (Illus 24.46); the RCAHMS record may have reflected clumsy repairs. The other early records cited above are all very schematic but appear to show the east face of the stone completely free of the wall. All early sources show the east face of the north-west orthostat completely free of the wall. The recumbent stone in the chamber on Area H was the original north-east orthostat, not the one upright in 1980 (see Chapter 12.2.3; Illus 12.12). It was slimmer than the false orthostat, giving more of the appearance of a facing stone. If we were correct in interpreting a slight feature in the clay under the AD1857 and later chamber fills as a cast of the toe of the orthostat then the recumbent stone must at some stage have been in the angle of the chamber wall, not in the AD 1923 position recorded by RCAHMS. The south-east orthostat was also free of the wall. Discussion and conclusions \ 1026

Calanais Survey and Excavation, 1979-88

Illus 24.43Â The kerb and chamber from the south with the outer cairn in Area D largely removed [Film 1981.12.22]

Illus 24.44Â The later kerb slab, chamber wall and partially removed cairn in Area H with the chamber in the background, seen from the north [Film 1981.5.11] Discussion and conclusions \ 1027

Calanais Survey and Excavation, 1979-88

Illus 24.45Â Reconstruction of the plan of the original chambered cairn Although there is no compelling reason to suppose that the original chamber lacked orthostats the possibility thus remains open, as depicted on Illus 24.46. Some implications of this are discussed below. 24.12.3 Dating and finds from the primary cairn A radiocarbon date was obtained from a single piece of hazel charcoal from a reddened area 736 in the primary outer cairn body (sample 2243/81, AA-24967 4050+/-45). It calibrated to between 2860 and 2470 cal BC. An age from a piece of birch charcoal (sample 116/81; AA-24966; 4210+/-50) from the infill 732 of the cairn at a stratigraphically slightly lower level provided a significantly earlier date of between 2910 and 2630 cal BC. It should not have been deposited

earlier than the charcoal in context 736, so must have been residual. Details of the finds from the cairn and related contexts are tabulated in Appendix 12. They are summarised here. All of the pottery from general pre-cairn contexts (10 catalogue entries) was E/ MN corky or probably E/MN corky apart from Cat 48_53 which included a Heb Inc sherd and Cat 1032_1033 which might instead have been corky fine Beaker. The bottom fill of slot 730 cut into the green clay platform immediately under the cairn included alongside neolithic sherds two possibly thoroughly burnt sherds probably from a fine Beaker and one from a domestic Beaker (Cat 671, Cat 672; see Table 12.9). Clay patch 885 by the slot where it lay under the passage wall included a sherd probably from a domestic Beaker (Cat 736). Basal cairn layers included an E/MN Heb Inc (Cat 281) and an indeterminate one. Discussion and conclusions \ 1028

Calanais Survey and Excavation, 1979-88

Illus 24.46 Reconstruction of the plan without orthostats The primary chamber wall contexts had 8 E/ MN corky catalogue entries and one for an E/ MN Heb Incised sherd. Primary cairn layers included mostly E/MN sherds (10 catalogue entries), but they also included an E/MN Heb Inc sherd (Cat 174_175), E/MN or Beaker sherds (Cat 168_171 and 1036– 1038), probable Beaker sherds (Cat 541_542), and a much worn International AOC Beaker sherd or sherds (Cat 351_352). The later dilapidation of the cairn did not reach this level. Analysis of the plans of the layers of the cairn showed that the International AOC Beaker sherd pair probably did not make its way down into context 736 long after the cairn was built. Admittedly the sherd pair might have reached the position in which it was found after the cairn had been partially robbed if someone lifted two layers of large stones, lit a fire, placed the sherd where we found it (it showed no obvious signs of burn-

ing) and then carefully replaced the stones. But if that happened, one would expect a large amount of debris to have fallen into the hole when it was created. That no such debris was found is a fairly strong argument against the possibility. The Beaker sherd pair was in a similar abraded condition to many of the E/MN sherds found in context 736 and other parts of the primary cairn. The possibly and probably Beaker sherds from immediately pre-cairn and primary cairn contexts provide support for that idea. The maths is simple enough; all of the propositions described above have to be wrong if the cairn dated before Beakers arrived in the area So if the three main ‘probably Beaker’ identifications are each given a 1 out of 3 chance of being wrong and if each of those probabilities is independent of the others then they together define a 1 in 27 chance that the cairn was pre-Beaker. If on top of that the likelihood Discussion and conclusions \ 1029

Calanais Survey and Excavation, 1979-88

that the definite Beaker sherd was not deposited when the primary cairn was built is also taken to be 1 out of 3 (although my own less conservative assessment is more like 1 out of 5), then the total probability that the cairn pre-dates introduction of Beakers to the Calanais area is 1 in 81. Despite the risks in using subjective estimates of probabilities, and despite the difficulties in defining what is meant by independent assessments of pottery types, those odds are good enough to affirm that the cairn was built after 2500 BC. The secondary cairn is described and discussed along with erection of the Phase 13 massive kerb slabs in Areas B and H. 24.12.4 Chambered cairn studies, grand narratives and grave doubts In the second half of the 20th century Henshall demonstrated that whatever the origin of the practice of burying people in stone chambers most Scottish chambered cairns belong to broad regional types (Henshall 1963; Henshall 1972; Davidson & Henshall 1989; Davidson & Henshall1991; Henshall and Ritchie 1995; Henshall & Ritchie 2001). Although very few of them are adequately excavated or dated these broad groupings retain validity for some purposes; but on the one hand there is a lot of local variation and on the other hand some studies have long suggested small-scale social differentiation only broadly linked to distance, for instance Müller’s study described below (Müller 1988, 1). I do not wish to return in any detail to the mid20th century controversies about the origins and significance of the various types of chambered cairn, controversies cogently described by Stuart Piggott (1954, 124-129). They took place in the context of a belief that farmers first settled in Britain about 2000 BC and that the spread of chambered cairns and their use was over by about 1500 BC. In that mistaken time-scale the belief in an over-arching commonality in culture was entirely credible. The difficulty remains as it was over 50 years ago that in the areas of Scotland where chambered cairns survive there is (except in central Orkney) little good data about all those other aspects of

material culture which would allow us to test and challenge the significances attached to typologies (Piggott 1954, 126). Indeed it is conceivable that the chambered cairn at Calanais was built for reasons quite different from those which underlay earlier cairn building. Tilley (1998, 155) was right to complain that using identical terms for disparate things creates false links and that our terminologies constrain our thinking. I wholly agree with his comments about abuse of such labels (Ashmore 2003), but paradoxically the same thing can happen when ideas like Tilley’s are misapplied as when labels are misused. That said, and returning to Calanais, even if the chamber was used for burials, it could have been conceived primarily as a space for rituals associated with ancestors; or with initiation; for curating totems; for maintaining mystiques associated with power-games and social elites; for bonding; or for some of these and other purposes. Or it could have been a trophy house for the bones of enemies; that might give a different significance to the inclusion of ‘ancestral’ material in the pre-chamber slot. If the Ring was conceived as separating a spiritual place from the outside world the slot and the chamber might have been seen as providing a secondary defence against the spirits of old enemies whose bones had been put in the chamber rather than protection against restless ancestral spirits. 24.12.5 Comparanda In 1954 Piggott compared the Calanais chamber’s orthostats with those in the Camster type of chambered cairn (Piggott 1954, 225). Audrey Henshall’s contributions in her magisterial survey of the Scottish cairns and material related to them were described in Chapter 3: Previous Studies (see also Technical Note 24.12.8). Here it is worth recalling her comment that in a Hebridean context it was odd that the whole of the interior of the Ring was not used for the cairn, for if it had been the Ring would have served as a peristalith like those round Hebridean style cairns (Henshall 1972, 138). She saw the little chambered cairn as unique within the Western Isles, reminiscent of the cairns of Caithness and Orkney (Henshall 1972, 125, Discussion and conclusions \ 1030

Calanais Survey and Excavation, 1979-88

138-9, 206). Also the eastward orientation of the Calanais chambered cairn is characteristic of Camster cairns (Davidson and Henshall 1991, 79-80) but of few other sets of large monuments of the 4th or 3rd millennia (Ruggles 1999, 130). it is indeed interesting, even possibly significant, that the Camster cairns in general tend to line up around due east (Davidson and Henshall 1991, 79-80; Ruggles 1999, 130, 148-9). The chambered cairn at Calanais is far younger than all but one of the few scientifically dated Camster-style cairns. The exception is the second phase of Embo in Sutherland, discussed below. Megaw and Simpson, in their survey of British Prehistory, simply assigned Calanais to a broad group of passage graves of the period 3000 to 2000 cal BC within a model of regional development of styles (1979, 78,135). In his Magister thesis for the University of Freiburg, published in 1988, Johannes Müller used the information provided by Henshall to analyse the chambered cairns of Shetland, Orkney and the Western Isles (Műller 1988). He applied cluster analysis to try to understand the similarities and differences between the cairns. He considered how information might have been passed between the people who built them, and the reasons why they were built in their particular places. He showed that in at least one way the Calanais cairn is typical of much of the Western Isles. All of the chambered cairns in the long island are dispersed, with each large cairn or pair of cairns in an area defined by natural boundaries (Műller 1988, 26). Calanais too is isolated from other chambered cairns with the exception of the possible cairn at Breasclete (Henshall 1972, 460). One of Müller’s cluster analysis techniques suggested that in the Western Isles the Calanais chambered cairn is most like those at Loch Glen na Feannag (his Ust 23) and Craonaval (Ust 13) (Müller 1988, 20). Another analytical technique, combining evidence from three different clustering methods, showed it as an outlier closest to Unival (Ust 34), Garrabost on Lewis (L 6) and Sig More (Ust 27) (ibid 19, Fig. 10). But despite these local comparisons he saw the cairn at Calanais as exceptional in the context of other cairns of the Western Isles (Muller 1988, 26). More broadly he

saw a similarity between Calanais and Pettigarth’s Field (Zet 32) in Shetland, Cubbie Roo’s Burden (O 11) on Rousay and, in my view inexplicably, with the cairn (O 15) at Eday Church in Orkney (ibid, 30 Fig 17). He suggested that cruciform chambers like that at Calanais provided evidence for a direct connection between the Western Isles, Orkney and the Boyne. He argued that such chambers were to be found in the main ritual focal areas at the Boyne, Calanais, Brodgar (Maes Howe) in Orkney and Gruting Voe in Shetland (Műller 1988, 34-6). Had he broadened his analysis he might have seen some of the chambers at Camster in Caithness as cruciform (Henshall 1963, 96, 103); they are no less so than the cairn at Calanais. Many of them (and I would argue also those in the Gruting Voe area in Shetland) are not in obvious ritual foci. But the comparisons cited by Piggott, Henshall and Műller may not be valid for the original chamber. It had a round back end and the possibility that the orthostats were added in a subsequent phase cannot be excluded. Without the orthostats and square back end the similarity of the chamber to those of Barpa Langass and Marrog in the Uists increases, and that to the chambers in the cairns of Orkney and Caithness decreases. It may have been as similar to other small simple passage cairns in the Western Isles as to any regionally specific type of chambered cairn elsewhere. It could be compared to Achaidh near the Dornoch Firth which contained both inhumations and cremations (Henshall & Ritchie1995, 101-2). But in truth the chambered cairn at Calanais is far from identical to any these comparanda. Undoubtedly some small chambered cairns remain to be discovered in the Western Isles. For instance, one may be the 7m by 8m cairn at Stacaiscal, Barvas with small northern and western chambers compared by its recorder to cairns ‘such as’ Calanais (Campbell 2010, 11). So the Calanais cairn may not be as unusual in the Western Isles as it currently seems. That said, for the moment its final form appears to be exotic. Burl has suggested that it was added to the Ring as a physical retaliation against alien ideas (Burl 2000, 40). He detected a fairly common phenomenon of inserting ‘native burial places’ Discussion and conclusions \ 1031

Calanais Survey and Excavation, 1979-88

into ‘outlandish shrines’ (ibid, 132). Although the cairn as it was understood before excavation was of a form alien to the Hebrides, if it was originally bag-shaped and lacked orthostats it conforms better to his idea. But other interpretations may be preferred. The inclusion of cairns or cists within stone circles, or the erection of stones round cairns, is a common feature in Britain and Ireland. At one extreme is New Grange (O’Kelly, 1971), at another the Clava cairns (Henshall, 1963, 21-7; 1972, 271; Bradley 2000; Bradley 2010), at another the small encircled cairns of Ireland and western Britain. Conversely the huge peristaliths surrounding some of the round-cairned chambered cairns of the southern Hebrides (Henshall, 1972, 127-8), particularly those where the peristalith encloses more than the cairn (ibid), may bear some relationship to stone circles. It has been suggested, somewhat speculatively, that the practices which resulted in the combination of burial monuments and rings of large stones may have been the result of an extremely long-lasting belief that large upright stones were good symbols for people (Bradley 2010). If so it may be very rash to assume that

the conjunction of stone rings and burial places was specific to any particular local belief-set or had a narrow chronological significance. In the next section this idea is taken forward. 24.12.6 Dating small chambered cairns The radiocarbon dates available in 2006 for burials in chambered cairns in Scotland are listed in Technical Note 24.12.9 and summarised in Illus 24.47. The immediate stimuli for building the chambered cairns came from different sources judging by the various styles of chambered cairn in different parts of Scotland. Burials in a similar manner to that used for primary burials continued in some areas until the middle of the 3rd millennium, and secondary burials in different styles were subsequently added. But it is instructive to consider the sequence of styles of burial in other parts of European similarly distant from the areas where chambered tombs were first built. For instance in Western Switzerland the first (4800 to 4300 BC) funerary ritual involved cemeteries containing mainly individual graves. A second phase (4300 to 3300 BC) was marked

Illus 24.47 Radiocarbon dates for burials in chambered cairns in Scotland Discussion and conclusions \ 1032

Calanais Survey and Excavation, 1979-88

by the appearance of double or multiple, simultaneous or successive collective burials in cists. Megalithic tombs came into use only afterwards; subsequently Beakers were deposited in them with burials (Desideri & Besse 2010, 158-9). I do not for a moment wish to suggest that the Scottish sequence followed an identical path; my point is solely that it may have been equally complicated. There are indeed suggestions that the sequence in Scotland was complex. Probably some small chambered cairns in Atlantic Scotland did not belong to the main floruit. As part of their analysis of the introduction of farming, pottery making and other practices to Britain and Ireland, Sheridan (2009) and Pailler and Sheridan (2009, 34-7) have shown that simple passage tombs and small closed megalithic burial chambers like that in the cairn at Achnachreebeag may have been constructed in the west of Scotland (and in Ireland) before 4000 BC (Pailler and Sheridan 2009). There was a period of chambered cairn building in the area round Inverness significantly after the construction of main floruit chambered cairns ceased. Jones and Thomas (2010, 293) have speculated that it may have been much more widespread than previously supposed. They have suggested that the entrance graves of the western tip of Cornwall, a few of the miscellaneous small graves in the Isles of Scilly, those forming the group of wedge tombs near Waterford in Ireland and possibly some of the Bargrennan-style tombs of southwest Scotland may reflect the activities of travellers at a time when exotica such as gold lunulae, Beakers and copper alloy objects were also appearing. They suggested that the idea of building small chambered cairns was taken up afresh amongst diverse communities who responded in local idioms. The surviving evidence suggests that the burial ritual involved cremation although most of the cairns they discussed were built in areas of acid soils (which break down raw bone somewhat faster than well-cremated bone) and they considered that the lack of inhumed remains might be due merely to unsuitable preservation conditions. Jones and Thomas also emphasised that all of the scientific and artefactual evidence for use of these cairns could have arisen from secondary activities. But their preferred interpretation was that some

small tombs were established in coastal areas in the late 3rd and the 2nd millennium BC ( Jones & Thomas 2010, 292). Their model has more of the status of a speculation than a theory, because of the acknowledged lack of dating evidence for the primary construction and use of the cairns, but Calanais would sit fairly well within the framework of a resurgent interest amongst those familiar with the use of Beakers in building small chambered cairns and depositing cremated human bone in them. They also pointed to the late dates for some Clava cairns and viewed them as part of the same phenomenon ( Jones and Thomas 2010 292). Radiocarbon dates from hazel charcoal on the old land surface under the cairn at Balnuaran of Clava, in a valley debouching on the inner part of the Moray Firth, imply a date for construction after 1900 cal BC (Technical Note 24.12.9b; Bradley 2000, 115-6; AA-24234, AA-25433). As described in Chapter 21.1.1 above, Bradley had already hypothesised that they may have been built because of a perceived lack of â&#x20AC;&#x2DC;ancestralâ&#x20AC;&#x2122; monuments. If Jones and Thomas are right there may have been more to it than that, and the evidence from Culbin Sands cited by Bradley fits as well with their hypothesis as it does with Bradleyâ&#x20AC;&#x2122;s. But it is necessary to emphasise that, as freely admitted by Jones and Thomas, their model was based largely on inference, that dangerously attractive methodology, and requires the test of more excavation and more scientific dating. Even if the ideas of Jones and Thomas are not correct, interesting comparisons and contrasts can be made with the south chamber and passage in a small oval chambered cairn at Embo near the east coast of Sutherland. The cairn itself was built and used for burials much earlier, as described below. But the south chamber (at least) appears to have been built or greatly modified at some date after 2320 cal BC, probably by Beaker-users. Neither of the chambers was close in form to the one at Calanais (Illus 24.48). Even if the Embo antechamber be taken as equivalent to that part of the Calanais passage within the chamber wall the main differences between the two tombs were substantial: the Calanais chamber was faced by a wall and the chambers at Embo by orthostats up to 1.2m tall, although Discussion and conclusions \ 1033

Calanais Survey and Excavation, 1979-88

the passage was wall-built (ibid 12, Fig 4). The longer dimensions of the chambers at Embo were at right-angles to the passages. But there were some similarities. The south chamber at Embo was about the same size as the reconstructed original chamber at Calanais; its ‘antechamber’ was about 1m long and its passage was 1.6m long (Henshall and Ritchie 1995, 136), dimensions not far different from those at Calanais. Four radiocarbon dates came from contexts supposed to be contemporary with its construction. An age from two adult vertebrae found behind a stone under a corbel stone (GrA-772; 3720+/-70 BP) calibrated to between 2320 and 1900 cal BC. Small animal bones from a deposit contemporary with construction of the chamber provided an age (BM-442 3870+-100) calibrating to between 2580 and 1990 cal BC. On the other hand, infant bones from under corbel stones produced two ages (GrA-770 and 771, 4010+/-70 and 4340+/-70) calibrating to 2865 to 2310 and 3305 to 2700 cal BC (Henshall and Ritchie 1995, 75). Four more technically reliable dates were recently obtained from single human bones in the north and south chambers (UB-6876 to 6879). Two from fills of the south chamber calibrated to between 3520 and 3350 cal BC; and an age from another bone from fills of the south chamber calibrated to between 3330 and 2920 cal BC. But one from the floor of the much damaged north chamber calibrated to between 2620 and 2470 cal BC. (Sheridan 2006, 205). Parts of an International Beaker were found redeposited in a secondary cist at Embo and an insular Beaker came from the south chamber. The question must be whether the latter was built in the 4th millennium or in the second half of the 3rd millennium when skeletal material and Beaker pottery were deposited. The most straightforward interpretation of the reported stratigraphy is that the chamber should be dated to between 2320 and 1900 cal BC by the latest radiocarbon-dated bones in constructional contexts. In that interpretation the 4th millennium bones will have been ancestral material from the original cairn. Opposition to this interpretation can be based on the propositions that the late 3rd millennium bones were simply tucked behind stones in an al-

Illus 24.48 Calanais and Embo, the latter simplified from Henshall and Wallace 1963, 10, Fig 2

Illus 24.49 Early Beakers found at Calanais Discussion and conclusions \ 1034

Calanais Survey and Excavation, 1979-88

ready ruinous construction and that all but the four dates obtained recently are unreliable. Such reservations cannot be dismissed. But if the reported excavation results are accepted, the least complicated explanation is that re-use of a pre-existing chambered cairn involved construction or major reconstruction of a chamber and passage very broadly similar to those at Calanais. Thus the south chamber and passage at Embo bore a similar symbolic relationship to the cairn to that between the chamber and passage at Calanais and the Ring. In this interpretation Bradleyâ&#x20AC;&#x2122;s suggestion that the Clava cairns remedied a perceived lack in ancestral sites can be extended in other areas to include the appropriation of ancient tombs either by people influenced by incomers or by incomers seeking to demonstrate a spiritual relationship to the ancestors of local people. More generally the speculation set out by Jones and Thomas (2010) provides a broad context for construction of new small chambered cairns. 24.12.7 Early Beakers and the cairn at Calanais Refining the possible date range for building of the chambered cairn at Calanais requires an examination of the likely earliest date for Beakers. Sheridan (Chapter 18 and as described below) has set out the solid evidence from direct dating of human bones found with British Beakers that burials started in the third quarter of the third millennium BC. But if the western sea ways continued to serve as a conduit for new ideas there may have been even earlier Beakers in domestic and funerary contexts on the shores of the Long Island. The matter has even more direct relevance given the presence of three very early Beakers at Calanais. Most of the sherds from these Beakers came from the area to the southeast of the entrance to the Ring, pre-cairn enclosures and chambered cairn (Chapter 18 Illus 18.16). One sherd of ASH 38 came from a primary layer (736) of the chambered cairn on Area H. There was a scatter of sherds in Area B to the east of the Ring. A sherd of a fourth early Beaker (ASH 40) was found to the west in a thin soil outside the Ring, far from the other early sherds.

It seems on balance most likely that ASH 37 and ASH 38 at least came from burials (which remain undiscovered) at Calanais. Those burials may have been in a single grave or graves, or even in the chamber of the cairn. ASH 38, however, cannot have been deposited in the chamber of the cairn; it may have been brought to Calanais in soil and clay used to bed the stones of the chambered cairn or accompanied a burial in a grave contemporaneous with one of the pre-cairn enclosures. Sherds of similar Beakers have been found in many funerary and domestic contexts in the west of Scotland. Sheridan (Chapter 18 The Pottery Assemblage 18.7.4) finds it likely that the Calanais international-style Beakers date to the third quarter of the third millennium as early as the 25th century BC. Her scholarly exposition is an important complement - possibly even a salutary correction - to the somewhat more speculative discussion here. There are various models for the dating, origins and spreads of Beakers in Europe. They may have first been made in northern Iberia during the first half of the 3rd millennium BC (Desideri and Besse 2010 quoting Giulaine 1998). But Needham (2005, 176) preferred a more cautious assignation to Iberia or south-western France and a more specific origin for Maritime Beakers (a category to which the Calanais Beaker ASH 39 belongs) in the Tagus estuary in Portugal. In discussing the Beakers of southern France Lemercier (2012, 133, 138, 140) suggested that their likeliest origin was in Spain and Portugal. Lemercier dated the appearance of Beakers in his study area to between 2500 and 2400 BC. One long favoured idea is that early western Beakers were disseminated to northern Europe where Beaker-making potters were influenced by other contemporary styles of vessel, notably Corded Wares (Needham 2005, 176-182). Beakers in new styles were then carried back westward, influencing local artisans. But probably things were far more complex in the centuries after metal tools and ornaments reached Atlantic Europe, allied as the spread of Beakers seems to have been initially with flint trading routes (Needham 2005, 177) and subsequently with metalworking skills (Needham 2005, 177). Discussion and conclusions \ 1035

Calanais Survey and Excavation, 1979-88

In some areas of Europe Beaker-users settled amongst local indigenous communities and were involved changes in both population and ideas (Desideri & Besse 2010, 158, 171). That seems to be a generally accepted model for what happened around the middle of the 3rd millennium in parts of southern Europe; the first Beaker-users were members of a sprinkling of incomers bringing intriguing novelties into local cultures and the consequence was a fusion of the practices of indigenes and settlers (in some cases perhaps involving resuscitation of earlier indigenous ideas). That model may also apply in Britain, where it is probably wrong to regard burials of Continental incomers as exceptional; Sheridan (Chapter 18: The Pottery Assemblage) points out that Beakers of AOC and E type are particularly widely distributed over Britain, and most of the earliest dated examples are from graves of non-local individuals, as revealed either by strontium isotope analysis of their tooth enamel or by the Continental style of their graves. She speculates that the users of the early Beakers at Calanais might have been Continental immigrants, or were first or second generation descendants of incomers. It seems quite likely, taking into account both Sheridanâ&#x20AC;&#x2122;s points and the hypothesis proposed by Jones and Thomas, that the first phase of contact between Beaker users and the mid 3rd millennium inhabitants of the Western Isles resulted from the activities of an early trickle of Beaker-using people, possibly traders or prospectors, perhaps fascinated by whatever the great monuments of the Boyne valley and Orkney (and by extension Calanais) represented. It is even possible that the first landfall of Beaker users in the northern parts of the Western Isles took place before 2500 BC; for given the evidence for the diverse ways in which Beaker users interacted with the people of various parts of Europe, I see no reason why Beakers should not have been put to domestic or ritual use before single burial with Beakers became common enough to show in the British archaeological record. Indeed one of the earliest British burials, on the island of Coll, was cut through layers containing domestic Beakers (Ritchie & Crawford 1978). But it must be emphasised that the reliable radiocarbon dates currently available for burials

associated with Beakers in Scotland all fall significantly after 2500 cal BC and that is currently the most reasonable absolute early limit on the date range of Beakers in northern Britain (Sheridan Chapter 18 The Pottery Assemblage 18.7.4). It thus currently provides the best absolute terminus post quem for the chambered cairn at Calanais; a date half a century later provides a more likely limit. Because of the considerable similarity of early Beakers along the coasts of Europe it is not ridiculous to draw broad comparisons between their appearance in NW Scotland and southern France, where all the earliest Beakers were fine and in Maritime or International style. Settlements were small, including 1 to 4 oval houses. The burial tradition was generally collective burial in megalithic structures or local caves or cavities (Lemercier 2012, 133-135). But it is clear that the other artefacts associated with Beaker users reflected local and regional indigenous traditions, varying from west to east. It is possibly no coincidence that the description given above fits the (scanty) evidence from Lewis and Harris. But what specifically stimulated the building of the chambered cairn (and, indeed the earlier timber structure)? Could it have been a period of extreme hardship induced by some drastic change in the environment? Looking at a very different source of evidence, tree ring studies suggest a period of widespread stress on trees starting in 2354 BC and culminating in 2345 BC (Baillie 1999, 146). It is conceivable that this coincided with either the transition from Pollen Zone CaN-3ai to 3aii or to that from CaN-3a to 3b. After each transition there was increasing reliance on pastoralism, perhaps reflecting crop failures during 9 years of devastating weather. If the coincidence was with the earlier transition, it may be that the latest timber structure or the cairn was built in response. However, for the moment, any relationship must remain highly speculative. Baillie related this environmental event to an eruption of Hekla in Iceland (ibid 146, 205) and if further excavation takes place at Calanais it would be interesting to study clay and soil samples for particles of volcanic glass. Discussion and conclusions \ 1036

Calanais Survey and Excavation, 1979-88

Illus 24.50Â The Stage 1 enclosure stake-holes with the area of the spread Stage 2 bank and entrance silts in light colours Summing up, the chambered cairn at Calanais very probably dated to after 2500 cal BC, its architecture was similar to that of some other Scottish chambered cairns used for burials, and fragments of cremated human bone were found in its chamber in AD 1857. Although a possibility remains that it was built with very different aims it was most probably primarily intended for curating the remains of dead people. More speculatively it can be proposed that it was one realisation of a more general phenomenon: new building of burial chambers and passages stimulated by an infusion of Beaker-using people from south west Europe who in an initial phase of settlement either took up local traditions, causing a new vitality to their expression, or used them to affirm their own ancestorsâ&#x20AC;&#x2122; relationship to the ancestors of local people.

24.13 Phase 9a and 9b: The Stage 1 and Stage 2 enclosures 24.13.1 The Stage 1 enclosure There were four main stages of enclosure. The earliest was defined by stakes (Phase 9a), the second by an earthen bank with an entrance to the east (Phase 9b). It may have had more than one substage. It was succeeded by two successive wallbased enclosures (Phase 11a and 12a). At a low level in the north of the Stage 2 enclosure was a greenish-brown layer, 836 lower. In it were numerous small round features (Illus 24.50). They were interpreted as stake holes. Five similar stake holes were found in the south of the enclosure under patches of clay forming later floor levels. Discussion and conclusions \ 1037

Calanais Survey and Excavation, 1979-88

The stake-holes provide evidence for a separate enclosure stage, Stage 1. They were all shallow. They were invisible in layers at only a small absolute height above them. In Illus 24.50 iron-panned stake-holes have been omitted. They are included on Illus 24.52 Some may have gone with the Stage 1 enclosure.

possibly the stakes left very shallow impressions because they were merely set on the levelled down surface as part of a frame for a turf wall, rather than being hammered deep into the ground.

Illus 24.51Â The scooped western part of the enclosure [Film 1981.15.5]. They lay inside a scooped cut through layers of greenish clay and turf lines east of the Ring area (Illus 24.51). A Food Vessel sherd (Cat 567) may have been in upper soil 836 (see Technical Note 7.6.8 for problems with context assignation of this sherd). If so it must have been introduced by ground-working. 24.13.2 Discussion of the Stage 1 enclosures The stake holes formed a pattern asymmetrical to the Stage 2 enclosure (Illus 24.50) and highly asymmetrical to the Stage 3 enclosure. The southern Stage 1 stake holes were covered by the stub wall of the Stage 4 enclosure by Ring stone 43. The stake-holes cannot have been internal fittings of any of the later enclosures. So taken in conjunction with the scalloped edge of the enclosure to the west, they define a Stage 1 enclosure. The stake holes may have been truncated if the ground inside the enclosure was lowered during building of the later Stage 2 enclosure. But the preferred interpretation is that the ground was levelled down outside the Ring (the eastern part of which, it will be recalled, had earlier been built up with green clay) as part of making the Stage 1 enclosure. Just

Illus 24.52Â The Stage 1 enclosure and ironpanned features of ambiguous date on Plans 46/81 to 48/81 [NMRS DC38080 to 38082]. There was too little evidence to reconstruct the overall shape of the Stage 1 enclosure. It may have been circular, or bag-shaped like one of the enclosures on the green clay platform under the cairn. If it was not circular and some of the iron-rich stake-holes and features around the entrance to the Stage 2 enclosure in fact went with the Stage 1 enclosure the latter might have had a main axis running slightly north of east and just under 4m long (Illus 24.53). If so it could have been very similar in size and shape to the smaller and probably later of the timber-built enclosures on the green clay platform under the cairn (Illus 24.39 upper). The match is far from perfect and it is not suggested that the Stage 1 enclosure necessarily had the exact shape of the pre-cairn one. But their broad resemblance gives credibility to the idea that the Stage 1 enclosure was built at the same time as the chambered cairn and its function was to allow the practices followed in the pre-cairn enclosure to continue, albeit outside the Ring. Discussion and conclusions \ 1038

Calanais Survey and Excavation, 1979-88

Illus 24.53Â Overlay of reconstructed outline of the small sub-cairn enclosure on Illus 24.50 and the Stage 2 enclosure 24.13.3 Phase 9b: The Stage 2 embanked enclosure A small sub-oval enclosure was built in much the same area as the Stage 1 enclosure, immediately to the east-north-east of the Ring (Illus 24.54, 24.55). It was slightly less than 4m across internally. It had a broad eastern entrance, in which were broad shallow depressions containing thin silty deposits. .Illus 24.54 shows the area of the Stage 2 enclosure. The bank defining it suggests a more substantial structure than the Stage 1 enclosure. It may have been the slumped and ploughed relics of a relatively narrow turf and timber wall. There was no evidence to suggest roofing. In the west, near Ring stone 43, there was no surviving bank, but the area included several iron-encrusted stake holes which appear to have lined the higher ground to the south into which Ring stone 43 had been set. Their line bifurcated near Ring stone 43 suggesting that the enclosure was modified at some time. More posts in the palisade slot, which is shown in magenta on Illus 24.55, helped to define a turf wall. There is a possibility that the more southerly line of the palisade continued under turf lines 164 and 162 (see 24.7.1 and 24.9.1) but that is not the preferred interpretation of the field evidence

The middle fills of the enclosure in the area shown on Illus 24.57 were amongst the most heavily iron-encrusted at Calanais. The possible significance of this will be discussed along with the fourth stage enclosure. A large black patch of turf line on part of the outer edge of the bank in BII may suggest a phase of stability. It was cut by an ard, which had left marks to the west of the surviving patch of ground surface. This reflected at least one period of ploughing at a date before the latest wall-base 103 was added. Ambiguous linear features (not shown on Illus 24.55) to the east of the enclosure were interpreted as the much-ploughed remains of cultivation beds, but it is conceivable that they had also lined an entranceway to the enclosure. Bank makeup at Calanais included only Early/ Middle Neolithic sherds, one a possible Hebridean incised sherd. They had presumably been in soil and turf used to build the bank. The lack of later pottery allows a (non-preferred) interpretation that the first enclosure was built soon after the Ring and dated to between 2850 and 2750 BC. Two barbed and tanged arrowheads (Illus 24.58; CAT 209, 210), a sherd of a fine Beaker (Cat 771) and three other Early/Middle Neolithic or domestic Beaker sherds (Cat 829-831) in the silts of the entrance-way independently suggest a date for its active use after the introduction of Beaker pottery to the area after 2500 BC. Many different explanations could be given for their presence but one possibility is that they were purposeful deposits of complete arrows. Given that bone did not survive well at Calanais, another possibility is that they were part of funerary deposits in the entrance. Yet another is that they reflect a conflict. Central pit 149, about 0.55m across, which might be of this stage, contained a possible fine Beaker sherd (Cat 677); the sherd showed signs of marked abrasion (wear category 3) so was probably residual. But it must be emphasised that the pit and a round -bottomed pit 180 in an ambiguous relationship to it (boteh pits were much damaged by the Victorian pit 129) could have gone with any stage of enclosure except perhaps the first.

Discussion and conclusions \ 1039

Calanais Survey and Excavation, 1979-88

Illus 24.54Â The second stage of enclosure in 1980 with remains of the fourth stage wallbase outside it [Film 1980.11.23] 24.13.4 Analogies If this early embanked enclosure was rectilinear and dated to around 2500 BC it can be compared to some of the houses associated with grooved ware pottery in various parts of Scotland. For instance House 10 at Skara Brae (thought to have been late in the sequence there and to have had a special function within the settlement) and House 1 at Barnhouse in Orkney are of roughly the same size. the latter however probably dated to around 3000 BC. In some ways a more interesting comparison is with house-like structures at Beckton Farm (Pollard 1997). Beckton Farm is near Lockerbie, in the valley of Dryfe Water which runs north from near the head of the Solway Firth. It is not really beside the main western seaways, but it lay only two weeks leisurely travel away from Calanais by sea and river. The complex included at least two buildings (Illus 24.59), pits with cremated bone, two of which contained several Grooved Ware sherds, and four-post structures measuring c. 3m by 3m.

It was interpreted as a transient settlement, possibly occupied several times, at which four post structures might have been used for excarnation or as raised granaries. Some of the pits may have been used as receptacles for cremation pyre debris. Structure 136 was interpreted as bounded by stake holes and possibly a daub wall enclosing an area about 4m across (Pollard 1997, 75-6, Illus 3). Structure 111 was defined by a shallow ditch lined with stake holes defining four sides of a rounded pentagon and enclosed an area about 4m across. The entrance, formed by the fifth side of the pentagon, faced a little south of east (ibid 77-8, Illus 5). Clearly these structures differed in detail from the Stage 2 enclosure at Calanais, but they were of similar size and shared the general idea of building with clay or turf and stakes. The house-like structures were not directly radiocarbon-dated but dates (from mixed charcoal which may have included pieces old at the time) from features with Grooved Ware nearby provided a range of date, 3298-2980 cal BC, 2923 to 2617 Discussion and conclusions \ 1040

Calanais Survey and Excavation, 1979-88

Illus 24.55Â The Stage 2 enclosure Discussion and conclusions \ 1041

Calanais Survey and Excavation, 1979-88

cal BC and 2910-2500 cal BC (ibid, 89). Perhaps in part because of oak and residual charcoal, they suggest an earlier date than that preferred for the Stage 2 enclosure at Calanais. Despite the presence of older charcoal, the main period or periods of use at Beckton Farm probably dated within the span 2900 to 2300 (Pollard 1997 Table 1) and the excavator thought that Structure 136 might have been one of its later features (Illus 24.57). If the Stage 2 enclosure was originally meant to be oval a loose analogy can be made with a houselike structure found at Tormore, Arran (Barber (ed) 1997, 7-21). The first phase of the structure had Beaker pottery and barbed and tanged arrowhead associated with it. But the area enclosed by the bank of the Tormore was considerably greater at about 9m by 12m so the analogy is not close (Technical Note 24.13.4). The Beckton Farm structures could be characterised as ‘special purpose domestic’ like the best analogy for the Calanais Stage 4 enclosure, a building at Dalmore, Lewis, interpreted in that way by Edwards & Ralston (1997).

Illus 24.56 Part of the bank sectioned in Area BII [Film 1980-10-4]

24.13.5 Ploughing of the enclosure The Stage 2 enclosure bank was ploughed down before the Stage 3 enclosure was built. None of the ard marks in Areas BIV and BV south of the enclosure was datable to before the ritual deposition and ground-working of Phase 10a. So it seems most likely that ploughing of the enclosure succeeded the earliest part of Phase 10a. A piece of Pomoideae charcoal found in an ard mark in clay 123 to the north of the enclosure dated to between 2120 and 1770 cal BC and may reflect the date of some, but necessarily the earliest, of the ploughing of the bank. A pair of Food Vessel sherds (Cat 580_581) in bank-like soil filling a linear depression north of the enclosure may suggest a similar date, although it was possibly intrusive.

Illus 24.57 The southwest corner of the Stage 2 enclosure from the east, with scarped edge and iron-encrusted palisade stake holes

24.13.6 Summary The preferred interpretation of the Calanais enclosure is that it was ‘special purpose domestic’; by that term I mean (as I suspect that Edwards and Ralston also meant) a building used from time to

Illus 24.58 Arrowheads in the entrance silts, 209 quartz and 210 mylonite Discussion and conclusions \ 1042

Calanais Survey and Excavation, 1979-88

time while short term seasonal activities or periodic non-domestic activities took place nearby. It was stratigraphically later than the Ring but there was no direct evidence for its relationship to the timber structures under the chambered cairn, nor to the cairn itself. The preferred interpretation of its construction date is thus based on somewhat speculative ideas about the succession. Thus the preferred construction date, based on interpretation rather than hard evidence, is somewhere in the last half of the 3rd millennium BC. Its closure date is slightly better defined to the last few centuries of the 3rd millennium BC or the first few of the 2nd by finds in its entrance silts, which were probably contemporaneous with its active use. 24.14 Phase 10a: Ritual deposition and ground-working 24.14.1 Introduction As described in more detail below, the preferred interpretation of the evidence in post-cairn construction levels on Areas D and B is that ritual deposition took place over a long period, including that of soil and burial-related materials. Because of ground-working including ploughing, both during the period of deposition and subsequently, it is difficult to separate different periods of disturbance and Illus 24.60 is diachronic. Stones are shown in light grey and clay patches in various shades of brown. There were small patches of clay and soil deposition on top of the main turf lines 365 and 334 in Area D. They lay below ginger-coloured clay 320 which may have had a (generally poorly preserved) turf line separating it from subsequent plough soils 315 and 344. On Area B the (probably much later) patterns of stones found in and above old plough soil 141 may imply the existence of small rough enclosures prior to the latest ground-working. Much of the deposition may have been of domestic remains in soil. Much of it may have originated in Phase 3 material imported is soils during Phase 7. â&#x2014;&#x201A; Illus 24.59Â Calanais Stage 2 enclosure and Beckton Structures 136 and 111 adapted from Pollard 1997 Illus 3 and Illus 5 Discussion and conclusions \ 1043

Calanais Survey and Excavation, 1979-88

Illus 24.60Â Possible ritual deposition of material in Areas B and D Sheridan has identified a wide range of Neolithic sherds in corky and non-corky fabrics, and of both plain and Hebridean Incised types. She interprets them all as from Neolithic domestic assemblages of much the same character as each other. They were distributed widely across many stratigraphic layers. The single Grooved Ware pot ASH 61 (Illus 24.36) from scrapes outside the entrance to the Ring and to the pre-cairn enclosures in the Ring undoubtedly had a different origin, and was brought to the site as a complete pot, possibly in Phase 6 rather than Phase 7. Beaker sherds were abundant. Some, notably ASH 42 (Illus 24.62) entered the site as accompaniments to burials Ballin has noted that burnt quartz was also distributed widely across many stratigraphic layers. It included burnt flakes, cores and tools. The flakes were hard percussion by-products. He reported that

burnt quartz occurs on some domestic sites in the Western Isles but flakes scattered across ritual and burial monuments are usually simple bipolar flakes, produced by bashing up nodules on an anvil. Judging by these comparanda the burnt and unburnt flakes at Calanais, and other debitage from quartz working, are likely to have originated in domestic or domestic/craft contexts. Although it is in principle possible that ritual or ceremonial practises demanded tool production in the Ring there is no evidence for this at Calanais (Ballin Chapter 17 Lithics and Ballin pers. comm. May 2009). This suggests quite strongly that much of the quartz was not associated with remains imported from cremations conducted elsewhere, as might be suggested to explain other aspects of the evidence. However, the lithic assemblage undoubtedly contains elements with various origins. If the distinctive artefacts were mostly associated with Discussion and conclusions \ 1044

Calanais Survey and Excavation, 1979-88

burial-related activities, as seems very likely for the two arrowheads loosely associated with a probable burial in Area BV (Illus 24.63), or one-off events as seems possible for the arrowheads in the silting at the entrance in Area BIII to the second-stage enclosure, then their date need have little bearing on that of other elements of the assemblage. Overall, when combined with Sheridanâ&#x20AC;&#x2122;s and Ballinâ&#x20AC;&#x2122;s analyses, the stratigraphic evidence implies that much of the artefact assemblage originated in domestic activities elsewhere in the 4th or early 3rd millennium BC, and was imported during Phase 7 (clay and soil spreading at some date after Ring erection and before chambered cairn building); but others were brought in with ritual deposits of soil and clay during Phase 10, and a subset including the datable artefacts originated in burial-related and other activities in the late 3rd and early 2nd millennia BC. 24.14.2 Deposition, burials and groundworking on Area D On Area D shallow pits and low spreads of clay were detected at the base of the main plough soil. Numerous Beaker and other decorated sherds were found in the plough soil itself. The first deposits appear to be earlier than insular Beakers. One of the deposits in a pit may have been marked by a post. The pollen in some of the seemingly lowest pits on Area DI was comparable to that of Phase CaN-3c at Calanais Leobag, which should date after 1900 BC. It suggests that some of them had been dug from higher up than their surviving tops, their upper part having been disguised by soil processes which helped to produce ginger clay 320. That clay overlay many of the small spreads of brownish clay which are interpreted in part as the spoil from shallow pits or scrapes and partly small dumps of clay soil. The most intriguing feature in the area south of the cairn at this level was a squarish stone setting about a metre across in a dip in the old turf line under the plough soil (Illus 24.61). It may have been a small burial enclosure or a repository for ritual deposits.

Illus 24.61Â Squarish stone setting 376 south of the cairn (see Illus 24.60) [Film 1981-8-1] Several potsherds were found in its immediate vicinity. One was E/MN Corky (Cat 122). All of the others may have come from domestic Beakers, although two they might instead have been E/ MN Corky (Cat 595, 714, 765-766). Three quartz flakes and a miscellaneous piece of quartz were also found. However although they were ascribed to this context they might have been residual. A sherd from a fine Beaker (Cat 954) and a piece of cremated bone were found in a chocolate brown, fine silty clay 377 below the plough soil south of the cairn, cut by a possibly post-cairn pit and spread nearby at a higher level. There is no evidence for cremation taking place at Calanais, so either burnt animal bone occurred in some of the basketfuls of soil deposited in the area or some of the deposits came from cremations conducted elsewhere. As explained in Chapter 15, the few fragments of cremated bone from Calanais are so poorly preserved that they have been reserved from analysis until sample preparation techniques become more advanced. The interpretation preferred in Chapter 9.3 (Area D) was that the area was used for digging of small pits and deposition of charcoal, potsherds and other material from elsewhere. Then, after some more deposition, soil processes caused oxidisation before use of the area for more deposition recommenced. The closest part of the cairn became more and more dilapidated as time went on, perhaps as stones were removed to mark deposits. The area was ploughed, possibly ritually, or worked

Discussion and conclusions \ 1045

Calanais Survey and Excavation, 1979-88

over by hand, perhaps to merge dumped material into the plough soil. This continued, perhaps for generations. Probably some stones were removed from the plough soil to make ploughing easier (or possibly to destroy their function as markers). 24.14.3 Deposition, burials and groundworking on Area BIV/BV It does not seem very likely that the charcoal-rich deposits in BV and the southern part of sub-area BIV were dumped in 1857. If they had been it is unlikely that the placements of the stones above them would have possessed even the hints of structure that did remain. It is also unlikely that the charred heather twig found in a deposit 139 on top of the plough soil would have survived unbroken. There was a succession of layers to the southeast of the entrance to the chambered cairn passage which reflect a build up of dumped or spread soils and clays and the digging of shallow pits. This seems to have started during a period when Grooved Ware was in use for a pot was deposited probably in Phase 7b. Other pits or scrapes were later than the erection of Stone 30 but clearly earlier than some of the black deposits and plough soils with charcoal and pottery. They may be evidence for the lighting of fires outside the cairn and east of the Ring, but seem more likely to have been individual deposits of soil with charcoal and in some cases potsherds. If the early black deposits in and just above dark iron-panned layer 160 (below plough soil 141) originated in the chamber then the grey-green pottery-rich but charcoal-free clay 812 overlying them in Sub-area BV may have come from the chamber floor. But that is not a preferred interpretation; instead the origin of the green clay 812 is admitted to be obscure. Whatever its immediate origin it is conceivable that the greenish sandy clay 812 was used to seal over the spread dumps of charcoal-rich material. Above the greenish sandy clay a succession of black layers containing charcoal and potsherds included a plough soil 141. Some of the artefacts probably reflect disturbed burial deposits. As on

Area D the ground was worked over with a spade or ard or both several times during this period. The material in the main plough soil was similar to that found in Area D, in that it contained abundant charcoal and potsherds. Similar soils were mostly absent elsewhere on Area B.

Illus 24.62Â Beaker ASH 42

Illus 24.63Â Arrowheads near burial remains

Discussion and conclusions \ 1046

Calanais Survey and Excavation, 1979-88

The exception to the general pattern of deposition of mixed potsherds and charcoal is a group of Beaker potsherds (ASH 42; Illus 24.62) and two barbed and tanged arrowheads (Illus 24.63) adjacent to (not strictly associated with) the last vestiges of decayed human bones in a darker patch interleaved in but mostly below the plough soil 141 and at a higher stratigraphic level than the green clays. It is a reasonable candidate for a largely complete interment. Sherds from ASH 42 were found in 8 different contexts nearby, including the plough soil. The preferred interpretation is that it was a disturbed in-situ burial. But it must be admitted this pattern leaves open the possibility that the burial deposit was in a secondary position and had been removed from the chamber while the skeleton was complete or near complete. Sheridan (Chapter 18 The pottery Assemblage 18.7.12) compares it to pottery found elsewhere in the Hebrides, and suggests a date between the 22nd century BC and c 1750 BC. The hints of structure in the stones in the higher parts of the plough soil may reflect creation of and subsequent damage to slight enclosures built to contain burial-related materials (Illus 24.60). A piece of cremated bone was found in a line of stones running from Ring stone 44 in front of East Row stones 30 and 31 beneath modern turf and soil. The fragment may have been in secondary or tertiary positions, perhaps a relic of prehistoric

clearing of the chamber deposits, or possibly a part of deposits cleared from the chamber by Sir James Matheson’s men. One other pot, a domestic Beaker (ASH 48 Illus 24.64) was relatively unabraded and well represented (slightly less than 10% of the pot). It had a distinctive funnel-like neck (Sheridan Chapter 18 The Pottery Assemblage 18.7.7) and was decorated with lines impressed by a marine shell, probably not a cockle. Unlike many of the other domestic Beakers, the sherds of which were small and heavily worn, and which may have been brought to Calanais along with baskets of soil, ASH 48 possibly arrived as a complete pot. Its sherds had a very distinctive distribution, concentrated in plough soil 315 and disturbed Victorian context 326 on the cairn in Area DI. That area was otherwise dominated by finds of fine Beaker, ASH 48 may not be the only large pot at Calanais with a funnel-like neck although none of the other possible examples survive well enough for secure identification. Sheridan (Chapter 18 The Pottery Assemblage 18.7.7-8) notes that this form of funnel-necked domestic Beaker is known from elsewhere, citing as an example that from Cluntyganny, Co. Tyrone, Ireland: Given the distinctiveness of the form that raises the possibility of a connection with northern Ireland. 24.14.3 Deposition and cosmology The soils in Area H to the north of the cairn and in Area S to its northwest contained far fewer pieces of decorated pottery than those south and southeast of the cairn. The difference in distribution of pottery between the north-eastern and south-eastern parts of the Ring, and to a lesser extent (because less of it was excavated) south-western part of the Ring can be interpreted as evidence for a cosmological scheme. That pattern is supported by the distribution of possible ritual deposits (Illus 24.60). Pollard and Ruggles (2001, 81-2) listed 3 cosmological schemes involving radial divisions of the world.

Illus 24.64 Domestic Beaker ASH 48

—— Divisions of the world into two halves (not well attested in the ethnographic record). Discussion and conclusions \ 1047

Calanais Survey and Excavation, 1979-88

—— Divisions of the world into four quarters or more strictly parts. Their examples included hearth orientations in Orcadian houses, clustered round the four intercardinal directions (Richards 1990), and the use of different types of stone preferentially used in different quarters of the (solstitially oriented) passage cairns at Balnuaran of Clava (Bradley 1998c). —— Divisions into more complex sets of radial sectors (the example quoted is from the Inca Cuzco). They identified four quartering schemes; a cardinal scheme, and 3 schemes with divisions at 45 degrees to the cardinal directions, of which two are solar and lunar schemes (Pollard and Ruggles 2001, 82-4 Fig 4a). Superficially the deposits in the northern half of Area BV and those in Area D seem to have originated in much the same way as each other. Because BV was outside the Ring that would suggest that the Ring was not seen as a microcosm of the world. Instead it sat in a world divided into quarters from its centre. But fundamental to this interpretation is the question whether one should emphasise the similarities or the differences between the deposits in the two areas. Some of the lighter coloured patches on BV below the worked soils and those on D below the main plough soil were at least generically similar to one another. On the other hand, very few pots left sherds on both Areas D and B. Nevertheless the preferred overall interpretation is that similar processes did operate in the south-east quadrant of D and in BV during the earlier periods of deposition. Subsequently there was another depositional phase on the northern part of Area BV leaving abundant very dark soils, amongst which were skeletal material and many sherds of fine Beaker ASH 42 representing about 15% of the vessel. There were no identifiably similar discrete burial deposits on Area D, apart from, just possibly, the squarish stone setting referred to above (Illus 24.60) although it seems quite likely that if similar ones had been there they would have been broken up by ground working. If that analysis is right the events reflected by the archaeology of Areas BV and D followed similar trajectories over a long period; but the later history of the two areas was different.

The stone rows at Calanais (including the hypothetical north row consisting of the southernmost three stones of the west side of the avenue) suggest an approximately cardinal scheme of division of the world into quarters. But the burial-related deposits on Area BV extended further north than the line of the East Row (Illus 24.60), suggesting that the two schemes were not identical. Therefore the preferred working theory is that the deposits on Areas D and BV reflect a cardinal quartering of the cosmos which was subsequently superseded by the less precise quartering suggested by the stone rows. The relationship of the first deposits of pottery and charcoal to the Ring and chambered cairn are obscure. Some of the small pits on Area D were first noted after the prevalent complex turf lines had been completely removed. At face value this implies that they preceded the Ring and central monolith. However it is possible that all of them had been cut through the ground surface which formed after the chambered cairn was built and subsequent turf formation modified their upper fills. Others were first noticed above the turf lines but below ginger clay 320; the pollen in them and in clay 320 suggests that some of them were cut from a higher level and that soil processes had disguised the fact that they were originally cut through the ginger clay. The preferred overall interpretation is that many of the early deposits were not related to cremations or inhumations but to dumping of basketfuls of soil from domestic contexts. On the other hand the acidic soils at Calanais are likely to have dissolved most cremated bone, and some of the other deposits may have been charcoal-bearing soils which had been associated with cremations elsewhere. In Scotland overall, cremated human remains have been radiocarbon-dated to around 3700/3600 cal BC, as early as inhumation apart from earlier stray human bones in middens and caves slightly after 4500 cal BC (Technical Note 24.14.5). Known examples from the first half of the 4th millennium come from a pyre under a ring mound at Midtown of Pitglassie in Aberdeenshire, from a mortuary enclosure at Pencraig Hill, East Lothian and from a large cist at Moleigh (also known as Cleigh) in Argyll. These geographically widespread burials suggest that the rite was Discussion and conclusions \ 1048

Calanais Survey and Excavation, 1979-88

practised by different groups of people from the earliest centuries of farming in Scotland, and the very various contexts may reflect equally different local beliefs; but the data set is tiny and the few dated examples may instead reflect different aspects of a widespread tradition (Technical Note 24.14.5). Only three cremations have been dated to the following thousand or so years, at the long barrow at Fordhouse in Angus, in the hole of a standing stone at Orwell in Perth and Kinross, and in a Beaker cist at Dornoch Nursery, Highland (Technical Note 24.13.5). Deposition of cremations and cremation-related material only became common in Scotland from around 2100 to about 1500 cal BC, peaking roughly from about 1800 cal BC (Technical Note 24.14.5). In the 2nd millennium BC cremations were put in a wide variety of new and old structures; but the dated examples were mostly found associated with urns in flat cremation cemeteries. The evidence from Calanais contrasts with that from Stonehenge where cremations are judged to have ranged in date from the 30th century cal BC to the 24th century BC (Parker-Pearson et al 2009, 36). Nearly two thirds of them belonged, according to their model, to between the 26th and 24th centuries BC. Some cremated bone at Calanais was found in contexts underlying the chambered cairn. Charcoal from the same contexts was dated to the last quarter of the 4th millennium. Pottery included possible and probable Beaker sherds. So the cremated bone could be from animals or humans and could belong either to the late 4th or the late 3rd millennium BC. The survival of a few other pieces of poorly preserved cremated bone at Calanais, and the 19th century record of cremated bone in chamber deposits could represent the phenomenon seen at Stonehenge, but the current rarity of dated cremations of the 26th to 24th centuries BC in Scotland could be taken to suggest that the cremated bone at Calanais is more likely to date after 2500 cal BC. Resolving this, should, as detailed in Chapter 15: Introduction to the Specialist Reports await improvements in dating techniques because the surviving pieces of cremated bone are small and poorly preserved.

24.15 Phases 10b and 10d: The Rows and Avenue Strictly speaking too little dating information was obtained to put the Avenue and South and West Rows in the phasing scheme. That difficulty is accentuated by the possibility, discussed below, that they may have been built sporadically over a fairly long period. Their phasing should therefore not be trusted overmuch.

Illus 24.65Â The Avenue and Rows in 1980 during excavations 24.15.1 The Rows The Rows have no obvious overall symmetry (Illus 24.69). The South Row points almost due northsouth but the East and West Rows do not line up with each other. As noted above, the three southernmost stones of the west side of the Avenue could conceivably have originally formed a North Row. Discussion and conclusions \ 1049

Calanais Survey and Excavation, 1979-88

24.15.2 The East Row The east row is not straight and it is orientated a little north of east. It consists of five stones, including that found on Area C where Palmer showed a stone in 1857 (Illus 24.66-24.68; see Chapter 3 Previous Studies). The fallen stone was found there (initially through documentary research and probing by Gerald and Margaret Ponting) along with a stone-hole with surviving packers, into which the foot of the stone fitted exactly (Chapter 8 Area C). It was set into a much earlier cultivation bed. It could not be related directly to a small deposit of potsherds with decoration reminiscent of that on Food Vessel 75 (Cat 558_562) under a peaty mound nearby; in any event, the sherds were abraded and may have been collected much later during nearby cultivation and deposited in a mound of weeds and stones. Something like that seems to have happened on Area E at the edge of a cultivation soil, albeit involving Hebridean incised wares. At the opposite end of the row, near the Ring, part of the pit for Stone 30 was exposed in Area B. Although the stratigraphy in this area was highly complex the pit cut an area of black soils and deposits of material related to burials. It was overlain by a dark soil 141 interpreted as a plough soil. It cut an earlier shallow dark soil 160.4. The stone was erected at some date between 2560 and 1690 cal BC, stratigraphically significantly later than the Ring (see Chapter 7.6 Area B). The stratigraphy of Stone 31 was complicated, probably by an earlier pit. This hinted at a prospect of tighter dating than we achieved if further excavation is undertaken. But its immediate consequence was that there was no independent dating evidence for erection of Stone 31; and although their stratigraphy was broadly similar there was no evidence to show whether Stones 30 and 31 of the East Row were set up as the same time as each other or decades apart. 24.15.3 The hypothetical North Row The hypothetical North Row consists of the three southernmost stones in west side of the Avenue. The resistivity survey did not reveal any anomalies

Illus 24.66Â The Glasgow 1975 plan with lines drawn between adjacent stones of the avenue and rows Discussion and conclusions \ 1050

Calanais Survey and Excavation, 1979-88

for at least one more stone at a distance similar to the average distance between the other stones. As noted by RCAHMS (1928, 26), however, the westernmost stone is blocky where the others are slabs aligned with the row, and it looks as if it was intended to define its end. By chance the West Row is aligned almost precisely on national Grid East-West (Illus 24.67). That provides a salutary counterbalance to some theories of the reasons why stone rows were aligned as they were. That said, its orientation is unusual for Scotland; it is discussed in the context of the East Row below. 24.15.5 The South Row

Illus 24.67Â The Rows including a hypothetical North Row between the 3rd and 4th stones (Chapter 5 Illus 5.1, 5.10). It thus provided no support for any idea of a North Row extending beyond the three stones. The hypothetical row formed by the three stones was on almost the same line as the South Row (see Chapter 24.14.10 for discussion). 24.15.4 The West Row The west row is nearly straight and consists of 4 stones (Illus 24.67, 24.69). It is much shorter than the East Row. There is a rock outcrop to the west not far beyond it but there would have been room

The south row today consists of 5 irregularly spaced stones. It is nearly but not completely straight. It runs on average almost geographical north-south and points fairly accurately at the central monolith (Illus 24.66, 24.70). Bedrock outcrops in the middle of the row and an area of high resistivity readings suggests that rock is near the surface elsewhere, although the soft overburden deepens again between the southernmost surviving stones (See Chapter 5 Resistivity, in particular Illus 5.1 and 5.13). Excavation by Edinburgh University on the other side of the wall at the south end of the area in State Guardianship revealed complex remains between it and the natural rock outcrop Cnoc an Tursa It has been suggested that it was a kerb cairn (Neighbour 2005, 4, referring to an unpublished report of 1999 by Campbell and Coles, held in the Department of Archaeology at Edinburgh). The other published accounts of the excavations do not

Illus 24.68Â The East Row from the North after re-erection of Stone 33A [Film CAL Gen8] Discussion and conclusions \ 1051

Calanais Survey and Excavation, 1979-88

mention a kerb cairn (Coles 1993, 110-11; Coles and Rees 1994, 96). Instead the 1993 account refers to a platform below 5 large boulders. Investigation of the platform revealed large pits, four shallow post-sockets and a shallow ditch running northsouth cut by post holes along with other post holes. The 1994 account described a focus of prehistoric activity around a fissure in the c. 2m high natural rock face below Cnoc an Tursa. The lower layers in the fissure consisted of black charcoal-rich deposits including hearth-like settings of flat stones. They partly covered a large pit which was cut by some of the pits discovered the year before. It might be best to interpret most of the stonework as a flat cairn or platform bounded by a long arc of larger stones (Campbell pers. comm. 2009). Another large pit, also partly excavated the year before, appeared to contain packing stones. It lay directly on the line of the South Row about 10m north of the fissure and was interpreted as a removed standing stone. Thus the South Row may once have been longer than it is today.

Illus 24.69 The West Row from the Ring photographed by Richard Strachan of Historic Scotland [Film DSCN 0981]

24.15.6 Interpreting the rows In a sample of 300 Western Scottish stone settings studied by Clive Ruggles only two sites have broadly east-west indications, Calanais and Blashaval in North Uist where three stones are placed in a 50 m long line, far longer than the other short rows included in the sample (Ruggles 1999, 75). The façade of the Ring was designed to face very nearly due east and the East Row was set at a significantly different angle so it was probably created for a different purpose from those underlying creation of the Ring. This reinforces the evidence from archaeology that the East Row was significantly later than the Ring. However the East Row was set on the slight ridge formed by Cultivation bed 2 (Illus 24.10). That may not have been an accident and although it does not provide an explanation for why the row was set up in an eastward direction it does explain why, when the decision to put up a row of stones in that general direction was made, they were set in that particular direction. As described below (24.15.11) in the context of the Avenue, Curtis and Curtis (2009) have shown

Illus 24.70 The south Row from the South photographed by M Brooks of Historic Scotland (Film BR.3.6) that the stones of the East Row could have been used to mark out significant moon rising points as viewed from near stones on the east side of the Avenue. Whether that was actually done is one matter; whether it was ever done by more than a few individuals finding a new use for an old monument is another; and whether it was a material consideration when the stones were set up is yet another. Lacking any independent knowledge of what people in the 2nd millennia and late 3rd millennia regarded as significant we cannot prove the matters one way or another. The orientation of the West Row is of course as unusual amongst prehistoric rows as that of the East Row. Despite there being room for another Discussion and conclusions \ 1052

Calanais Survey and Excavation, 1979-88

stone between it and the sharp drop to the west the blocky plan of its westernmost stone suggests that it was intended to end where it does today. The same blockiness can be seen in the re-erected end stone (Stone 33a) of the East Row (Ponting and Ponting 1984, 12). On the ground the South Row more obviously points to the central monolith than plans convey, for at ground level the row is slightly ragged (Illus 24.71). It is conceivable that it was intended to link the Ring and Cnoc an Tursa, the prominent outcrop to its south, particularly given the evidence described above for activities at the base of the outcrop. That would fit with ideas about its being one side of an avenue (RCAHMS 1928, 24). But excavation in Area E along the supposed east side of such an avenue produced no evidence to support the idea.

Illus 24.71Â Pattern-making and Pit 917 in Area D The excavation evidence from the East Row suggested strongly that at least the west end of the East Row was later than the Ring. There was no evidence that parts of rows were removed from the area of the Ring when the latter was built.

Only one pit (Pit 917, coloured green on Illus 24.71) could possibly bear this interpretation. But as described in Chapter 9 (Area B) it was probably neither a posthole nor a stone-hole. Nevertheless, Pit 917 could be said to lie on an extension of the east end of the East Row, if one allowed oneself license in drawing an alignment, ignoring the westernmost stone. If the central monolith is included with the hypothetical North Row they and Pit 917 line up fairly well. Thus Pit 917 was nearly at a point where the extended alignments of two existing Rows and the hypothetical North Row (nearly) met each other. It could even be argued that the South Row and the hypothetical North Row once formed a single somewhat ragged chain, hardly more irregular than the existing East Row or the sides of the Avenue. However, in Area S to the north of the central monolith excavation of the top 0.15m of strata did not reveal any sign of a similar pit near the line of the hypothetical North Row, although a pit was discovered well to its west. Pit 917 also lies almost on a line drawn between Stones 9 (outside the SW quadrant of the Ring) and 34 (outside the NE quadrant of the Ring). It is shown as a fainter red line on Illus 24.71. It is conceivable that stones 9 and 34 were set up before the Ring, although equally possible that they were later; indeed the evidence from excavation by stone 34 suggested (but did not prove) that it was later. Pit 917 is the best candidate amongst the several pits recorded in Area DI for a post- or stonehole potentially recording a pre-Ring alignment of the kind suggested by some authors, but given the lack of evidence that it was dug to support a post or a standing stone, and given its similarity to the post-Ring pit under the passage in Area BIWX which almost certainly was never intended to support a standing stone or timber, it should instead be compared to the pits found on several ceremonial sites including Machrie Moor, Arran and Balfarg Riding School, Fife. In sum, I must admit to regarding the exercise illustrated by Illus 24.71 as purely modern pattern-making. I suspect that those who like such things will admire the (near) coincidences described above and make something of them, Discussion and conclusions \ 1053

Calanais Survey and Excavation, 1979-88

while others, with whom I side, will argue that coincidence is precisely the right word to use. Overall the preferred interpretation is that, as has long been suggested, all the rows at Calanais were later than the Ring. 24.15.7 The distribution and dating of stone rows

Illus 24.72Â Stone rows after Burl 1993, 92, 148, 149 Although the Rows at Calanais may be, perhaps should be, regarded as different from free-standing stone rows, the latter may have influenced the creators of the former. On Illus 24.72 long single rows are in red, 4- to 6-stone rows are in blue and

3-stone rows in black at a smaller dot size. Where there are strong local concentrations numbers are under-represented because of the scale of the map. For detail see Burl 1993. The East, South and West Rows at Calanais fall into the 4 to 6 row category. The mainly western distribution of Scottish 4 to 6-stone rows is from Calanais to Kintyre although the example in Caithness is given added interest by the style of the Calanais cairn. The distribution of 3-stone rows is similar; they are more abundant, are found further south and they spread further east in the Midland valley. There is a notable lack of surviving single rows of any length in north-east Scotland and the Orkneys (Illus 24.72; Burl 1993, 147-9 Fig 37-38). But it cannot be demonstrated that the intensity of land clearance in those areas is not a factor in that. Burl contrasted the overall distribution of short and long rows and suggested that their difference means that the short rows developed from the long as â&#x20AC;&#x2DC;economical versions of the longer rows that existed in adjacent areasâ&#x20AC;&#x2122; (Burl 1993, 151). He suggested that unlike the longer rows, short rows fairly commonly formed alignments on astronomical phenomena. That is discussed below when considering the various claims for astronomical alignments at Calanais. The date range proposed by Burl for long stone rows was 2100-1600 BC and that for short rows was 1800-1000 BC (Burl 1993, 91, 147). Since Burl reviewed the evidence more has been uncovered. The earliest dated stone row in Britain is on Cut Hill, Dartmoor. It was longer and sparser than the rows forming the Avenue. Dating of peat above and below fallen stones suggested that two of them fell or were laid flat in the 4th millennium BC (Technical Note 24.14.2; Fyfe and Greaves 2010, 55, 59, 62-67). Fyfe and Greaves were careful to point out that the stones of the row might never have stood upright, and of course none of the rows at or around Calanais count as a long row, apart from the sides of the Avenue if they are considered in isolation from one another. But Cut Hill provides an admittedly very general precedent for the existence of stone rows in the 4th millennium, only a few weeks travel away along the western seaways, reDiscussion and conclusions \ 1054

Calanais Survey and Excavation, 1979-88

ducing the most obvious obstacle to free speculation about the relative dates of the main linear and central elements of the stone setting at Calanais. The distribution of short rows suggests that at the time the rows at Calanais were built the main contacts of the people there were along the western seaways to the south. There is no suggestion of the Aberdeenshire connection hinted at by the avenue at Broomend of Crichie. Perhaps the distribution mapped by Burl is only part of a wider story. Timber single rows may once have been common in Scotland. The best dated examples I know of are two long rows found at Eweford East, East Lothian (Lelong and MacGregor 2007, 53-68). Both ran roughly east-west. Judging by several radiocarbon dates obtained from single pieces of charcoal both rows were probably built (at least in part) between about 2500 and 2250 cal BC. Grooved Ware was found in some of the pits. Of course, technically, the radiocarbon dates and pottery provided only termini post quem but the lack of later charcoal or pottery suggests that the dating is probably at least roughly right. One pit produced a much earlier date which the excavators thought might reflect use of the structure for 600 years but in my view probably reflects the presence on site of charcoal which was old at the time the pit was dug. Charcoal from hazel and willow, and in the southern alignment also Rosaceae, suggest that the rows may have been panelled. The southern row consisted of up to eight straight segments, between most of which were gaps, which strongly suggests several episodes of construction. These rows were interpreted as complex creations with the digging of each pit perhaps an individual episode in a long process of creation. 24.15.8 The Avenue On Illus 24.73 the stone pits on Areas A have been included. The avenue runs roughly 9 degrees east of true north from the circle. Its sides have been described as splaying out slightly but the main contribution to that effect comes from the three southernmost stones of the west side of the avenue. They do not themselves form a straight line but are, on average, along a different line to the stones further north. The rest of the west side

seems to be simply slightly irregular, with the northernmost stone to the east of the approximate line of the rest. The east side was probably no straighter; in general it pointed at the side of the Ring, and roughly at Cnoc an Tursa, the knob of rock beyond the south end of the setting. Stone 34 (connected to the rest of the east side by a dotted line on Illus 24.73) was offset from the general line of the east side and oriented at a markedly different angle to the Avenue stones, so it was probably set up for different reasons Excavation of part of its pit (in Area BIN) produced no satisfactory relative or absolute dating evidence. Modern sub-soil weathering had obliterated relationship between the top of the pit and the surrounding mineral soils. To erect the stone superficial soils and stones were cleared over a larger area than required for the deepest part of the stone-pit. Then the latter was dug. The stone was placed in the pit. Finally head-sized stones were placed round the base of the stone. What appeared to be packing stones stuck up into the weathered soils above and in Chapter 9.5 the subjective impression of a relatively late phasing (perhaps somewhere between Phases 10 and 12) is recorded. But there was no objective evidence. The gaps between surviving stones are very irregular. It has been suggested that the gaps may have been original; but discovery of at least one stone pit (003) in Area A where a stone was planned in the early 19th century AD (MacCulloch 1819) confirms that some gaps were the result of removal of stones in modern times (See Chapter 3: Previous Studies and Chapter 6: Area A). Pit 003, the northernmost of the two pits on Area A shown on Illus 24.73 and 24.76 was close to the line between the two northernmost existing stones; twice as long as it was broad it was oriented along the line. If the stone in Pit 003 had the same proportions as the pit it would have been a slab twice as long as it was thick. The other pit marked in Area A, Pit 004, may also be the remains of a stone-hole although it had two phases of fill. A stone matching the whole pit would have been about the same width as it was long. The occurrence of two pits so close to each other suggests that some stones were erected and then dismanDiscussion and conclusions \ 1055

Calanais Survey and Excavation, 1979-88

tled in prehistory. Alternatively it is conceivable that the west side of the avenue was marked out only by pits at one stage. 24.15.9 Timber avenues and similar structures in Scotland It is conceivable that the stone rows forming the avenue at Calanais were preceded by a row of pits (Chapter 6: Area A). If it can be proved that the charcoal dating a pit alignment at Crathes Warrenfield in Aberdeenshire to well before the inception of farming in the area was not residual (Technical Note 24.14.3; Murray et al 2009, 1617), by analogy the hypothetical pit alignment at Calanais could even be the earliest feature there. However the nearest excavated double pit alignment to Calanais was probably much later. It lay at Upper Largie in the Kilmartin Glen amongst a palimpsest of timber structures including a cursus, a pit/post ring complex, a large timber ellipse and many burials. It had been partially destroyed by quarrying and it is unclear whether it was an avenue as defined by Burl (1993, 4). It dated to after c. 1600/1400 cal BC (Technical Note 12.14.3; Cooke et al 2010, 174-175, 193, 195, 202). Timber avenues forming approaches to large enclosures have been excavated at Forteviot and studied from air photographs in conjunction with the excavations at Meldon Bridge in Scotland. The avenue at Forteviot was orientated almost exactly north-south. It was 4m wide and some 30 m long; it was built of up to 19 massive timbers which would have stood 3 to 4 m tall. It has been provisionally dated to the 27th or 26th century BC (Noble 2009, 231. No distinctive pottery was associated with it. The Meldon Bridge avenue was oriented roughly NW/SE; it was 4m wide and about 27m long. Judging by the air photographs the 16 posts were massive (Speak & Burgess 1999, 24-5). The enclosure itself dated to after 2600 cal BC and could have been as late as 1900 cal BC (ibid, 110). Similar enclosures with avenues are known from elsewhere in Scotland and from England and Wales (Barclay 2001, 149-151). These timber avenues were clearly intended to provide monumental entrance-ways to the enclosures. The enclosures towards which they lead

Illus 24.73Â The Avenue Discussion and conclusions \ 1056

Calanais Survey and Excavation, 1979-88

nue at Calanais. Similar timber double-alignments at Cowie Road near Stirling belong in the first half of the 4th millennium (Ashmore in Thomas (ed) 2007, 248-254). The main formal difference between the early long timber enclosures and avenues is that the former were closed at each end. But if the late 3rd millennium activities at Holm and Holywood reflect a continuing or renewed interest in the enclosures there may have been some connection between the two phenomena. 24.15.10 The distribution and dating of stone avenues

Illus 24.74Â Pits 003 and 004 in Area A looking southward [Film 1980.10.16] were very different from the Ring at Calanais and there is no reason why there should be any direct connection between them. The only similarity, I would argue, is that they lined approach routes. The double rows at Holm Farm in SW Scotland formed a complex pattern (Thomas (ed) 2007, 201 Fig 22.1). Their elements date from the first quarter of the 4th millennium BC. Sporadic later activities at Holm include what seems to have been creation of a pit-defined enclosure or avenue at the end of the 3rd millennium, and the digging of ring-ditches at substantially later date. The long timber enclosure at Holywood North lined the inner side of a ditch defining the cursus there, and seems to have been set up in the mid 4th millennium. It saw some activity, the cutting of a small ditch and the re-cutting of part of the cursus ditch, in the second half of the 3rd millennium BC, dates not far from those proposed for the ave-

Illus 24.75Â Stone avenues and double rows after Burl 1993, 42, 79

Discussion and conclusions \ 1057

Calanais Survey and Excavation, 1979-88

â&#x20AC;&#x153;Classical avenues are double rows that are unequivocally attached to a henge such as Stonehenge or a stone circle like that at Callanish. They were popular in north-west and southern England. It is the conjunction of the row with a ritual enclosure that determines its status as an avenue.â&#x20AC;? (Burl 1993, 4). On Illus 24.75 definite examples of avenues attached to rings and other settings are shown in red. Definite double rows are shown in blue. The concentrations on Exmoor and Dartmoor contained many more examples than can be shown at the scale of this map. For detail and for other possible sites see Burl (1993). The nearest certain stone avenue to Calanais as the crow flies is at Broomend of Crichie, Aberdeenshire. It was about 10 days away by sea (see Chapter 24.1.6) with a day or so of subsequent overland travel. Broomend shares another feature with Calanais: at one stage it included a central monolith (Bradley 2011, 88). It had been inserted into the top of a refilled shaft grave and the grave or the monolith may have been the focus of the two parts of the avenue. The smaller northern part of the avenue lay between the monolith/shaft grave and a recumbent stone circle (Burl 1993, 59), or probably so (Bradley 2011, 84). The southern part of the avenue ended near but to one side of a Beaker cist cemetery (Ibid, 85). The henge also surrounded a shaft grave, a stone setting and cremation burials in Collared and Vase urns; a later timber circle lay outside the henge and athwart the line of the west side of the southern part of the avenue (Bradley 2006, 19-20; Bradley 2011, 85). The central shaft grave and the cemetery at the south end of the avenue probably belong somewhere in a period between 2450 and 2150 BC. The shaft grave may well have been dug to take a Beaker burial (Bradley 2011, 73-74). The terminus post quem for building of the henge lies somewhere between 2150 and 1900 BC (Ibid, 74). The timber circle may date somewhere between 1850 and 1500 BC (Ibid, 74). The relationship of the avenue to the other parts of the complex is somewhat ambiguous (Ibid 75-78) although the henge can be seen as a late structure post-dating the northern avenue (Bradley 2007, 26; Sheridan 2007, 221).

Bradley thought that building of the avenue could date to any time between 2450 and 2150 BC. Given that the timber circle dates after 1850 to 1650 BC and before 1650 to 1500 BC (Bradley 2011, 74) and intrudes into the avenue (Ibid, 85) the latter had probably fallen out of use by then Ibid, 80). Bradleyâ&#x20AC;&#x2122;s data did not allow him to say whether the two parts of the avenue were built at the same time as one another or at different times. He thought, however, that the avenue parts each and together showed signs of a unitary design with alternating massive stones and slender orthostats spaced between 18.5 and 20m apart (Ibid, 78). The available evidence was skimpy and although there is no positive reason to reject his interpretation, in my view further excavation of the avenue is required to test it. Inland from Broomend there may have been five avenues attached to recumbent stone circles (Burl 1993, 59). The stone avenues shortest away in time from Calanais (about 8 days by boat and then a day or so overland) are in England and Ireland. Six were attached to rings in Cumberland including that at Shap where large boulders once defined an avenue about 3 km long (Burl 1993, 47). It is comparable to the avenue at Calanais only in the sense that both can be interpreted as consisting of stones erected to either side of a route way. Three examples near Broughderg in N Ireland consist of relatively small stones (all below 0.9m tall) and the stones on the western side of the avenue are markedly lower than those on the other (Burl 1993, 57). In southern England there is a concentration of 8 certain avenues on Dartmoor. They are very narrow with an average breadth of 1.5m and most of the stone used were small; the few much larger stones seem to have formed portals (Burl 1993, 50-54). They do not provide a close analogy with the avenue at Calanais. Three possible examples occur on the south coast; one may have had an avenue comparable to the Avebury avenues (Burl 1993, 50). The other avenues in southern England are associated with unusually large and complex ceremonial sites. At Stanton Drew the two surviving avenues point to stone circles; one at least of comparable dimensions to the Calanais Avenue Discussion and conclusions \ 1058

Calanais Survey and Excavation, 1979-88

(Burl 1993, 41-5). At Avebury the Kennet Avenue was at a much more massive scale than Calanais and seems to have been built in segments. It points to the massive henge rather than to a stone circle (Burl 1993, 45-7) At Stonehenge the earthen avenue (not included on Illus 24.75) may have pointed to the sarsen ring of stones in Stage 3 of the scheme proposed in 2009 by Parker-Pearson and colleagues, which they date to between 2580 and 2470 cal BC (Parker-Pearson et al 2009, 26). It may be slightly later, between 2580 and 2280 cal BC, if the date provided by an antler pick from the bottom of the ditch near Stonehenge is preferred to that implied by the 2009 phasing (Parker-Pearson et al 2007, 628). It is important to emphasise that the purpose and date of the ditch-defined avenue at Stonehenge cannot be related unquestioningly to those of the orthostat-defined avenue at Calanais. However both run from a stone ring towards water, albeit in the case of Calanais seemingly only for a small part of the distance to Tob na Faodhall, the Bay of the Ford near the standing stone at Cliacabhaig (Illus 24.33), and the date range independently proposed for the Calanais Avenue overlaps with that suggested by the radiocarbon date from the Stonehenge Avenue. Between the Cumbrian and Wessex concentrations, Arbor Low in Derbyshire may have included an earthen avenue (not shown on Illus 24.75). The only other avenues attached to rings are those in Brittany near Landouec and Kergonan (Burl 1993, 42, Fig 7). In addition to these, Richards has recently recorded a very short avenue at the complex setting at Na Dromannan (Calanais 10) formalising ‘a preferred direction of approach from the S’ (Richards 2006c, 184). It hardly seems comparable to that at the main setting, at Calanais even if the avenue started off as a ‘short’ one. Burl has also documented detached double stone rows in Caithness, Cumbria and Wales and tangential avenues in the northern part of Ireland. There are many examples on Exmoor and on Dartmoor of double stone rows, much narrower than avenues with an average breadth of 1.3m. The three known double rows in Caithness were built of low stones and are fairly short (Burl 1993, 90). There is no reason to connect them with Calanais.

Burl suggested in 1993, on basis of the Beaker associations, that avenues attached to ritual enclosures date to between 2600 and 2000 BC; he noted for instance the N2 Beaker burial in Avebury’s Kennet avenue stone 25a pit contemporary with erection of the stone, quoting I F Smith (1965, 210) (Burl 1993, 67-8). The earliest Beakers are currently held to date at earliest to around 2500 cal BC and the insular Beakers associated with avenues are probably significantly later, but that detracts little from his general argument. It fits adequately with the chronology suggested for Calanais although it is worth stressing that there none of the Beaker sherds there were directly associated with the avenue. 24.15.11 Interpreting the Avenue Aubrey Burl’s has suggested that the Avenue was originally short and later extended. He cited the increasing height of the stones nearest the circle and the increasing height northward of the rest of the avenue stones “causing the avenue to sag like a hammock at its middle” (Burl 1993, 61). Indeed it does (Illus 24.76), although the hammock would have been uncomfortably lumpy. The prominence of hornblende ‘eyes’ on the southernmost three stones of the east side of the Avenue (Curtis and Curtis 2009, 30-31) provides more support for the idea that the avenue was originally short. Other Rings seem to have had portal stones framing an approach to them (Burl 1993, 35-9). A variant on Burl’s idea that the Avenue was originally short is the possibility that the three southernmost stones of the west side of the avenue originally formed a north row. That proposition, discussed in more detail in Chapter 24.15.3, might mean that the setting originally had four rows pointing roughly in the cardinal directions before the rest of the avenue was constructed. A superficially appealing alternative to the idea that the avenue started off short is that some of the northernmost stones were put in place first. Then shorter stones might have been inserted between them and the Ring. That notion is built on the concept promulgated by Curtis and Curtis that the northern end of the avenue was used to observe Discussion and conclusions \ 1059

Calanais Survey and Excavation, 1979-88

Illus 24.76Â The heights of the stones (data based on Pitt-Rivers 1882; east side of avenue red, west yellow) the moon setting into the Ring at the southernmost lunar standstill. Burl also compared the avenue to Breton examples, saying that the stones on the eastern side were consistently only three quarters as tall as their western counterparts. Illus 24.76 shows indeed that when considered as pairs of stones none of the eastern stones is taller than its (approximate) counterpart on the western side, but it does not support the idea of a consistent 3:4 ratio. In a similar vein he thought that the layout of Calanais may have been influenced by â&#x20AC;&#x2DC;high and lowâ&#x20AC;&#x2122; rows of stones on the hillsides in N Ireland (Burl 2000, 145). The possibility that the Avenue was built over a long period of time does mean that different ideas could have predominated when different stones were erected, and as Burl suggested Irish influences may have played a part during some periods of stone erection. The funnel-neck Beaker ASH 48 found inside the Ring does suggest links with Ireland in the centuries around 2000 BC (Chapter 18 The Pottery Assemblage 18.7.7) Most avenues seem to have lined routes to large monuments. At Calanais the modern township

lies alongside one of the likeliest northward routes. Stones possibly emplaced in prehistoric times have been recorded in the modern settlement (Margaret Curtis pers. comm. 2009) but no firm candidate stones for a continuation of the Avenue have been published. The surviving Avenue points downhill towards Tob na Faodhall, the Bay of the Ford, as noted by Burl (2000, 207). Given the changes in sea level since the avenue was built that ford may have been above high water mark around 4500 years ago (Illus 24.3) but there is no reason why a route continuing north from the Avenue need have been straight, and it may have led to a narrow neck of land between salty and fresh water. In that respect it may have been similar to the hypothetical ancient route between the Stones of Stenness and the Ring of Brodgar in Orkney and to the avenues leading from Stonehenge and Durrington to the River Avon (Catling 2009, 23-4). Colin Richards has proposed that standing stones in some complex settings were erected from time to time by locally important people mobilising the community in celebration (Richards 2005, 217, 224). That seems particularly applicable to Discussion and conclusions \ 1060

Calanais Survey and Excavation, 1979-88

avenues. If the Calanais Avenue lined (part of ) an ancient route to the Ring it may have been built over a long period and one need not suppose that its stones were ever particularly regularly spaced. The exact positioning of individual stones might have been determined by reason like those put forward by Curtis and Curtis. However, as so often when dealing with stone settings the evidence allows this interpretation without proving that it is correct. Another possibility is that the Avenue was continued northward in a different style from the surviving part. The Kennett Avenue at Avebury, for instance, was built in segments probably over a long period (Burl 1993, 67-8). Perhaps at Calanais smaller stones were used in periodic extensions. The Avenue at Calanais may have served more than one purpose, and they need not have been mutually exclusive. In addition to lining a route it may have been related to celebrations of moonset into the circle at the lunar maximum, which occurs every 18.6 years. As Curtis and Curtis have documented, this phenomenon is easy to observe from the north end of the avenue. Curtis and Curtis (2009, 28) have also suggested that no stones are missing from the east side of the avenue and the stones were used along with a horizon marker for observing extreme settings of the moon. This idea is discussed below in the section on astronomy. But the preferred interpretation is simply that the Avenue lined a route from the north to the Ring to the northernmost visible stones, and that individual stones were erected ad hoc. The local orientation of the route may have reflected some astronomically related beliefs, as may the placement of some individual stones; but testing that idea scientifically will be difficult if not impossible. The sizes of the later stones may have been chosen with those of the already erected stones in mind; but my suspicion is that for the most part size reflected the wealth or status of individual or group organising erection.

a unique design’ with no more - or less - mystery in the stone setting than in other sites’ (Henshall, 1972, 138). Although the motives of the builders of prehistoric structures are opaque, some different possibilities can be suggested. Műller linked the cruciform stone setting at Calanais (as well as what he claimed to be the cruciform shape of the chamber), to the cruciform chambers of the Boyne Valley monuments (Muller 1988, 24). Henley (2005) also thought that the avenue may have corresponded symbolically to a chambered cairn passage, and the south, east and west alignments to end and side chambers. This idea has a long history to it. John Stuart put it forward in 1866, comparing Calanais to New Grange (Ponting & Ponting 1984, 36). Downes and Richards, in a different context, have suggested that the furniture in houses in Orkney built by those who used Grooved Ware was laid out in a cruciform pattern (Downes and Richards 2005, 57-59, 126). They suggested that the house interiors were designed according to cosmological principles with each element in the cruciform providing a reference to ‘the key points in the annual cycles, which govern both agricultural and social practices’ (Downes and Richards 2005, 58-9). Given the occurrence of a grooved ware vessel at Calanais similar to that from the Stones of Stenness and some of those from Barnhouse, perhaps similar ideas underlay the layout of Calanais; or the avenue and rows may even have been created to turn it into a symbolic ‘house’. These insights are difficult to assess; the East Row at least was much later than the Ring so while such ideas may have played a part in the final form of Calanais they probably did not determine its initial shape. The same applies to another possibility, discussed above in the context of ritual deposition and below in the context of the Avenue and Rows, that Calanais was a material expression of a cosmological quartering of the area centred on the Ring.

24.15.12 Interpreting the Avenue and Rows

24.15.13 Astronomy and Cosmology at Calanais

Those who seek to understand why the stone setting at Calanais was built must contend with the reductionist view that Calanais is ‘… only the familiar forms of avenue and circle, but combined in

Theories that astronomical knowledge was incorporated in the stone setting run through 17th to 19th century discussions of Calanais. Many Discussion and conclusions \ 1061

Calanais Survey and Excavation, 1979-88

Illus 24.77 The movements of the moon (after Ashmore 1995, 39; developed from a diagram published by G and M Ponting 1984, 45) of them including Somerville (1912), Hawkins (1965), Thom (1967, 1971 and 1978) and Thom, Thom and Burl (1980) were summarised in Chapter 3 (Previous Studies), and they and others have been ably presented by Burl (for instance Burl 1993, 13, 14, 64, 65; 2000, 202-6). I shall review only a few of them here. It hardly needs to be said that Burl’s two major books on stone settings (1993, 2000) are indispensable treasuries, and he is responsible for important pioneering work on systematic studies of orientation amongst regional groups (Ruggles 1999, 130). But in what follows I shall rely most heavily on Clive Ruggles’ ‘Astronomy in Prehistoric Britain and Ireland’ in discussing matters of fact and in making value judgement (Ruggles 1999). I shall also depend on

the wide-ranging surveys and investigations of the landscape around Calanais undertaken by R and M Curtis (e.g. Curtis and Curtis 2008). Their labours have been conducted very much in the spirit of the ‘the pressing need to examine further evidence on the location and design of monuments in relation to the contemporary landscape … in a systematic way’ (Ruggles 1999, 156). Over the years there have been many claims that the avenue indicated the rising of particular stars, and similar claims have been made for the rows; but not one of these explanations for the avenue’s orientation or those of the rows stands up to scrutiny (Ruggles 1999, 136). That said, he did emphasise that ‘We can never know: all we can ever have is a degree of belief in a certain idea Discussion and conclusions \ 1062

Calanais Survey and Excavation, 1979-88

according to the evidence available’ (Ruggles 1999, 76). On the southward orientation of the avenue, he noted that claims of very precise alignments fail because Cnoc an Tursa hides Clisham. His work casts a shadow on claims of precise astronomical or calendrical significance in the lining up of stones at Calanais; but his dismissal of claims that the stones provided scientifically robust evidence for precise alignments was not intended to exclude more general ideas, such as that of the importance of southernmost moon-set to those who lived in the area in the years when Calanais was in use. Although the precise astronomical claims contain errors, in 1999 Ruggles wrote ‘Remarkably, perhaps, we seem to have come full circle, from criticising the inherent biases in the work of Thom and others, through seeking rigour and objectivity, to recognising the shortfalls of that approach and discovering the need to readmit subjectivity as part of a controlled approach which involves ‘continuous dialectic between ideas and empirical data’’ (Ruggles 1999, 162). I continue to distrust subjectivism in any study that purports to be scientific (Popper 1976, 1956), except in dreaming up new ideas to test; and if subjectivism does not allow rejection of non-trivial ideas through as objective tests as are possible I am genuinely unsure what a “dialectic between ideas and empirical data’’ means. I suppose it just means checking your ideas against the data and making sure that the data informs your ideas; but if the criteria for rejection of ideas are purely subjective the approach is as dangerous as it always was. Nevertheless, I have here taken just one of Ruggles’ points in isolation and must stress that I agree with Ruggles’ points far more than I disagree. During much of the 3rd and 2nd millennia BC in Scotland there may have been an interest in the southernmost limits of the moon’s motion. The evidence for such an interest in movements of the moon or the sun is particularly strong amongst the recumbent stone circles of the north-east and the stone rows of the north-west (Ruggles 1999, 109; Welfare 2011). From many of these sites the moon at its major standstill limit in the lunar node cycle moves just above the southern horizon (Illus 24.77; Ruggles 1999, 159). This is a phenomenon which last occurred at Calanais in 2006 and will

occur again in 2024-5 when the moon will glide just above the southern horizon once a month for a few months around the precise lunar extreme. Ruggles did not quarrel with the data established by Gerald and Margaret Ponting and documented more fully by Margaret and Ron Curtis (see below), and speculated that the intention behind the stone setting may have been to create a dramatic relationship between the monument, the landscape and the heavens, reserved for rare and auspicious occasions, a generalisation of an idea put forward by Margaret Ponting (1988) and described below. Neither did he reject the possibility, raised by the Pontings (Ponting & Ponting 1981; Ponting 1988, 426-7) that the various monuments around the main setting formed a complex in which ranges of sacred hills were framed in relation to lunar rising and setting on the horizon. But he did point out that current rigorous methodologies do not test such possibilities, which means in essence that he could not fully accept them (Ruggles 1999, 136). As described in Chapter 3 Previous studies 3.1.2 Toland in his History of the Druids produced the first surviving published claim that the tale told by Diodorus Siculus, quoting Hecateus of Abdera (c 330 BC), of the moon dancing along the horizon in a northern island referred to Calanais (Toland 1726, 188-91). Others have suggested that that reference is to the lunar node cycle of 18.61 years, the gap between reoccurrences of the lunar major standstill. Ruggles however has pointed out that the Diodorus anecdote referred to ‘the period in which the return of the stars to the same place in the heavens is accomplished’ which is the 19.0 year Metonic cycle, not the 18.6 year lunar node cycle (Ruggles 1999, 88-9). Ruggles referred also to the more rarely quoted Strabo, who wrote that the Celtiberians of NE Spain and their neighbours in the north ‘offer sacrifice to a nameless god at the seasons of the full moon, by night, in front of the doors of their houses, and whole households dance in chorus and keep it up all night’. Ruggles thought that if interpreted with caution the writings of Diodorus and Strabo might tell us something about practices in Atlantic Europe in the pre-Roman Iron Age, but ‘it is quite another question as to whether Discussion and conclusions \ 1063

Calanais Survey and Excavation, 1979-88

there is any evidence for a continuity of tradition extending back into the Bronze Age and before’ (Ruggles 1999, 88-9). There need be little doubt that the horizon-skimming phenomenon was known to those who built the stone setting at Calanais, even though there is no evidence for how it fitted with their other world-beliefs. It is well seen from near the north-easternmost stone (Stone 8) of the avenue (Curtis and Curtis 1994, 23). From there the moon at its major standstill sets behind the outcrop Cnoc an Tursa to the south of the stone setting and then reappears momentarily close to the central monolith. M Ponting has suggested that Calanais was specially positioned in the landscape ‘to take maximum dramatic advantage’ of lunar phenomena (Ponting 1988, 424, 427, 431). She observed that if the north end of the avenue had been intended to allow observation of moon-set on the distant horizon, there would have been no physical impediment to building the avenue a few metres further west (Ponting 1988, 429). She also showed that there were other parts of the setting from which lunar events could be seen. She thus suggested a positive intention on the part of the builders to include both the disappearance and the reappearance. Ponting and Ponting’s work has also confirmed that Somerville was right in supposing that a line between the two stone outlying the circle, stone 9 and stone 34, indicates the northernmost full moonrise at its extreme position in the 19.81 year lunar cycle (Ponting 1988, 430) although that the alignment should be regarded as approximate because of the large size and irregularity of the stones (Ponting 1988 fig 19.5). Their work has also shown that the avenue can be used to indicate the midwinter setting of the sun behind a mountain and its subsequent gleam through a notch in the horizon (Ponting & Ponting 1984a, 52). If this phenomenon is due to more than chance it demonstrates an interest in sunset at midwinter which can be paralleled at, for instance, Maes Howe in Orkney around 3000 BC and the cairns at Balnuaran of Clava after 2000 BC (Bradley 2000, 122-5). And as described above (Chapter 24.9.5) the layout of the

the ring at Calanais suggests an early interest in the equinox (rare in Scotland, reflecting the visual anonymity of the event) which may have lasted for several centuries. More recent work by Curtis and Curtis has shown that the stones on the east side of the Avenue can be used in conjunction with a foresight marker for assessments of how close moon-set was to its extreme. They have suggested that this explains the positions of all of the stones in the east side of the avenue except Stone 5, and possibly stone 6 (Curtis and Curtis 2009, 28). That concept is of the sort acceptable to those who believe that sophisticated astronomical observations played a large part in the lives of an elite group, but without demonstration of the existence of a contemporaneous foresight most others will suspend judgement; for accepting that it could have been done does not entail accepting that it was done. As described above (Chapter 24. 14.3), deposition patterns within and just outside the Ring suggest a cardinal quartering of the cosmos. The lines along which the rows were set up may also have reflected cosmological ideas although the direction of the East Row seems more likely to have been determined by the orientation of earlier cultivation beds (Illus 24.10). In this interpretation initially abstract lines divided the whole of the area inside and outside the Ring into quarters. The rows, perhaps initiated generations after deposition started, suggest a somewhat less precise cardinal quartering of the cosmos. These changes seem to fit reasonably well with Pollard and Ruggles conclusions on cosmological ordering in Britain during the third and early second millennia. ‘… solar and lunar-derived cosmological schemes were not mutually exclusive, and … it was through long-term ritual practices that the motions of the moon came to be increasingly referenced through deposition. It is perhaps of significance that those deposits most closely allied to the motions of the moon are of human bone (principally cremations) (Pollard and Ruggles 2001 87). What happened at Calanais will not have been identical to what happened at other stone settings in Britain. But the evidence does corroborate the existence of a belief in cosmological principles similar to those deduced at other sites. Discussion and conclusions \ 1064

Calanais Survey and Excavation, 1979-88

24.16 Sacred and ritualised landscapes The integration of geography with beliefs about the cosmos is widespread amongst historically recorded and modern non-western societies. It is in principle possible that all landscapes in Britain and Ireland were regarded as sacred in that sense in the fourth to second millennia. And as many authors have suggested (Ruggles 1999) it seems likely that past peoples did not see a clear separation between sacred and domestic landscapes but a continuum. 24.16.1 Some ideas about sacred landscapes A sacred landscape is ‘a numinous landscape with an importance that might extend beyond the immediate locality’ (Ritchie 1997, 79). Although Ritchie’s phrase sums up the two most important characteristics of sacred landscapes, I shall rely in what follows largely on Ruggles summary of ideas about them. He declares that they are culturally charged; people experience them against a background of memory and associations; they embody ideas and symbolic meanings; and they help to harmonise people’s activities with the cosmos (Ruggles 1999, 120). Although many historical and modern indigenous groups draw no distinction between ritual and mundane activities (Ruggles 1999, 262 Archaeology Box Notes 1) some places were regarded by most people most of the time as more suited to ritual or ceremony, or demanding more observance, than others. There is a considerable literature in which words like ‘secular’, ‘ritual’ and ‘ceremony’ carry specialised meanings. But such words do have useful ordinary (and thus not very precisely defined) senses. By a ceremonial place, in what follows, I mean somewhere designed for ceremonies, celebrations or festivities involving expressions of belief and carried out by groups of people. In this commonplace usage ceremony and ritual can overlap. What I mean by a ritual centre is a place customarily used mainly for carrying out rituals, or formally interacting with the sacred aspects of life, whether actively or passively, privately or publicly. In contrast to that is a place,

for instance a dwelling or a field, where rituals may well have been embedded in daily activities but the primary goings-on were domestic or agricultural. The place of faith and beliefs in such activities is brilliantly evoked in an episode attributed to AD 1932 in Claude Michelet’s family saga, set in the Correze, France: ‘When all the crops were sown Pierre-Edouarde, following an old custom, fixed a small cross, shaped roughly out of straw, in the middle of each field. Mathilde [his wife] attached importance to this practice. She saw it firstly as a sort of tribute to the land and to the work, and also as a discreet appeal to Heaven, a little wink at the Lord, as if to say to him: ‘Look, we’ve done all we should do, as well as we could; now its up to You to do the rest.’ (Michelet 1994, 79). . A distinction has been proposed between two modes of religious practise: infrequent intense events (imagistic) and frequent low-level ones (doctrinal) (Whitehouse 2000). It is very tempting to suppose that large central sites such as Calanais 1 were used mostly for ‘imagistic’ events and the small rings around it for ‘doctrinal’ ones But I do not in general intend a precise equation of ‘ceremonial’ and ‘ritual’ with ‘imagistic’ and ‘doctrinal’ a formal distinction would not be very useful because the current evidence from most Scottish stone settings is inadequate for exclusion of either mode. The same may not be true of timber settings if Thomas is right in supposing some were purposefully built for destruction by fire, suggesting predominantly ‘imagistic’ events (Thomas 2007). Darvill has argued that a site-oriented approach (in landscape studies in general) has its dangers; the implication is that one should concentrate neither on ‘imagistic’ nor on ‘doctrinal’ sites. Landscape studies should be considered in social terms at larger spatial scales. Subtle landscape features and apparently empty spaces could sometimes have been socially significant. Archaeological surveys do not usually record all the sensory information available. He advocated more attention to assessing what the basis was for actions which did leave traces. He also pointed out that all landscapes may have been in a constant flux but that social constructs may have survived physical change, perpetuated by myths, legends and placenames (Darvill 1997). Discussion and conclusions \ 1065

Calanais Survey and Excavation, 1979-88

Illus 24.78 The landscape to the southwest of Calanais (Mike Brooks, Historic Scotland) These strictures are well justified. However, I am firmly of the opinion that interpretations of the human dimension of past landscapes must be evidence-based if they are to improve understanding of the past. Proposed social dimensions must often remain speculative because there is often no oneto-one correspondence between traces of activities (let alone absence of traces) and the reasons why the activities were undertaken. For field archaeology the most important point is that while built foci may be recognisable the same may rarely be true of the natural, let alone the transitory. In what follows, while bearing Darvill’s points in mind, I shall concentrate on landscapes where there seems to be enough material evidence to suggest that the area was regarded as special. Stone settings represent a concrete expression of some ideas about the spiritual aspects of landscapes and the natural objects in them. Their creation may usually have formalised and normalised pre-existing practises. The buildings did not make the terrain any more or less spiritual. Instead they facilitated,

or for some people imposed, preferred ways of interacting with it. But that will not have happened in isolation from other changes. Ritualisation of the landscape may have reflected increased social group sizes and possibly also contacts with powerful external groups of people. The monuments probably reflected attempts to transmit and reinforce doctrines, strengthening social solidarity. Perhaps frequent visitors tried to strengthen commonalities with local groups by introducing religious beliefs; or local groups may have tried to strengthen religious bonds between each other in the face of disturbing contacts with external groups. In what follows I shall use the term ‘ritualised landscapes’ for areas where it can be shown that cosmological symbolism was important in the design or placement of non-domestic structures. But it is important to remember that such ideas were in a sense tools to a different end; whatever the detail behind them, they probably reflected the use of religion to increase (and perhaps to impose) social coherence (Norenzayan 2012, 43).

Discussion and conclusions \ 1066

Calanais Survey and Excavation, 1979-88

24.16.2 A ritualised landscape round Calanais Observations of the moon dancing over the horizon and of moon-set at major standstill can be made from several stone settings near Calanais; and although the southern horizon has attracted most attention other views have great natural beauty, at least by modern standards (Illus 24.78). The area can be regarded a ritualised sacred landscape. Indeed, using a largely superseded terminology Henshall suggested some four decades ago that the east shore of the inner reaches of Loch Roag had some special sanctity in the same way as had Crinan (the southern extension of the Kilmartin Valley) and the area around Stenness in the middle of Orkney ‘with their concentration of temple sites’ (Henshall 1972, 150). The evidence from the main stone setting at Calanais appears to show that the people there were, at least in parts of the third and the second millennia BC, interested in coincidence of the southern limits of movements of the moon at major standstill with landscape features and the stone setting. Ponting and Ponting (1981, 1984) and subsequently Curtis and Curtis (1993, 2008, 2009) have photographed and surveyed horizon profiles seen from there and many other stone settings in the neighbourhood and shown that at least some of them could be used in the same way. That suggests, in Whitehouse’s terminology, that the setting were built for ‘imagistic’ events; but the current evidence cannot exclude their use for ‘doctrinal’ ones. One aspect of that landscape is that the profile of the distant hills south of Calanais looks to some observers like a woman lying on her back. Further to the west the Clisham Hills profile includes the deep valley of Glen Langdale (partly obscured by a foreground hill in Illus 24.78, but lying directly above the islet in the middle body of water). As viewed from several parts of the area around Calanais at the south lunar extreme the moon rises from the former and sets (after temporarily disappearing behind the hills) in the latter (Ponting 1988; Curtis and Curtis 2008, 4-5). Curtis and Curtis record at least one indication of major standstill southern extreme moonset provided by lines between stone settings; when viewed from

the fallen stone at Druim na h-Aon Chloich the stone ring at Ceann Hulavig is in line with the Glen Langdale valley and thus with final moonset at its southern extreme. They have demonstrated that others of the six rings in the area could have been used for observation of the same phenomenon although paradoxically it is not visible from the main stone setting itself (Ponting 1988). Single stones inserted into the landscape may have marked places important within a melding of beliefs with perceptions of the physical geography. The work of Curtis and Curtis, building on that of Ponting and Ponting, has made a substantial case for there having been a ritualised landscape round Inner Loch Roag, natural but with its symbolic nature accentuated by artificial structures. But how can the specifics of their interpretations be tested? The usual way to investigate such ideas is to form a null hypothesis, that there was nothing special about the placing of the sites, and test it against data. Such approaches have been explored, for instance by Bradley and his colleagues in their examination of the rock art of Galloway (Bradley et al 1993) and of an area around the Clava cairns (Bradley 2000). If the data does not allow acceptance of the null hypothesis then there was something special about the placing of structures. That does not necessarily mean that the reasons for the siting were the same as those supposed when formulating the null hypothesis. There may have been undreamt-of reasons. But testing the null hypothesis is an important first step in assessing whether landscape patterns were significant. In practice, that means comparing the areas around ritualising constructions with a random set of apparently empty areas. If after that there are no better explanations why the settings were where they are in Calanais landscape, then Curtis and Curtis’ thesis will gain more weight. If the areas immediately round stone settings turn out to have few characteristics different from those of other randomly chosen parts of the landscape, interpretation of the placing of the stone settings will be open to many other ideas. Particularly if their precise locations were not strongly defined by the visibility of astronomical phenomena, the positions of the stone settings were presumably linked to other uses of the land. Discussion and conclusions \ 1067

Calanais Survey and Excavation, 1979-88

One model might be that the single standing stones from which striking moonsets could be seen were close to small settlements while rings and other more complex settings from which the same events could be observed were close to larger settlements. Alternatively, if people moved regularly through the landscape smaller stone settings might reflect less used routes and larger ones those used most frequently. Or it might turn out that archaeologically minor sites were stations on twisting landscape-determined routes between more substantial ones. Terrain modelling might allow predictions about where settlements lay, or how people travelled about the area. All that said, Curtis and Curtis have made a very good case that the landscape around Calanais was sacred in the sense defined by Ruggles (1999) and ritualised in the sense used here. It is the preferred working hypothesis. 24.16.3 A variety of ritualised landscapes elsewhere in Scotland What follows is a short version of Appendix 10 Scottish Ritualised landscapes. There may have been ritualised landscapes in Scotland several centuries before Calanais was built. For instance the large chambered cairn clusters of Caithness, at Loch Calder, Loch of Yarrows, Sordale Hill and Dunbeath (Davidson & Henshall 1991, Fig 5) may prove to be more than members of a set of paired cairn and settlement. In Rousay in Orkney too each cairn may represent a territory belonging to a single household or small community. But the first obvious evidence for more complex ritualised landscapes appears in the centuries around 3000 BC. The most prominent ones with dominant stone structures of much the same period as Calanais are at Machrie Moor on Arran, and in the middle of Orkney The monuments around Balnuaran of Clava and the several clusters of recumbent stone circles in NE Scotland, up to a millennium later than Calanais, may be an expression of integration of landscape and cosmos more like that of the early chambered cairns of Caithness. The Crinan/Kilmartin Valley, complex includes both stone and timber monuments and thus pro-

vides a bridge between the stone-based ritual landscapes and the timber and earth ones of the plough-lands of Scotland Examples are found at Forteviot (Noble 2009, 14-17; 2010; Brophy, K & Noble, G 2011), Meldon Bridge in the Borders (Speak & Burgess 1999), Dunragit in Wigtownshire (Thomas forthcoming). and Balfarg/ Balbirnie in Fife (Barclay & Russell-White 1993; Barclay 2005; Gibson 2010). Some of these were probably far more impressive in their time than Calanais. The ceremonial centre at Machrie Moor on Arran may have lain in a landscape of extensive field systems with far more abundant neolithic settlement than has previously been suggested (Barber 1997, 144-145, 149). If the rings of tall stones at Machrie Moor are of much the same date as Calanais they should also be of a broadly similar date to the timber circles on Machrie Moor 1 and 11. This conjunction does not match well with the idea proposed by Pearson and Ramilisonina (1998) that (in southern England) timber structures were built in one area for the living and stone ones were built for the dead in a non-overlapping nearby idea There may have been considerable continuity in the ides expressed at Machrie Moor. The low stone circle on Machrie Site 11 is very probably 250 to 900 years later than its timber predecessor in exactly the same place. It is quite extraordinary; given that ploughing intervened, that the precise site of the timber circle was remembered for at least ten generations, or that precise rules for the placement of structures relative to other features survived that long. Barnatt and Pierpoint concluded that there was no evidence in the siting of the ceremonial complex for an interest in precise alignments on the main solar or lunar events, but that the rings may have been placed to exploit a horizon notch indiDiscussion and conclusions \ 1068

Calanais Survey and Excavation, 1979-88

cating midsummer sunrise (Barnatt and Pierpoint 1983, 112). But there are problems with the methodology of their study (Ruggles 1983, 116; 1999, 133) and their conclusions have to be treated with caution (Ashmore 1999c). As things stand the concept of a widespread ritualised landscape has to be employed to an even greater extent at Machrie than at Calanais. From currently available evidence it seems that people looked towards a single ritual focus rather than imbuing several related locations with connections to a wider cosmos. Around and after 3000 BC a rich society expressed itself in settlements and monuments around and between the southern part of the Loch of Harray and the Loch of Stenness in Orkney. For ease of reference I shall refer to this area as the Heart of Neolithic Orkney (Historic Scotland 1998, 39). It is sufficiently well known and publicised that I shall keep my account very brief; indeed the sometimes indiscriminate use of the archaeology of Orkney in discussions of the prehistory of other Scottish and indeed also English regions has raised protests which merit some sympathy (see for instance Barclay 2000). From an Orcadian perspective also the singling out of this area has been seen as invidious. Much of Orkney contains exceptionally well endowed 4th to late 3rd millennium landscapes and the Heart of Neolithic Orkney may be only part of an archipelago-wide ritualised sacred geography. But the Stones of Stenness, the hall at Barnhouse, Structure 10 at Ness of Brodgar and the chambered cairn at Maes Howe are currently exceptional in illustrating different but related aspects of the beliefs of a coherent social group. The extraordinary buildings at Ness of Brodgar, some seemingly non-domestic, add depth and complexity to this assemblage of ritual and ceremonial sites. And the multitude of large and largely unexplored monuments, and

the division of the landscape by substantial built boundary walls contributes to the case for this area’s being special part of a ritualised landscape. My present impression is that monuments for the living and the dead were set close to each other, rather than set in large and distinctive areas as has been suggested for the Stonehenge Durrington landscape in southern England (Parker-Pearson and Ramilisonina 1998, Figure 7). The ritualised landscapes at Calanais and the Heart of Neolithic Orkney differed considerably from each other despite the existence of tall-stone rings in both areas. Central Orkney contained two large settlements, and several large funerary mounds; and despite the amazing archaeological landscape it did not have (or more strictly speaking, it does not now have) the small rings and stone settings found round Calanais. Nor does it seem to have links with the Beaker-using people who introduced change to many parts of Europe. But it seems to have been very much richer. That the Kilmartin Valley supported what I have here called a ritualised landscape has long been recognised. Not only was it ‘… certainly a centre for ceremony and burial for an extended period …’ (Ritchie 1997, 83) but ‘The concentration of sites and the intervisibility of many within the confined valley floor allow us to conjure notions of a numinous landscape with an importance that might extend beyond the immediate locality’ (Ritchie 1997, 79). The valley contained both timber and stone structures. The latter have been long known, with a linear cemetery of massive cairns, some chambered, along with a henge, standing stones including the decorated stones at Nether Largie and Ballymeanoch, and the stone rings at Temple Wood. It also contains numerous decorated rock outcrops (Ritchie 1997, 77- 84) and has produced fine artefacts, including highly decorated pottery of many periods and jet jewellery (Ritchie 1997, Discussion and conclusions \ 1069

Calanais Survey and Excavation, 1979-88

80-84). Major timber monuments were discovered near Upper Largie in the late 1990’s. They included a long rectangular timber structure; a pit-defined avenue; an elliptical timber ring; a ring complex; Beaker and other graves (Terry 1997, 19-21 fig 8; Ellis C 2000, 16; 2002, 145; Ellis and Crone pers. comm.; Cook et al 2010). The long rectangular enclosure probably dated to the second quarter of the 4th millennium BC. Building of the stone ring at Temple Wood is not radiocarbon-dated but its plan is similar in size and shape to that of the Ring at Calanais so it may be of approximately the same date. It was subsequently embanked and its interior was filled with boulders under and among which were small kerb cairns and burial cairns. Recent radiocarbon dating of cremated human bone suggests burial between about 1420 and 1270 cal BC (Sheridan 2008, 202). The Beakers from Kilmartin show many similarities to Dutch ones. Presumably that reflects contacts with the eastern areas of Britain. Irish Bowl Food Vessels from three cairns suggest long-distance western connections and Whitby jet suggests contacts with NE England in the last few centuries of the 3rd millennium BC; (Sheridan 2008; Sheridan pers.comm). Possibly during the early period of structure-building, and certainly in the later 2nd millennium BC, there is evidence for cosmological interests. Many of the isolated short stone rows and stone pairs of the Kilmartin Valley are, at minus 30°, closely aligned on moonrise or set at the major standstill limit. A smaller number of alignments lie between minus 21° and minus 26°. Each of these is close to a row or pair with an alignment of minus 30°. Ruggles has suggested several explanations for this pattern including the possibility that rows and pairs of the latter group were aligned on midsummer full moon at a random point in the 18.61 year lunar node cycle but when the moon was seen to set further and further south another ‘more correct’ alignment was built (Ruggles 1999, 109). That said, claims of alignments relating to the earlier stone settings at Temple Wood and Nether Largie need reassessment in the field (Ruggles 1999, 59, 231 note 79). The apparent contrast between the location of the timber monuments in the upper part of the

valley and those of the cairns and stone settings downstream raises the question whether the concept of zones for the living and zones for the dead can be applied to the Kilmartin Valley. The sacred landscape was on present evidence created and perpetuated by societies which were materially much richer and more widely connected than those around Calanais. Clava cairns lie mostly along the coast of the inner Moray Firth and the rivers running south with a concentration in Strathnairn (Bradley 2000, 2, 184). The monuments at Balnuaran of Clava include two passage cairns, a ring cairn and a small rough kerb cairn, while in an area of about a square kilometre around them there are at least six other monuments (Bradley 2000, 5, 175-8). The cairn at Balnuaran of Clava was probably built between 1920 and 1740BC (Bradley 2000, 115-6; AA24234, AA-25433), or in round terms about 900 to 600 years later than Calanais. The two passage cairns are orientated towards the point where the midwinter sun sank below a nearby hill (Thom 1966, 18; Bradley 1998, 136, 142; 2000, 122-5). The passages of the two cairns still point at midwinter sunset, although the alignment of the northern one would have been better 4000 years ago, midwinter sunset from southern cairn is currently blocked by trees (Bradley 2000, 122). The more southerly cairn lies on the line indicated by the passage in the north cairn (Bradley 1998, 142). There may also have been orientations on midsummer sunrise but lunar horizon phenomena were not positively indicated. Bradley perceived Balnuaran of Clava as unusual in Scotland, because he thought that indications of interest in solar movements were uncommon (Bradley 2000, 126). By and large Clava cairns each had a local significance; they dominated their immediate localities but were neither placed to be visible from afar nor to have unusually good views outward Discussion and conclusions \ 1070

Calanais Survey and Excavation, 1979-88

(Bradley 2000, 178-80,182). A sample of 25 cairns suggested that their orientations clustered round midwinter sunset and the major and minor lunar standstills (Bradley 2000, 181-2). Thus in the lands south of the Moray Firth (with outliers on the Black Isle to the North) there seem to have been many small ritualised landscapes. In places they were so close to one another that they probably formed a single large one. The pattern is far richer and more extensive than that visible around Calanais. T h e re c u m b e n t stone circles of Northeast Scotland have a distribution largely confined to Aberdeenshire. On the limited reliable evidence currently available, at least some of them were built over half a millennium earlier than the radiocarbon-dated cairn at Balnuaran of Clava. Beaker sherds have been found in some abundance at them- and they were used as cremation cemeteries around 1000 BC (Welfare 2011, 162-167). Welfare has contrasted their distribution with that of large burial cairns, some ‘found deep in the glens that penetrate the mountain hinterlands’ (Welfare 2011, 64-67). They share many traits with other Scottish ‘Bronze Age’ monuments and Welfare, noting how many sites had wrongly been called recumbent stone circles, deprecated the tendency to lump sites together, obscuring the true variety of cairns and megaliths in the area (Welfare 2011, 252, 258). In 1984 Ruggles concluded that there was a highly significant general orientation from the interior of the circles outward over the recumbent stone towards horizon points centred on the SSW, generally avoiding local horizons (Ruggles 1984, S77). He and Burl showed that overall there was no conclusive evidence for precise observations of the sun or the moon (Ruggles and Burl 1985, 578). Welfare (2011, 213) has argued that it is more

likely that the general orientations of recumbent stone circles and Clava Cairns reflect an interest in the setting sun. No strictly contemporaneous settlements have been demonstrated although Welfare has suggested that they had settlements close to them. In that respect they are like the Clava Cairns. The existence of the circles with the interest they demonstrate in the general direction of lunar and solar setting does imply that observance of some contemporary rituals was focussed in particular places and to that extent the landscape was partially ritualised. But despite that there were large areas round them where, whether or not the landscape was thought of as sacred, there is little evidence for its ritualisation. It must be stressed that Illus 24.79 is based on very incomplete information because of the small number of excavated sites and the limited amount of archaeoastronomical survey. I suspect that the differences which it encapsulates will diminish with future archaeological discoveries. For example the position of ‘Early Kilmartin valley’ on the ‘astronomical’ line may be over-cautious. It is based on reservations expressed by Ruggles (1999) about claimed early astronomical alignments. Certainly the orientations of some of the stone rows there suggest that by the second half of the 2nd millennium BC local people had a well developed interest in astronomical events. Also, I have not included the evidence from Lowland Scotland, although the monuments so far excavated, for instance at the Balfarg complex in Fife, at Forteviot in Perthshire, Broomend of Crichie in Aberdeenshire, Dunragit, Holywood and Holm in the south-west represent different patterns again. Nevertheless, even the few examples of ritualised landscapes discussed above suggest that there was no overarching template to their creation. Barnatt (1989, 5) has suggested that stone circles and related monument forms may ‘to some extent form parts of a continuum’, particularly the smaller ones. He thought that small stone circles, ring cairns, kerb cairns and other round structures may, functionally, have expressed very similar ideas in different ways in different places and times (Barnett 1989, 9-13). Welfare expressed a similar idea after his exhaustive study of recumDiscussion and conclusions \ 1071

Calanais Survey and Excavation, 1979-88

Illus 24.79 A variety of ritualised sacred landscapes bent stone circles in northeast Scotland, going so far as to write ‘ ... similarities as might appear to exist are better seen as shared traits in contemporary manifestations of country-wide beliefs’ (Welfare 2011, 259). There is indeed some indication that there were broadly favoured belief systems in Scotland even if local expressions varied. The examples discussed above suggest that there was a stronger interest in building structures referencing the movement of the moon than to the sun, with Maes Howe and Balnuaran of Clava the most obvious exceptions. (although as noted above Welfare (2011, 226-8), preferred to explain their orientation as referencing the sun). Midwinter orientations seem to have predominated over midsummer ones, although how far that is due the relative lack of investigation of timber monuments is an intriguing question given the idea that stone monuments were for the dead and timber ones for the living (Parker-Pearson & Ramilisonina 1998). 24.17 Phase 11a: The Stage 3 enclosure After the Phase 9b Stage 2 earthen enclosure had decayed, and following a phase of ground-working a wall-base (102/804) was constructed outside the bank. It followed a slightly different line from both

the earlier bank and the later wall-base 103. It was not identified close to the Ring. It survived best in the slightly lower areas outside the better preserved parts of the upper wallbase. Illus 24.80 shows the stones left after digital removal of those definitely part of the upper wallbase and those in the modern ditch. Ignoring (for the moment) a possible fragment of a wall-base near the East Row the remaining stones include disturbed stones of the upper wall-base, stones of the lower wall-base, and a few stray stones. Where it survived best the wall-base consisted of two lines of slabs with gritty grey material between them (113 on Illus 24.81). This enclosure appears to have been roughly oval. Its entrance was in the same position as that of the Stage 2 enclosure. A single line of stones crossed it on the line of the interior face of the wall-base. No finds could be assigned specifically to the Stage 3 wall-based enclosure. Interpretation of a radiocarbon age (AA-24968 3575± 45) from a piece of willow charcoal in greasy clay 738 outside the cairn on Area H, together with the inverted pollen zonation of an old turf line 751 and overlying soils, led to a hypothesis that the material forming the northern part of the cairn body was pulled down onto the area Discussion and conclusions \ 1072

Calanais Survey and Excavation, 1979-88

to the north and the stones removed for building elsewhere. Perhaps the stones were taken for the wall-base of the Stage 3 enclosure. It does not seem quite as likely that the robbing was for the Stage 4 enclosure because by hypothesis the cairn was backfilled with material from the third stage one when the massive kerb slab was set up on Area

H. That would allow the age, which calibrated to between 2040 and 1770 cal BC, to provide a terminus post quem for the third stage enclosure. But the preferred interpretation is that the Calanais enclosure had the same function as the Stage 4 enclosure. That function is not clear but may have been a partially ‘domestic’ one, albeit with a specialised purpose. 24.18 Phase 12a to b: The Stage 4 enclosure and subsequent cultivation

Illus 24.80 The Stage 3 enclosure and stones near the east alignment

A new wall base (103) was subsequently built, mostly on the worn down crest of the old bank (Illus 24.82). Some of the stones from this were displaced to form the tumble fraction of stone spread 102 (the rest of which formed part of the second phase enclosure). The latest of the possible floor levels (167) included a sherd of Food vessel ASH 75 (along with three Early/Middle Neolithic Hebridean incised sherds and two possible Beaker sherds), so the Stage 4 enclosure was probably still in use after Food Vessel ASH 75 had reached Calanais, perhaps a few centuries after building of the Stage 1 enclosure. If, as suggested in Chapter 7: Area B, the area east of the Ring was cultivated after clearance of vegetation between 1940 and 1690 cal BC, and the ploughing turned up stones 134 which were then dumped near the east row, that may explain the incompleteness of the excavated enclosure.

Illus 24.81 The best preserved part of the Stage 3 enclosure wall-base Discussion and conclusions \ 1073

Calanais Survey and Excavation, 1979-88

Another possibility is that stones were robbed to build up the cairn when the massive kerb slab in Area BIWX was erected, although the preferred interpretation is that the Stage 3 enclosure wallbase was robbed for this purpose. A non-preferred explanation is that we were misled by the stub wall near the Ring. It is just about conceivable that the Stage 4 enclosure circuit continued much further south to join the stones in the southwest corner of BIII (Illus 23.83). If so access to the chambered cairn would have been through the latest enclosure. However this interpretation is not favoured because of the orientation of the hypothetical wall-base between East Row stones 30 and 31(Illus 24.85). Illus 24.84 combines Somerville’s plan, published in 1912, and a plan of the stones found during excavation in 1980. The overlay of the main stones is slightly imperfect. His plan suggests a southern enclosure and if so its line might once have passed between East Row stones 30 and 31, which in this hypothesis would be later, and returned to the Ring stone south of the passage entrance. But there is really nothing in the resistivity survey or excavation plans to support this suggestion and in particular there was no trace of the south-western part of Somerville’s suggested wall-base or bank at the south end of the south extension. Even so, creation of the East Row might have involved the destruction of earlier features; and if there was a southern enclosure it need not have been the same shape as the northern one.

Illus 24.82 The Stage 4 wall-base and stones near the east alignment

24.18.2 The purpose of the enclosure Perhaps the enclosure served some part in burial rituals prior to deposition of bones in the chambered cairn, or after the chambered cairn went out of use. Or perhaps it was not used directly in funerary rituals and had a semi-domestic function. By far the closest morphological parallel for the fourth stage enclosure at Calanais is nearby on the coast at Dalmore, Lewis. It was almost identical in size and had a lined entranceway (Illus 24.85). Like the Calanais enclosure its back was formed by a pre-existing feature. The excavations there produced many Food Vessel sherds (Sheridan pers.

Illus 24.83 A non-preferred alternative interpretation Discussion and conclusions \ 1074

Calanais Survey and Excavation, 1979-88

Illus 24.84 Part of Somerville’s plan in red superimposed on a plan in green of the upper stones.

Illus 24.85 Calanais Stage 4 and Dalmore the latter after Hunter & Ralston 1999 Fig 5.7 comm.). Loosely ascribed to the Early Bronze Age, its purpose, pending publication of the excavations, is unclear; a specialised and perhaps seasonal use has been suggested because of the paucity of the faunal assemblage and its shaded position (Hunter & Ralston 1999, 85). A loose analogy can be drawn between the Calanais enclosure and a house at Ardnave, Islay

(Ritchie & Welfare 1983 Fig 4). The structure at Ardnave was at least partly subterranean with its wall holding back sand. Part of the early phase was squarish with rounded corners and measured about 6m across, half as big again as the enclosure at Calanais; but in Period 2 the inner area was reduced to a more closely comparable size of 4m by 3.3m. The approximately contemporary structures at Kilellan (near Ardnave) had been so damaged by erosion that their nature is obscure. The closest parallels for the Calanais Food Vessel come from Kilellan and Ardnave where they appear to date to between 2150 and 1750 BC (Chapter 18 The Pottery Assemblage 18.7.11; Ritchie & Welfare 1983 Fig 8; Ritchie 2005, Illus 60-61)). These analogies suggest that the fourth stage enclosure may date to sometime between c. 2150 and 1750 cal BC, and may have had a ‘special domestic’ function. Analysis of other 2nd millennium radiocarbon-dated round and oval domestic and funerary structures in Scotland has not produced any convincing parallels for the Calanais enclosure (Technical note 24.18.2). It is technically possible that the latest stage enclosure, was much later, perhaps belonging in the very late second or early to mid first millennium BC. Hingley has suggested that the placement of a roundhouse in front of the Quanterness chambered cairn, and in front of that at Howe, both in Orkney (Renfrew 1979; Ballin Smith 1994) was intended to control access to the entrance to the chambered cairn at each of those sites. However if the shape of the enclosure at Calanais has been interpreted correctly all stages of the enclosure lay entirely north of the approach to the passage. Or it may have been used for excarnation and ‘use of bones in significant activities related to the identification of the community’ (Hingley 1999, 238-9) or for non-funerary practices because of the potential of human bone for ‘symbolising a variety of concerns central to Late Bronze Age communities (Bruck 1995, 250). The paucity of bone fragments could be explained by the acidity of the soil; but it has to be said that there was no evidence to support either suggestion.

Discussion and conclusions \ 1075

Calanais Survey and Excavation, 1979-88

24.19 Phase 13: Kerb slab erection and secondary cairn Two massive kerb slabs were erected after a long period of robbing and collapse of the cairn. Stones and soil were then dumped behind them. The kerbstone 122 in Area BIWX was found prone. That in Area H (711) was found leaning slightly outwards.

Illus 24.86Â Positions of fallen kerb stone 122, prone stone 148 and kerb stone 711 24.19.1 The kerb slab on Area B In Sub-area BIWX there was a massive prone slab 122 (Illus 24.86). Once it had been removed excavation beneath it and slumped cairn material revealed several shallow slots (Illus 24.89). The highest, 856, overlay the bounding slot 858 of the green clay platform. Two other well formed slots, 852 and 855, lay some 0.25m east of 856

and traces of another (857) were found still further out. Slot 855 was covered by cairn collapse and the others cut it. Slot 855 was interpreted as having held the base of stone 122 when it served as a massive vertical kerb slab. The other slots probably reflect its re-erection so it seems to have fallen and been set up again at least twice.

Illus 24.87Â Sketch view of the slots for the kerb stone and the bounding slot of the green clay platform There was no dating evidence for the first erection of the kerb stone, except that it was later than the green clay platform under the cairn and underlay what was interpreted as cairn collapse. 24.19.2 A massive prone slab 148 on Area BIN Prone stone 148 in Area BIN was similar in plan dimensions to the two kerb stones (Illus 24.86). A pit around its south end suggests that an atDiscussion and conclusions \ 1076

Calanais Survey and Excavation, 1979-88

tempt was made to investigate it at some date well before peat covered Calanais. The attempt was abandoned. 24.19.3 The kerb slab 711 on Area H The massive kerb stone on Area H had been set up in a shallow slot cut into plough soils. It was not vertical; it leant to the north (Chapter 12 Area H Illus 12.28). Deep pockets of ministry grit (706) lay amongst the uppermost secondary cairn stones behind it. It seemed that pressure from the cairn had levered the kerb stone out. The setting-up of the kerb slab must post-date layer 738, which lay below the uppermost plough soil and above green clay. Layer 738 produced a radiocarbon date between 2040 and 1770 cal BC.

constructed cairn immediately behind it and dated to c. 1525 to 1420 cal BC; but that layer was not securely stratified below any prehistoric layers and could have been added at any of several periods before the 20th century. It is also tempting to associate re-building with the sherds of a Food Vessel (ASH 75) found in an upper secondary layer of the cairn; but that layer is interpreted as made of material gathered up from nearby areas so the pot may well have been residual. 24.19.4 Interpretation and speculation The most conservative interpretation is simply that the kerb slabs on Areas B and H were part of remodelling of the cairn, probably in the 2nd millennium BC. More speculatively, it may be that the robbing of the cairn in this area was undertaken partly to build the third stage enclosure wall-base. The stones of that enclosure were in turn robbed and provided the fill of the secondary cairn behind the massive kerb slabs. Manoeuvring the slabs will have required a non-trivial effort. But if the intention was to emulate a kerb cairn the work was very incomplete. Perhaps instead the work was done in conjunction with building of the fourth stage enclosure. The slab on Area BIWX could be seen as forming a back wall. But that does not explain the slab on Area H. 24.19.5 Dating and finds from the secondary cairn

Illus 24.88Â The kerb stone on Area H from the north [Film 1981.5.30] It is tempting to associate erection of the slab with grain found in the topmost layer of the re-

The secondary cairn dates to the 2nd millennium BC and iits discussion here is thus somewhat out of chronological sequence. Details of the finds from the secondary cairn and chamber wall are tabulated in Appendix 12. They are summarised here. Secondary cairn fills had 10 catalogue entries for certain, probable or possible E/MN corky sherds, one catalogue entry for a corky E/MN or Beaker sherd and another for a non-corky one, one for a fine Beaker and 11 for the Food Vessel ASH 75 (Illus 24.89). There were also several pieces of quartz and one piece of flint. No glass or modern ceramics were found, in this way contrasting quite sharply with chamber contexts. Discussion and conclusions \ 1077

Calanais Survey and Excavation, 1979-88

Illus 24.89 Food Vessel ASH 75 Secondary chamber wall contexts had 35 catalogue entries for certain or probable E/MN corky sherds and 2 for probably E/MN non-corky sherds. There were 2 entries for E/MN non-corky or domestic Beaker sherds. Other finds were quartz and glass. The two glass finds came from under the inner face wall stones of layer C of context 769 and suggest that the upper wall face had been re-set or rebuilt after peat was cleared from the area. It is not possible to assess what proportion of the secondary fills of the cairn and chamber wall came from fill left on the cairn after stone-robbing and what proportion was thrown back onto the cairn when the massive kerb slab was emplaced or after peat was removed from the top of the structure. Certainly some soils and clays were added when or later then when the kerb slab was set up, as were the Food Vessel sherds and charred cereal grains. 24.19.6 The deposits in the chamber Prehistoric pottery was almost completely absent from chamber fills (the only possible fragment, 81.244, was not seen during pottery cataloguing)

and there were abundant Victorian to modern finds (Appendix 12). All of the original deposits inside the chamber were removed in 1857; or at least they had completely gone by the time of our excavation. As detailed in Chapter 4 (Introduction to the Fieldwork) Sir James Matheson enclosed with his letter about excavation some minute fragments of bones found in the chamber, and a specimen of a black unctuous substance in which these fragments had been contained (Innes 1860, 110-12). They were identified by Professor Anderson of Glasgow as human and probably subjected to the action of fire (Innes 1860, 112). There was no mention of any pottery. It is conceivable that some of the deposits found outside the passage in Area BV were dumped there by Sir James’ workmen, although the stratigraphy disfavoured this possibility. Hingley (1999, 2389) has suggested that the contents of the chamber may have been cleared out in the ‘late Bronze Age’ (the late 2nd and early 1st millennia BC), on the basis that there seems to have been a renewed interest in the chambers of chambered cairns during that period. The stratigraphy at Calanais does not preclude that possibility but because Matheson’s workmen found some cremated bone in the chamber any earlier clearance cannot have been complete. A charred heather twig from deposit 139 overlying the remains of burials outside the cairn produced a radiocarbon date (AA-24957) between 1940 and 1690 cal BC. Because the charred twig was very fragile it seems likely to have been contemporaneous with its context, so both a ‘late Bronze Age’ clearance of burial deposits and dumping by Matheson’s workmen seem unlikely. Very small crumbly pieces of cremated bone (human or animal) were found in a variety of contexts at Calanais, most in the slots and related features cutting the green clay platform under the cairn and a few in later contexts including plough soils (Chapter 15 Cremated bone). One or two were in late contexts and might relate to clearance of the chamber deposits in 1857. The pieces were so small and eroded, and the assemblage overall so small, that the significance of the bone fragments even if identifiable as animal or Discussion and conclusions \ 1078

Calanais Survey and Excavation, 1979-88

human seems likely to be deeply ambiguous; the fragments appear to be degraded, and I believe their examination should await advances in sample preparation techniques. 24.19.7 Late modifications to the chamber Both the drawing prepared by Captain Thomas in 1857 or slightly later updating Palmer’s plan (Chapter 3 Illus 3.28) and the sketch included with Innes’ 1960 account of the peat-clearance (Chapter 4 Illus 4.8) are stylised but both show a neat square back-end. Sharbeau’s sketch (Chapter 3 Illus 3.35 and Chapter 4 Illus 4.7) does seem to suggest a flat back but it is largely obscured by the western orthostats. As discussed in the Introduction to the fieldwork (Chapter 4) Sharbeau’s sketch may have been later than some (unrecorded) ‘tidying up’ of the cairn although Sir James Matheson the proprietor averred that the chambered building was quite undisturbed with even the uppermost tier in its place and that the superincumbent material was ordinary peat (Innes 1860, 110-12).

rounded). The north wall of the end chamber is at a completely different angle to that implied by Sharbeau’s sketch and the drawing prepared by Captain Thomas. They look like the consequence of a casual repair job. We found no evidence showing that there had been a square back end before peat grew over the site. The small amount of stonework above the original basal courses looked like that planned by RCAHMS and was of very poor quality. Nevertheless some credence has to be given to the early accounts and it must be supposed that the back end of the cairn was rebuilt either in prehistory or immediately after peat was removed from the cairn. 24.20 Phase 14 to 15: Late ploughing and abandonment Ground working including ploughing with an ard was a frequent activity at Calanais. The setting may have been used for agriculture several times, although some phases of ground working seem rather to have been part of a long-term strategy for dealing with ritual deposits. If the preferred dating for the end of the last phase of use of the enclosure, some date in the period 2000 to 1800 BC, is accepted then subsequent ploughing and use of the ground for pasture may have been the main activities for about a thousand years. 24.20.1 Phase 14: Late farming

Illus 24.90 RCAHMS Field drawing of 1923 NMRS RCD/13/12 The RCAHMS field plan of 1923 (Illus 24.90) is the first which can be trusted in its detail. It shows an irregular back end (certainly not

The barley and wheat grains found in the topmost secondary layers of the cairn on Area H were dated to between 1525 and 1420 cal BC (See Chapter 23, Radiocarbon). Given the widespread evidence for ploughing in various areas at Calanais 1 it seems likely that this cereal-growing and charring was local. The dates from the grains loosely match that of a peak in cereal pollen in the pollen columns at Calanais Leobag (see Chapter 21, Palaeoenvironment). They also hint at the date of some of the destructive ploughing in and around the Ring after the cairn had become dilapidated – although, of course, this must remain a hypothesis given their context. That date however would not be unreasonable, falling in the period after use of the latest enclosure had ceased. Discussion and conclusions \ 1079

Calanais Survey and Excavation, 1979-88

24.20.2 Phase 15: The growth of peat Combining a date for basal peat in Area C and an estimate from the pollen there, the true date for the start of peat growth at Calanais probably lay between 920 and 400 cal BC, the earlier limit from the pollen zone 3d to 3e transition date and the later from the direct radiocarbon date GU-1403 (see Chapter 23 Radiocarbon 23.14.2). 24.21 Phase 16: Possible disturbance of the cairn in the first half of the 1st millennium BC The peat around the chambered cairn may have been about 0.5 to 0.75 cm deep at the time or times between c. 100 and 500 AD when possibly late Iron Age pottery found its way into some apparently pre-peat contexts This ‘Plain Style’ pottery came into use around the middle of the 6th century AD and continued in use after Viking pottery appeared in the Hebrides during the 10th century (Chapter 18 The Pottery Assemblage 18.8). But if the cairn itself was better drained than the surrounding ground, peat would have been slower to start accumulating on it. It seems probable that it was still vis-

ible then. If building stone was in demand in a mainly peat-covered environment it would have been simple to dig away peat and rob the cairn. That may, for instance, explain the damage to the facade stone on the south side of the passage and the absence of a complementary facade stone on the north side. If there was robbing for building stone in the second half of the 1st millennium BC that may also help to explain why Sir James Matheson’s workmen had to trench the south side of the cairn in AD 1857 to find the kerb, although the main robbing of the cairn had probably taken place in the early 2nd millennium BC. But this explanation is not very satisfactory because the pottery, along with a hodgepodge of distinctive earlier sherds, was found in the fill of the pit 859 for the facade stone. The sherds must have been put in it when the facade stone was erected (or, conceivably, re-erected). If the date for the Plain Style pottery is correct I have no sensible explanation for how it got there. 24.22 Phase 17: Continued Peat growth and post-medieval peat clearance Judging by the peat-etching on the standing stones at Calanais (Pitt-Rivers 1885) there had

Illus 24.91 Kerr’s Excellent Sketch, the engraving which accompanied Callender’s 1857 article Discussion and conclusions \ 1080

Calanais Survey and Excavation, 1979-88

been a fairly long period during which the surface of peat stood at about 1.5m above present ground level. If it had grown at an even rate from say 750 BC to 1850 AD, reaching a depth of 1.5m it will have accumulated at about 5.8 cm a century. But there is no reason to suppose that growth was even. At Calanais Fields the lower 45cm of the 84cm deep peat sampled in Trench 5 ( Johnson et al in prep) grew between 390 to 180 cal BC and cal AD 340 to 540, a period of 500 to 900 years; the upper 39 cm grew between cal AD 340 to 540 and cal AD 1640 to 1950, a period of 1090 to 1610 years. So the lower peat there probably grew up to twice as fast as the upper peat. In round terms the Calanais Fields peat grew 74 cm between 390 to 180 cal BC and 1640 to cal AD 1950. That suggests a somewhat slower rate of growth than at the main Calanais stone setting. The difference in topography may have been responsible. The CN3 column at Calanais Leobag covered a period ending between 1410 and 1110 cal BC so no useful comparisons can be drawn, except perhaps that the lower 13.5 cm of peat there seems to have grown more slowly than the upper part. The CN1 column at Calanais Leobag grew 30 cm between 1800 to 1200 cal BC and 850 to 200 cal BC. The vagueness of the dates makes peat growth rate estimates very imprecise, so no useful comparison can be drawn (Chapter 21 Palaeoenvironment). 24.23 Phase 18a: 19th century excavations and presentation The interior of the enclosure was damaged by post-peat pits. It is possible that some of them were caused by the activities of the Danish archaeologist Worsaae in 1846 (see Chapter 3.3 Previous Studies). He dug test pits and noted the presence of iron panning. When we excavated iron panning was common in Area B but not obvious elsewhere at Calanais during our excavations. Worsaae’s account suggests that the holes were dug by the bases of standing stones (Chapter 3: Previous Studies Illus 3.19). Kerr’s ‘excellent sketch’ of the period immediately preceding clearance of the peat in 1857 (Illus 24.91) does not show peat disturbance in this

area but if Worsaae had thrown peat back into his sondage it may have healed over during the intervening decade. However it seems most likely that the pits were dug in or after AD 1857 when Sir James Matheson, the proprietor of Calanais (indeed, since AD 1844 of the whole of Lewis), encouraged by the Society of Antiquaries of Scotland, had the last of the peat cleared away (Callender 1857, 383; Innes 1860, 110). But it is not clear exactly what the workmen, directed by Sir James Matheson’s chamberlain Mr. Donald Munro found when they removed the last of the peat from the Ring. The main record comes from a letter which Sir James wrote to the Society of Antiquaries of Scotland (Innes 1860, 110-112). In it he described the chamber “... an erection which proved, as the work proceeded, to be the walls of a chambered building, consisting of three compartments. …” saying that “It is remarkable that the sides of the small chamber are quite undisturbed - not a stone even of the uppermost tier removed from its place …”. It sounds, with hindsight, that he was describing the chamber wall in isolation from the cairn and from the passage. It would otherwise be hard to reconcile his claim that it was undisturbed with what we found during excavation. Our excavation showed that the northern side of the cairn was subsequently capped with turf, that the south side of the cairn was subjected to trial trenching before being rebuilt, and that repairs were made to the chamber wall-faces more than once. Late 19th century photographs of the Ring such as those taken by Valentine’s in the St Andrews University Special collection show many loose stones lying around. Our excavations suggest that many of these were collected up into a rough platform which formed part of the base for Lady Matheson’s Path. It also seems likely that much of the area was returfed - or maybe covered with a layer of soil and seeded. There may also have been attempts to stabilise the surface stones of the fourth stage enclosure using turfs dug from an area with a high grit component indistinguishable from earlier grit from decomposed stones. Many of these repairs may well have been carried out for Sir James before 1882, but records do not seem to survive.

Discussion and conclusions \ 1081

Calanais Survey and Excavation, 1979-88

24.24 Phase 18b: modern conservation and excavation Less vigorously intrusive conservation measures continued after 1882, when Lady Matheson passed Calanais into State Guardianship, selling it to the State for a nominal sum. Low key conservation, now under the aegis of Historic Scotland, continues to the present day. Over the last century and many routine records of what has been done have disappeared in conscientious attempts to weed out unnecessary paper. The excavations described here are undoubtedly the most destructive activities undertaken at Calanais in the last century and I trust that this record will survive better. 24.25 Conclusions I am left with the frustrating knowledge that we recorded only a small fraction of what happened at Calanais during past millennia. The patterns we did observe were often ambiguous. One obvious example is provided by the slots and the ground into which they were set under the cairn; why, as they survived, were the features to the north so different from those to the south? The clays were interpreted as part of a platform and the slots as part of light timber structures. How long did they last and how often were their components replaced? The limits are provided only by construction of the Ring and construction of the chambered cairn. If the structures were used for a substantial proportion of that period they would have had to be refurbished many times. But the duration of the enclosure stages and the possibility that each stage represented several minor modifications is equally problematical. And even if we did distinguish all the major revisions to the observed structures, we undoubtedly did not capture the activities of individuals. On a slightly different tack, we did not use some sampling and analytical techniques which are now commonplace. For instance, reverting to Baillieâ&#x20AC;&#x2122;s (1999) ideas of environmental catastrophes leading to changed ways of life, we could have asked someone to check samples for occurrence of particles of volcanic glass. It is a dry compensation that our excavations left much evidence untouched for the future.

Some 18th and early 19th century writers used the monument to bolster ideas about religion and race. Religious, ceremonial and astronomical themes predominated in the later 19th and 20th century literature, with variations echoing the preconceptions of the periods in which they were expressed. The current fashion seems to be that major monuments encapsulated beliefs woven around the local environment. As with interpretations offered in previous centuries, that probably says as much about prevailing worries than about prehistoric Calanais. It is tempting to interpret Calanais using ideas about shifts in social structure, to talk about collective works and communal actions, segmentary societies, chiefdoms and states, administration, social ranking and other facets of how people organised themselves (Renfrew and Bahn 2008,200-215). But there is currently too little relevant evidence from the Calanais area for that to be worthwhile. Speculation would be ill-bounded. Current attempts within British archaeology to refocus attention on regional and local interpretational frameworks have many merits (albeit reflecting curiously similar current political ideas). Nevertheless the building episodes at Calanais did not arise as isolated phenomena but, as with similar centres elsewhere in Scotland, within the context of contemporaneous developments elsewhere (for the general argument see for instances Kinnes 2004, 139, Brophy 2006, 39 and Beek 2011, 25, 43-45). Judging by the structures and artefacts found at Calanais, local communities were well-connected to those in some other parts of Scotland (and possibly to parts of Ireland and England) from some date between 2950 and 2850 BC to perhaps some date between 2000 and 1800 BC. The excavations described here allow the suggestion that the Ring was the focus of one of the several regionally important ceremonial complexes built in Scotland between 3100 and 2600 BC, including Kilmartin Glen, Machrie Moor, Stenness/Brodgar, Broomend of Crichie, Forteviot and Balfarg. There is at least one alternative interpretation. The Calanais complex as presently understood seems to have been much poorer than other complexes such as Kilmartin, and they are dwarfed by some of the earthworks and timber settings Discussion and conclusions \ 1082

Calanais Survey and Excavation, 1979-88

of lowland Scotland. One could envisage Calanais as merely a staging post for people travelling along the western seaways, a place where religious and other practices could be expressed by a transient population. The presence of a locally made Grooved Ware pot and the complete lack of other Grooved Ware sherds would fit this idea quite as well as other notions. But the interpretation preferred here is that the Ring and cruciform setting at Calanais was built and used by local people under the influence or direction of travellers along the western seaways, for the only potsherds in levels roughly coeval with the Ring were from Hebridean incised wares. It is worth remembering that the latter were part of an assemblage of Neolithic pottery including types found in Orkney. The fact that they subsequently deposited soil and clay from ancestral settlements inside the Ring is more ambiguous; it implies a sense of local ownership but that might have reflected a wish to take possession of a structure created by others. The evidence for a belief in a cosmos which could be divided into quarters is fairly strong. The vast preponderance of ritual deposits was in the southeast part of the Ring and to the southeast of its entrance. And although it is possible that the East, South and West Rows were built piecemeal, the preferred idea is that (apart perhaps from the easternmost stone of the East Row) they were built at a single time to make material an ideological quartering of the universe. The preferred interpretation of the Avenue is that standing stones lined the sides of a pre-existing approach to the central complex. Perhaps the stones were set up by affluent individuals to help announce or cement their status. Thus the â&#x20AC;&#x2DC;permanentâ&#x20AC;&#x2122; feature may well have been the route, while the stones each reflect ephemeral events. A good case has been made by Ponting and Ponting and Curtis and Curtis that the Calanais complex was (in the terminology used here) part of a ritualised sacred landscape. Ideas about the sacred seem to have led to the placing of many stone settings, influenced by observations of the

rising and setting points of the moon. Those beliefs probably did not involve precise astronomical observations in the sense that we use those words in current Western societies, nor perhaps did people rely on them for precise calendars. But it seems certain that they took a more than passing interest in the movements of the moon, that they were aware of the cardinal points as special parts of the sky and, at a more subtle level, they may have discerned the equinoxes as special times of the year. It looks as if the main setting was not much used after about 1800 BC. It may have served as a familiar landmark but there is no evidence from the place itself for undeniably ritual or ceremonial activities after that time. Perhaps however the focus of activities leaving material traces shifted only slightly; our investigations omitted many parts of the site; excavations by others have shown that there may have been a small artificial platform or burial cairn at the south end of the South Row. Or the foci for activity may have become more widely disseminated while still being influenced by earlier monuments. Excavations at Breasclete, about 1.5km away, have demonstrated the presence of a burial cairn roughly on the extended axis of the main Calanais stone setting. The current lack of data need not persist long if recent fieldwork takes less time to publish than the excavations reported here. Excavations nearby by Coles, Cowie, Curtis and Curtis, Johnson and Flitcroft, Neighbour, and Ponting and Ponting, have demonstrated that the area round the main setting contained agricultural systems and cairns as well as the long-known stone settings, although no domestic sites demonstrably contemporary with building of the main setting have been discovered. Further, the area, including the sea floor in East Loch Roag, clearly has good archaeological and palaeoenvironmental potential. It has probably not suffered the same amount of marine erosion as the machair plains of the western seaboard and there is an excellent likelihood that sites of all periods survive under the peat and below the sea. Some ways forward are discussed in Appendix 13: Looking to the Future.

Discussion and conclusions \ 1083

Calanais Survey and Excavation, 1979-88

25. Notes In these notes the convention used for labelling sections is different from that of the rest of the report. The headings like ‘25.1’ each cover a chapter in the rest of the report. But the subordinate headings have the same numbers as the section of the main narrative in which their subject is most important. That number is used for all references to the note in earlier and later sections. 25.1 Notes for Chapter 1: Abstract There are no Notes for Chapter 1 25.2 Notes for Chapter 2: Introduction There are no Notes for Chapter 2 25.3 Notes for Chapter 3; Previous Studies There are no Notes for Chapter 3 25.4 Notes for Chapter 4: Introduction to the Fieldwork Note 4.4: Excavation technique for widespread soils Where we did attempt to investigate soils 112 and 117 by box sectioning the consequences were often negative. Often the box sections did not resolve relationships unambiguously. In addition, they cut through lower levels making it more difficult to understand the latter once their remains were more fully exposed. With hindsight it might have been better to have resisted the temptation to use box sections. They were particularly unhelpful where ‘later’ levels of these pervasive soils were the results of diachronic soil processes rather than anthropogenic features. 25.5 Notes for Chapter 5: Resistivity Survey There are no Notes for Chapter 5

25.6 Notes for Chapter 6: Area A There are no Notes for Chapter 6 25.7 Notes for Chapter 7: Area B Note 7.4.8 Labelling of Section F The labelling of Section F (Section 81/7) may have been misleading. The soils above and below black layer 802 should not both have been labelled 117 even if they were superficially similar. The description of 117 written on the section was ‘pinkish, clayey, some roots’. That is quite different from the general description of 117 as ‘ Red-brown fibrous peaty/rooty layer covering the whole site except the enclosure 120 and ditch fill 138’. That said, when moisture content varied the colours and textures varied too, so the difference in descriptions does not on its own prove a real original difference. The layer labelled 132 on Section 81/7 ‘rich brown, clayey, less grits, less roots [than 117]’ had a completely different colour and texture from that labelled 132 on Section 11 Part D ‘green sandy yellow clay’. The Context Sheet has 132 as ‘yellow clay/sand with small grits in which grey ard marks area cut’ which is not incompatible with the labelling of Section 11 Part D but is incompatible with that on Section 81/7. The soil labelled 132 on Section 81/7 might better have been compared to 112 which was ‘a red-brown clay firm soil’ (Context Sheet 112). Indeed, that would fit the hypothesis that parts of 112 were material ploughed down from the second stage enclosure bank. Be that as it may, layers 132 on sections 81/7 (brown clay) and 81/11 Part D (green sandy) are sufficiently different to demand separate context labels. Note 7.6.6 the stone spread 102 in BIN Although analysis of the medial baulk sections suggests that the absence of Stones 102 in this area was real I have a distinct memory that the Notes \ 1084

Calanais Survey and Excavation, 1979-88

topsoil in this area was removed very fast by an energetic person with a pick and shovel. At the time stones 102 were thought of as tumble. It is therefore possible that some stones were removed with topsoil. Note 7.6.8 Conflict in context labelling of Finds 164-166 Cat Find

Ctxt

Description

E/MN or Non-corky 784 81.164 BI 810.1 Chalco/ or dom Beaker EBA 294 81.165 BI 810.1 E/MN

Heb Inc

567 81.166 BI 810.1 EBA

Food Vessel

Finds numbers 164-6 are marked on plan 46/81 of the interior of the enclosure, dated 12/5/81 (the day before the date of 13/5/81 on the bags for small finds 164 and 165). The context is 836, a khaki strip near the N bank of the enclosure. One sherd is Heb Inc, one is FV and one is E/MN non-corky or Beaker. The Finds List and (I believe) the bags have BI 810, and independent of other problems either BI or 810 must be wrong. I had â&#x20AC;&#x2DC;solvedâ&#x20AC;&#x2122; this problem by renaming the context 810.1 and interpreting it as a variant of 812, just south of the ditch which ran from the mouth of the chambered cairn. But in truth it is impossible to know which of the records was correct. Note 7.9.2 Layer 132 A layer on Section 81/07 was labelled 132 and described as a rich brown clay. It was stratigraphically similar to 112/130 . It must not be confused with a different layer, labelled 132 on Section 11/80 Part D during excavation, but relabelled 132.1 during post excavation, which was ard marked green sandy yellow clay merging with a mottled green-brown sand in 1980 - it seems highly likely that the numbering of 132 on Section 11/80 Part D was an error for 123.

Note 7.13.3 Packers at the base of Stone 30 Plan 21 (DC38055) had an annotation that Stone 30 was shown inaccurately, and the packers will have been recorded relative to it. The plan appears to show the southernmost socket lay just west of other packing stones corresponding to the southernmost packer on Plan 19, but the depiction was ambiguous. Note 7.13.6 Interpretations of the stratigraphy round East Row stone 31 Interpretation 1 The marginally preferred interpretation of the record is that a pit and ramp were dug for erection of the orthostat through soils derived from ploughing down of the earliest enclosure and the turf line 157 formed over their fills. 1. cultivation beds; 2. level ground-working; 3. ploughing down of the early earthen-banked enclosure and soil spreading, forming soil 112; 4. digging of a pit and ramp; 5. erection of Stone 31 and backfilling (leaving packing stones proud?); 6. formation of a turf line 157 running up to the packing stones; 7. collapse and spreading of the turf element of the third stage enclosure; 8. ground-working forming 141 and perhaps the lower part of 117; 9. placing of a local wall base (stones 116) possibly with a turf superstructure; 10. collapse of the turf element of the wall 116; the turf subsequently formed part of soil 117 and modern topsoil.

Notes \ 1085

Calanais Survey and Excavation, 1979-88

Interpretation 2 The levelness of turf line 157 could suggest that it had been cut through by the pit for Stone 31. Had it formed after the pit was dug one might have expected it to lap up over clay around the packing stones. Thus a possible sequence in this area is as follows. 1. cultivation beds;

4. ploughing down of the early earthen-banked enclosure; 5. formation of a turf line; 6. dumping of weeds and clearance debris forming layer 160.1 (possibly actually the lower part of soil 141);

2. level ploughing;

7. spreading of plough soil from further north to cover 160.1;

3. digging of a pit unrelated to Stone 31;

8. erection of the stone with packers;

4. ploughing down of the early earthen-banked enclosure;

(Then as 9-10 above.)

5. formation of a turf line 157; 6. erection of the stone with packers; (Then as 7 to 10 above.) In this interpretation erection of East Row stone 31 took place after a turf line had developed over ploughing associated with flattening of the early enclosure. Elsewhere putatively contemporaneous turf lines seem to have survived only in depressions where they had been covered by material from collapsed turf walls or another plough soil. In the case of turf line 157 which was directly overlain by soils 117 and 160 it may be supposed that it was covered by soil-like material spread from collapse of a turf structure before any further ground-working took place. Interpretation 3 However, it is impossible to rule out a variant sequence, with digging of the stone-pit and erection of the stone after formation of layer 160.1, with packers pressed tightly into the side of the stone-pit. 1. cultivation beds; 2. level ploughing; 3. digging of a pit unrelated to Stone 31;

If this interpretation is preferred, and it is supposed that the pit was originally dug through soil covering the turf line, it provides no closing date for erection of the stone, for the only direct evidence for its history is that the area was subsequently ploughed and organic material, stones and charcoal were laid or dumped on it. Interpretation 4 There is, alas, yet another possibility. If soils around and above the pit had been completely removed some time after the stone was erected, and the turf line then formed on the surface thus exposed, that would explain the relationship of the turf line to the packers on Section 57-58. One obviously possible reason why a fairly large area of surface layers round the orthostat should have been removed is turf-cutting for building structures. There is still a fair likelihood that the layer marked 112 on Section A was formed from ploughing down of the first enclosure, but if that is rejected as unproven it would allow erection of Stone 31 before the building of the first enclosure. As a working interpretation requiring testing by future excavation I prefer the first interpretation which sees Stone 31 being erected between building of the first and second enclosures, and my second preference is for the looser interpretation which still sees it being erected some time after ploughing-down of the first enclosure bank. But I have to stress that the other two possibilities are not incredible, and the fourth interpretation allows Notes \ 1086

Calanais Survey and Excavation, 1979-88

one of the enclosures (perhaps the earliest) to be later than erection of the orthostat. In all of these interpretations the isolated (when excavated) stretch of wall-base 116 could well be of much the same date as the second or third phase of the enclosure although it seems unlikely that it actually formed part of either of them (see Chapter 7.4 and Chapter 24: Discussion).

Note 7.15.4 Excavation of the passage features

The basic sequence near the west section of Area BV (incorporating the southern part of BIV) was recorded in the Day Book as in the illustration above. But the context number 160 had been used in 1980 to refer to dark layers in BIII (renumbered 160.1 in the final form of the context list), and in BIV in 1980 it was used for a dark layer intercalated with 139 (renumbered 160.2 in the final form of the context list).

The excavation of the features under the north wall of the passage took place mainly in extremely rainy weather; at any one time most of the passage was covered by polythene sheeting. Most plans were therefore overlays covering only parts of the passage. Later plans were orientated at 7 degrees to earlier ones. In this report the plans have been overlain on a base plan of the passage wall facing stones to make for easier understanding. There were in some cases errors in relative placement of up to 0.15 m between features from different plans. Slot 883 does not seem to have been formally planned as a single slot but a sketch plan was made on Context Sheet 884 and in the Day Book (page 77). The clearest overall depictions are provided by Film 1981.18.18 and Film 1981.19.11.

Table 25.1Â Subdivisions of label 160

Note 7.15.6 West sections of BIWX

Label Subarea

Working from digital files it was not feasible to combine the overlays completely accurately with the main section, and in any event the digital overlays were restricted enough in their detail that some license has had to be taken. However I am confident that the results are not positively misleading.

Note 7.14.1 Layer 160

160

BIV, BV

160.1 BIII

160.2 BIV

160.3 BVWX

160.4 BVSX

Description

Retained for 1981 layer 160 in the stratigraphic position depicted in Illus 25.1

Dark layer dug in 1980 notably near East Row Stone 31

A dark layer dug in 1980 intercalated with 139 (in its original sense of a group of layers) near the top of the sequence in the west baulk.

The layer immediately outside the passage south of the ditch and in the passage, at the same level as 160 and sometimes still called 160

A dark layer in the southern part of trench BV

Illus 25.1Â Interpretative section of Area BV [Day Book 31]

Notes \ 1087

Calanais Survey and Excavation, 1979-88

Note 7.15.7: Pollen in slot 883 Three pollen samples were taken from the slot (2009, 2011 and 2020). The samples reflect vegetation near Calanais Leobag straight after the clearance of birch wood in the middle of CaN-3a (see Chapter 21: Palaeoenvironment). Pinus pollen was absent or barely present. Non-tree pollen values were dominated by grasses. Heather pollen varied between 7% and 15%. Sample 2020 should possibly be placed somewhat later in time than the others, during zone CaN-3aii which started sometime between 2770 and 2360 cal BC and ended sometime between 2560 and 2200 cal BC. Variations between the pollen samples could suggest a heterogeneous origin, which fits in well with the interpretation that the slot contained much old material. The samples may have contained pollen of the period when the slot was filled (after the Ring stones were erected but before the cairn was built) and also remnants of pollen from a pre-Ring period of activity in the late 4th millennium cal BC suggested by the radiocarbon dates. However, the preferred interpretation is that pollen grains contemporary with the movement of the clay greatly outnumbered residual pollen grains from the period indicated by the radiocarbon-dated charcoal, and fairly reflects the period of slot-filling.

the resulting rise in ground level is subsequently truncated by ground-working. The stone-pits at Calanais presented their own versions of the problems associated with interpreting turf lines. I have built several wooden structures (without using concrete) over the last few years. When erecting posts, I seek to jam stones tightly between the pit edge and the post. Then I fill the gaps with whatever I had dug up to create the pit and stamp it well down. The stones often peek above the final surface. The larger the packers the more tendency there is for the topmost ones to stick up. Patchy turf then forms round and over the packers. From this direct experience, and to a lesser extent from other observations, I believe that if a pit is dug through a thin clay turf, and the pit is over-filled with stones tightly packed against the turf, and then clay is put round the packers and a new turf line grows right up to them, one could not tell through normal archaeological techniques whether the pit had been cut through the lower or the upper surface (Illus 25.2).

Note 7.17.3 Much of the description of the Ring in Area B is about turf lines near to it. Turf lines have distinctly different characteristics from most anthropogenic layers because they form over a period of time and disturbances to them can heal over. If a pit is dug through a turf line and backfilled new turf will grow over the fill. It may soon become visually and tactilely indistinguishable from the older turf around it. Then the pit may seem to be completely earlier than the turf line, where in fact it dated to some date within the overall period during which turf grew. This can be particularly confusing if for some reason turf does re-form over one pit before it is covered by a subsequent layer, whereas it does not form over an earlier pit because it was covered by for instance clay earlier and the turf line on

Illus 25.2Â Interpreting turf lines: two possibilities In Illus 25.2 I have not added the effects of subsequent building operations, iron panning and other soil developments; but they were present at Calanais and made discrimination even more difficult. On the other hand (ignoring the many other ways in which the top of a cut can be very obvious), it can be easy to see the cut if the pit-digging leads to bits of ragged turfs flopping over the pit edge (Illus 25.3, 25.4). Turf curled over the Notes \ 1088

Calanais Survey and Excavation, 1979-88

—— Plan 12B/80 [DC 38032] undated May 1980 —— Plan 32/81 [DC38066] of 01/06/81 —— Plan 43/81 [DC38077] of 26/5/81

Illus 25.3 nterpreting turf lines: two possibilities edge of the Stone 42 pit and defined the turf line through which it had been dug; but alas, in this area, the two turf lines which existed near Stone 43 had merged by the time they reached Stone 42. The only useful thing the curled-over turf line edge showed us was that the pit had not been cut through the clay covering the turf line. Note 7.18.7 Variations between plans of BIII and the southern part of BI Minor variations in the orientation of depictions of baulks and temporary section lines made interpretation of some features difficult. In particular the depiction of small vertical stones 118 in slot 185 on Plan 12/80 did not line up neatly with the depiction of the palisade slot 876 further west on Plan 43/81. The relevant plans are: —— Plan 08/80[DC38027] used as a base plan; the undated last main plan of 1980

Illus 25.5 Plan 32, the southern part of the last full 1981 plan [DC38066 part] The absolute position of the plan 32/81 is inadequately recorded and its match with the other plans may be wrong by up to 20 cm. Also Plan 48/81 of 16 May 1981, showing the edges of cut-away turf lines as first discovered (in brown near Ring stone 43), was probably not in perfect register with the plan extract 31 of 28 May 1981, showing palisade elements.

Illus 25.4 Turf lipping over the side of Area H Notes \ 1089

Calanais Survey and Excavation, 1979-88

There are errors in DC38066, the last full plan of Area BIII, because it does not match DC38027, the last full plan of B in 1980, or DC38065 or DC38077 which match DC38027 quite well. The error in the central baulk south end position on DC38066 is of the order of 0.75 m (Illus 25.5); the error in the meeting point between the central baulk and the boundary with BII/III (south) is 0.2m, and when the east edge of the plans are lined up the central baulk orientation differs by 2 degrees. Also, the north end of the east extension of BIII is probably drawn 0.7 m too far north. It is not clear whether the error is in the original plan (which in many ways is of a higher quality than many others, in that it describes fills of features in a confidence-inspiring way) or in digitising, but the former, alas, seems more likely. However, once corrections to the south end of the medial baulk have been made, and ignoring the other two errors, the match between 38066 and other plans is not unusually poor. 25.8 Notes for Chapter 8: Area C There are no Notes for Chapter 8 25.9 Notes for Chapter 9: Area D Note 9.2.17 Layer 356 The layer under the chamber wall was wrongly labelled 360 on the digitised (and perhaps also the original?) versions of plans 44A and 44B, although correctly labelled on Section 43. Note 9.3.4 Plan 9a of 1980 and Plan 31 of 1981 Plan 31 is the first 1981 plan of DII. Because it bears the annotation ‘overlay of Plan 9 with additions’ it must be a revised version of the Plan 9a referred to in the 1980 Day Book, the last plan of DII that year. ‘Plan 9a’ does not seem to exist independently Note 9.3.12a Layer 320 Layer 320 appeared on several plans. It was found below plough soil 315 and above a black greasy

layer identified over most of the western part of area DI as the uppermost prevalent turf line 334. Over most of the southern half of DI it overlay the equivalent of turf line 334, turf line 365. More generally clay 320 was separated from the underlying turf line by patches of brown and black material. On Plan 34 of 9 May 1981 it was shown reaching as far as the south baulk section. Its stratigraphic equivalent was depicted but left unlabelled on the southern baulk section. It was labelled 950 on the section during post-excavation. In the easternmost part of the trench it was not present at all, judging by the main baulk section 62 where its stratigraphic position was occupied by reddish brown gritty charcoal-smeared 369 or the underlying 377 to the north and black-brown humic clay 389 to the south. The turf line shown under 369 in the north may therefore be stratigraphically equivalent to the turf line above 950 described below. A layer labelled 377 was at the same absolute level on the eastern part of the southern baulk near corner of the trench. During pollen analysis 950 was described as a humus-rich grit. It was about 5cm thick. It was at the same stratigraphic level as ginger-brown clay 320 on the area south of the cairn. A turf line (or at least a humus-rich band of material) lay between the plough soil and the underlying layer 950, although it was not noticed during excavation. Clay 320 was visible at the level of the surviving top of the original kerb and underlay the Victorian kerb line. For an understanding of its stratigraphic relationship to the prehistoric kerb, a useful comparison is between Photograph 1981.5.5 of 9 May, in which the Victorian kerb had been largely removed and the top of the prehistoric kerb was just starting to appear, and Photograph 1981.1.18 of 14 May in which the prehistoric kerb was better revealed. Despite interference from Victorian activities to find the line of the prehistoric kerb, represented on the photographs by darker material just outside the kerb line, and despite the complication of the remains of the lower part of the erosion hollow 376 described below, it looked as if clay 320 tailed off towards the prehistoric kerb.

Notes \ 1090

Calanais Survey and Excavation, 1979-88

Note 9.3.12b layer 332 The label “332” was added to the plan during post-excavation. On the original plan it was not numbered separately, nor was its extent precisely defined, but it was described as reddish brown and gritty, in contrast both to plough soil 315 and the more ginger brown clay 320. In the Day Book it was described as wetter and heavily smeared with charcoal. It coincided with an area subsequently labelled ‘315 base’ on Plan 34 of 9 May 1881. Although further cleaning led to the view that it was the same as the top of 320, and its superficial differences were ascribed to the protection provided by Lady Matheson’s Path, its original interpretation as basal 315 seems just as likely to be correct, in the sense that it was the top of 369 impinged upon by the ploughing which modified soil 315. Yet its identification with 320 suggests that the gingery colour of 320 and the reddish brown of 332 reflect soil processes including the washing down of iron oxide. It is also worth noting that clay 332 was close to the second Victorian intrusion in the cairn described in Part 9-2 and soil 369 just below 332 overlay the surviving remnants of the eastern part of the prehistoric kerb perhaps as a result of Victorian activity. It is possible that the Victorian activities modified clay 320 to form 332 in this area. Note 9.3.14: Section 50 recording problems There are problems with the original labelling of Section 50 (Illus 25.4). Elsewhere on Area DI soil

396 was described as a composite layer consisting of black greasy clay above brown greasy clay with silt. But 396 on Section 50 did not include black greasy clay. Worse, whereas the main layer of 396 lay under turf line 365, which is the exact equivalent of 334, the layer labelled 396 overlay turf line 334 on Section 50. The section also shows two lenses of brown clay wrongly labelled 388, immediately above the turf line and 396. The label ‘388’ was mainly used for a similar looking clay forming a bank below the turf lines 334 and 365 (Plan 60). For these reasons, the original labelling of layers 396 and 388 on Section 50 cannot be treated as following the context-labelling system normally used at Calanais. However, if it is accepted that it provides a shorthand description of the layers complementing that written on the section, it implies that the hollow was partly filled with material derived from layers under the turf lines. Note 9.4.2 The plan on Illus 9.91 The plan on Illus 9.91 is composite. The south-eastern kerb and cairn was drawn on 19 May and the northern part of the kerb, monolith and chamber wall on 24 May. Note 9.4.3a: Chamber plans on Illus 9.97 Plan 87 was numbered out of date sequence. On each plan in Illus 9.97 the SW chamber orthostat shape and position have been copied in red from Plan 64. The central monolith shape has been

Illus 25.6 Section 50 [NMRS DC38182] Notes \ 1091

Calanais Survey and Excavation, 1979-88

copied from Plan 87 and is shown in a graded grey on top of the shape shown on each plan. The grey lines meet at 6m grid north of Reference Point X. Errors were in general up to 50 mm; a few larger differences probably reflect variation in choices of what to draw. In the case of the SW orthostat the height at which it was drawn affected its planned position and cross-section. There were some errors in relative orientation, probably less than 2 degrees. Earlier plans also had offsets from one another. For instance Plans 58 and 69 fitted together with discrepancies of about 0.1m. Note 9.4.3b: Relationship of cairn to wall During excavation we were initially uncertain whether the two stones shown on Plans 71 and 72 of 26 May in black were part of the cairn body or the wall. If they were part of the wall at this level then the fact that they were intercalated with cairn stones would have demonstrated that the wall and cairn were built up at the same time as each other.

Between them and under them was orange-brown greasy clay. It was similar to the clay 356 found under the inner wall-face. The stones were below clay 318-4, which may have been undisturbed when the cairn became dilapidated around or before the ploughing with abundant Beaker sherds (Day Book 169). It seems clear that they were in fact part of the cairn rather than the wall and therefore the chamber wall was probably built to at least this height before the cairn was filled in. Note 9.4.5: North chamber baulk sections The north baulk of the DI chamber trench was drawn twice towards the end of the 1981 season of excavation, in Sections 97 and 104. Section 97 is ambiguous in that it does not show the point where the lower turf line (probably equivalent to 912 elsewhere) was cut by the chamber. Section 104 appears to be more a diagram of what appeared at the surface of the sloping back of the

Illus 25.7Â Ambiguities in the recording of layers 355 and 356 Notes \ 1092

Calanais Survey and Excavation, 1979-88

chamber than a true section. However it confirms that the patchy lower turf line was directly cut by the chamber, a point which Section 97 did not convey. A diagram in the Day Book confirms the existence of two turf lines cut by the chamber. Section 104 adds that at the back of the chamber the upper turf line was contemporary with the paving which covered the socket, a point which was not however supported by the diagram in the Day Book where another unlabelled layer intervened. Note 9.4.7: Ambiguities in recording of 355 and 356 The sequence by the chamber orthostat was originally described early in May as a 0.1 m thick layer of dark brown friable clay 356 between the bottom of the chamber wall and grey clay 355 forming a chamber fill. On Section 62 a layer labelled 356 overlay an unlabelled layer, above green clay labelled 372, a label generally used for undisturbed natural clay (Illus 25.7). The labelling contrasts with that from the short section running east-west (Illus 25.7) where the highest level was 0.1 m thick brown friable clay labelled 355. Under this on the short section was a 30 mm thick layer labelled 356, dirty green clay with iron panning overlying green clay labelled 372. The label 355 was originally used early in May for a layer of yellow-green-grey sandy clay with lenses of black and green-brown humic clay and organic material near the central monolith. I suspect that there was no connection between the layers called 355 and 356 at the west end of the chamber and those by the SW orthostat. It is difficult to escape the impression that the labelling of section A-B on Illus 25.7 contradicts that on section C-D. Note 9.5.1a Planning ambiguities It is not possible to say from the plans alone whether features under the cairn lined up neatly with one another vertically or horizontally at different levels. The weather towards the end of excavation was appalling. In those circumstances our

surveying method, using Fibron tapes to position planning grids at a considerable depth below our survey grid pegs, did not allow us to produce plans with good long-distance accuracy. Local accuracy remained good but, for instance, there were eastwest differences of up to 0.2 m over a two to three metre distance (6 to 10%) between Plans 94 and 101. In the middle of the area there were discrepancies of about 0.1 m (3 to 5%). The implications of this are not trivial. But I do not want to exaggerate the difficulties. Its main consequence was that one should not work the information from the plans too hard: no stratigraphic significance can be given to what seem from the plans to be minor overlaps. A more serious problem was that the multiplicity of similar and interleaving patchy layers led to uncertainties about which labels should be attached to which layers on drawings made at different times. In particular I think the main east baulk section was wrong where it assigned the label 912 to a turf line running over the top of slot 913. This turf line would better have been labelled 905 because the preferred interpretation is that the latter had developed in and filled gaps in the earlier turf line 912. This final account adopts nearly all of the onsite interpretations. This includes the conviction of the site supervisor that slots 913 and 915 were (as excavated in plan) at the same stratigraphic position as each other. My one disagreement is about the relationship of Slot 915 to the pit dug to erect the central monolith; the slot was very shallow and the supervisors view was that it was cut by the Monolith pit but my view was that it had originally risen slightly over clay spread from the mound over the monolith pit and had been truncated. That I discuss in some detail in Chapter 9: Area D and in Technical note 9.5.4 below. Most of the interpretations proffered in the main text can be tested if future excavation includes the area under Lady Mathesonâ&#x20AC;&#x2122;s Path between Areas D and B. Note 9.5.1b: Green clay ambiguities The green clay of the mound surrounding the monolith was not always distinguishable from Notes \ 1093

Calanais Survey and Excavation, 1979-88

green clay under the cairn. In the main narrative I have not described the many ways in which I have explored alternatives to the preferred interpretation described there. Illus 9.112 to 9.114 are based on Plan 101 of 30 May 1981 with the addition of some information from Plan 88 of 27 May 1981. Plan 88 showed the extent of grey-green silty clay 906 and turf line 905 in isolation from other features. Clay 905 occupied much the same area as the underlying slot 913. It was described as black greasy clay (a description often applied to a turf line) and 906 was described on plan as dirty greasy clay and in the Day Book (Page 203) as green silty clay. On other plans 906 was described as grey. Clearly it was heterogeneous; the differences cannot be ascribed simply to changes in weather conditions. My interpretation is that it reflects disturbance and weathering of green clay. Below the brown clays at the base of the cairn on Area D patchy green clay 398 was recorded as overlying turf line 905. It does not seem to have been planned although a 1982 annotation on Plan 88 (Illus 9.112) also records that turf line 905 appeared after 398 was cleared away. The green clay labelled 373 may simply have been spread from the mound round the base of the central monolith (Illus 9.115 Day Book 270 sketch plan). In the west it merged with the clay of the mound and in the northeast it was not recorded as present. Its eastern part was probably equivalent to the green clays of the platform under the base of the cairn on Areas BIWX, BINX and H. That means that the higher green clay 398 must have been spread above 905 by trampling. Certainly it did not form a coherent widespread layer. Slot 915 may have been visible in grey-green to dirty clay 906; the label was marked on Plan 101 in roughly the right position (Illus 9.115) but its outline was not drawn. The preferred explanation is that the slot top was obscured by clay trampled into it when people were working in this area at the start of cairn building. Note 9.5.4: Slot 915 At its western end slot 915 appeared to run up to paving stones associated with the central mono-

lith mound on Plan 94 (Illus 9.117). But the day book and context sheet record an interpretation that slot 915 was cut by the monolith ramp or socket. My observations and discussions on site convinced me that it petered out at this level and appeared to have stopped because it had originally risen over the paving round the base of the monolith, in the same way that the basal layers of the cairn wall rose up over the mound. It was thus actually later than the cut for the monolith socket. The fine brown clay 939 in the mound shown in yellow on Illus 9.117 and 9.118 may reflect its former presence. Note 9.5.5: Slot 920 and related soils Sections 95a and b, about 0.2 m apart, had apparently different evidence for the relationship of Slot 920 to earlier layers (Illus 9.111). On Section 95a near the middle of the slot its flat bottom sloped and its southern side cut through grey mineral soil (presumably 911) and an underlying mineral soil (presumably 914). A bifurcating turf line about 10cm to its south lay approximately at the junction between the mineral soils. The north side of the slot, on Section 95a, cut a patch or layer of clay, then two turf lines (presumably 905 and 912) separated from each other by about 20 mm, and then the underlying clay soil 914. On Section 95b the south side of the slot was partly overlain by a 0.2 m long patch of what was thought during recording to be turf line 912 and the overlying subsoil 911 of turf line 905 on which the cairn was built. The north side of the slot cut only the lower mineral soil 914, here over 0.1 m deep. There are three possible explanations for the patch of thin black material partly overlying the south side of the slot in Section 95b. It may have been an upturned turf; or it may possibly have been a remnant of charcoal-rich clay 390 or of a turf line in the cast of a basal cairn stone. That however leaves the problem of explaining why the northern side of the slot on Section 95a cut a layer of clay above two turf lines and their subsoils with a total thickness of 60 mm while on Section 95b the basal soil 914 survived to a height Notes \ 1094

Calanais Survey and Excavation, 1979-88

of at least 100 mm. The simplest explanation is that the soil 914 covered by the lower turf line rose quite sharply by 80 mm over a distance of 200 mm. Similar abrupt changes in level were recorded on Section 62. Local compression by basal cairn stones may have caused this. On the other hand soils and turfs of the underlying cultivation beds may have complicated the stratigraphy in this area. The slot was probably on the south edge of Rig 4, or almost in its gully. The turf lines recorded on sections 95a and 95b may have had nothing to do with those recorded on the main east baulk section 62.

81 because the section stopped just short of it. In Section 103C (undated) and in other versions of much the same section (Section 100 and 102C of 9:6:81) differences probably reflect cleaning back. A feature noted on Plan 40 of 9 May 1981 in the extreme NW corner of DV, and interpreted at the time as the edge of a turf line, may have been part of it. In Section 103B DC38237 of 4.6.81 the lower of two turf lines just overlaps the edge of the pit (Illus 9.126). Note 9.8.2a: More green clay ambiguities

Note 9.5.7: Slot 913 The slots 913 and 915 on Area D seemed to be of a different date from one another because on the main east baulk section slot 915 cut a turf line labelled 912 which was shown overlying slot 913. Yet on plan 88 (Illus 9.112) that turf line was shown in the area of Slot 913 and labelled 905. It seems almost certain that turf lines 912 and 905 had merged at this point, the earlier turf line having healed over. The upper part of the slot, although clear in section, was not well-defined in plan. When first noted it appeared to be fairly straight and of even width. At a slightly lower level, planned four days later, it appeared to be crooked and to kink sharply northward towards its east end. As described in the main text the completely different line of the lower expression of the slot in its eastern half may be due to confusion with a cultivation gully. At its western end the lower version of the slot extended beyond the upper one. Although there were planning discrepancies at this level they do not seem to have been sufficient to explain the difference, the cause of which remains unresolved although trampling during cairn construction may again have been to blame, spreading out the upper layers of the slot. Note 9.7.2: Pit 917 Pit (917) was revealed in the west baulk of DV and DI when it was cleaned back towards the end of the 1981 season. It had not been visible in Section

Illus 25.8 Section 92 of 2 June 1981 [NMRS DC38224] The original label 373 used for the green clay above and below the paving at the base of the monolith has to be treated more as a description meaning ‘green clay close to the monolith’ than a context number. The pit fill was numbered 922 but this label does not seem to have been used much. On plans 61 and 65 the green clay round the monolith was labelled 373. It interleaved with clay 391, interpreted as spoil from digging it (Illus 9.137). But the pit fill itself (922) underlay 391, cut from the top or near the top of mottled turfrich clay 394, and probably the green clay on Plans 65 to 74 should have been labelled 922 (see also Technical Note 9.8.2c). Notes \ 1095

Calanais Survey and Excavation, 1979-88

Note 9.8.2b: Problems with labelling of Section 92 Section 92 of 2 June 1981 was labelled with temporary numbers 01 to 08. Table 25.2 The concordance list on Section 92 01 dark greasy clay

373

03 Mottled brown clay

911

02 green clay Mon 29 socket 04

Dense black clay Old Ground Surface

373 918

05 Brown/green gritty clay

905

07 Brown/black mottled humic clay

390

06 Green/black dirty clay 08 Old ground surface

906 912

A concordance with context numbers was written on the section drawing (Table 25.2). These appear to be unjustified. Table 25.3 Descriptions belonging with the contexts listed on Section 92 when found elsewhere

Section 92 a layer 394 described as ‘a dense black greasy clay old ground surface with a grey-black gritty clay base’ where it occurred in DIV (Context Sheet 394). Context number 918 seems to have been used here only on Section 92 and 97 and must be the same as 394. The layer under mottled soil 394 in DIV rested on a patchy turf line which overlay a patchy grey mottled soil (904). Under the patchy soil 904 was mottled orange clay (903) which may have been equivalent to the low pile of orange clay stained soil 339c in DII around a metre away to the west, which underlay cultivation bed 4 (Table 25.4). Table 25.4 A modified version of the concordance list on Section 92

01 dark greasy clay 04

Dense black clay turf line

06 Green/black dirty clay 07

Brown/black mottled humic clay

Maybe related to 391, an orange yellow speckled clay in DIV? 394 (=918) Lower part of 394? 356 or 911

373 Green to green-yellow clay

03 Mottled brown clay

911

390

922

Charcoal impregnated clay, onto which the second layer of cairn stones was laid

green clay Mon 29 socket

Lower part of 394?

905 Dense greasy black turf line

05 Brown/green gritty clay

906 Green/grey silty clay.

08 Turf line

918 Upper layer of turf line

This tentative reinterpretation brings Section 92 into fair accord with the sequence in DIV.

912 Dense black greasy turf line 911 Mottled grey/brown gritty clay The lack of correspondence between the descriptions written on Section 92 and the descriptions given to contexts elsewhere (Table 25.3) strongly suggests that the context numbering in the list on Section 92 was wrong. Plan 87, drawn on 1 June 1981, the day before the section was drawn, shows at the west end of

Patchy turf line under 394?

Note 9.8.2c: layer 391 When 391 was discovered on 20 May 1981 it was described as spreading under the paving round the base of the monolith (Day Book 245). Plan 56 of 16 May 1981 (Illus 9.138) shows a paving stone Notes \ 1096

Calanais Survey and Excavation, 1979-88

overlay 391 and 373 where they met. Judging by Section 102b of 9th June clay 391 was interleaved with green clays called 373. But at least the bottom part of the lower of these should probably have been labelled 922 (green clay in monolith socket. Note 9.8.6 layer 394 Plan 98 (Illus 9.138) was numbered out of sequence. It preceded Plan 74. Layer 394 rested on a patchy grey mottled soil (904) which was probably the soil of a patchy turf line recorded at the base of 394. These were cut by the monolith pit. Its edge became unambiguous at the level of 904. Under the patchy soil was mottled orange clay (903). 394 may in part have been composed of turfs thrown on one side when the monolith pit was dug. The orange clay patch 903 may have been equivalent to the low pile of orange clay stained soil 339c in DII around a metre away to the west, which also lay directly on the natural and was partly covered by a turf line (316c). In DII the turf line 316c was interpreted as underlying a cultivation bed and surviving in its furrows. Extrapolating the rig from DII to DIV, it would lie in the eastern part of the area DIV, where clay 394 suggested turfs laid upside down on the turf line. The two explanations are not mutually exclusive: layer 394 may have been hybrid, partly a flattened old rig and partly the first turfs from digging of the monolith pit.

Note 9.9.14 Area DIV lessons for Area DI If Area DIV had been ploughed to the level of the bottom of Kubiena Box 1 then 904 itself would have been a strip of patchy silty grey clay, the earlier orange clay 903 would have been exposed to its south and the later grey silty soil 394 with halos and turf fragments would have appeared to its north. It is very doubtful that it would have been clear which levels were stratigraphically the latest, so it would not have been clear how to proceed. The complexities and ambiguities revealed by sections 95a and 95b in Area DI on either side of the temporary baulk preserved in the middle of DI (Illus 9.111) reinforce this point Note 9.9.16 The idea that there were cultivation beds on Area D arose with the excavation of Bed 4 on Area DII to the west of the central monolith. It was made immediate with the discovery of a ridge of grey gritty material 388 lying in plan between turf lines 334 and 365 on DI (Illus 9.164). It was given impetus by the sinuous turf lines outcropping in 334 because we speculated that they might reflect turf lines undulating between two beds. Illus 25.9Â Copy of Illus 9.176, the initial bed hypothesis

Note 9.9.1 Cultivation beds were clearest on Area C at the east end of the east row of standing stones. There they were orientated roughly east-west, had a wave-length of about 1.4 to 1.6 m and survived to a height of about 0.1 m. Excavation was not taken below their surface except in narrow exploratory slots. Dating was provided by ascription of the material overlying the cultivation beds to early zone CaN-3aii in the CN-3 column at Calanais Leobag sometime between 2750 and 2300 cal BC. That suggested that the underlying soil belonged before then. However despite their similar orientation to the beds on Area D there is no strong reason why they must have been of exactly the same date.

This became established as a loose working hypothesis. Illus 9.165 which illustrates this hypothesis is spuriously precise. In fact we were far from confident that beds, if they had ever been present, would be straight, or regularly spaced. The wavelength of about 1.6 to 1.7m, in this initial hyNotes \ 1097

Calanais Survey and Excavation, 1979-88

pothesis, was determined by the width and centre line of the bed in DII and the centre line of the ridge of grey material 388 (Illus 9.164). The idea was that seed beds about 1.4m across might be separated by gullies about 0.2 to 0.3m wide although as already explained these might be gentle depressions rather than sharp narrow features. The characteristics of Bed 4, which informed identification of beds elsewhere, were as follows. —— Stray patches of in situ turf lines, and low heaps of dirty orange clay, survived in some places under the bed; —— upside-down turfs, some with the turf line and the underlying green (or sometimes orange) clay subsoil retaining its colour and texture in each turf; —— areas between seed beds could simply be shallow broad declivities, but sometimes narrow gullies occurred, perhaps created through clearing out of soil; —— turf lines survived at one or more levels in the gullies and on the flanks of the bed; stretches of turf line were covered by soil as the gullies filled in; —— when beds were levelled the turf lines on the bed flanks were truncated. Soil filled the gullies; in many places on Site D the mineral soils above and below a truncated turf line were identical; —— although the gritty and quartz-rich components of the beds may have resulted in part from decay of stones (a process seen particularly clearly on the periphery of the later chambered cairn), they may also have reflected importation of sand on seaweed, and turf-rich material from grittier soils. —— detritus perhaps from weeding and removal of small stones from the seedbeds occurred in or at the edges of the gullies. The problem from the point of interpretation was, and remains, that many of these small-scale phenomena could have been caused by activities other than farming. Perhaps because the bottom 0.1m of the soils and clays were left unexcavated on Area DI we did not see beds and gullies in plan except where Bed 2 survived between two complex turf lines. There were stray examples in Area DI of the formation processes observed in the bed on DII,

but it cannot be pretended that every detail of every sequence of turf dumping, litter-layer formation and burial was fully recorded or completely understood. The bottom 0.1 m depth of soils in DI was not excavated except in a few narrow exploratory slots, and most of the evidence related to bed formation would be expected to survive in those soils. Illus 25.4 illustrates the problems; Sections 95a and 95b are almost too rich in variations to be interpretable. The beds, if they had ever been there, would have been damaged by subsequent flat cultivation or levelling, and the erection of standing stones and building of the cairn would have further disturbed them. Nevertheless, the information recorded on plans and slit sections allows interpretation of some of the features on Area DI as the remains of four beds, bed 1 underlying Ring Stone 47, about 1.5 m SSE of bed 2, bed 3 about 1.5 m to the NNW of bed 2 and bed 4 cut by the pit for the central Monolith 29 (Illus 165). The crook in the line of the lower part of Slot 913 does fit the idea of a gully between Beds 2 and 3. The feature labelled 943 on Sections 95a and 95b might be the westward continuation of the cultivation trough. However I suspect that 943 was in fact a non-circular pit rather than gulley created by cleaning out the trough. Perhaps the grey soils with fragments of turf line, particularly those on Section 95b (the upper section) are the fill of the trough between the two beds. But generally elsewhere such soils with turfs in them were part of the body of cultivation beds. Also I do not entirely trust the depiction of variations in grey soils in section 95a at this point because of the way that the lower turf line is shown as cut at a point well to the south of the end of the upper turf line. The grey soils rising at the south ends of Sections 95a and 95b may be the same. On Section 105 (Illus 9.161) the lowest soil had been labelled 388. Technically this was wrong because that label had been used for soil in Bed 2 to the north; it has been relabelled 946 in the Context List. But it is still likely that it was the soil of a cultivation bed. The most definite bed on sub-area DI, bed 2, was on the area least disturbed by later features, in the middle south part of the area. It was seen as a ridge of gritty clay running WSW to ENE Notes \ 1098

Calanais Survey and Excavation, 1979-88

Illus 25.10Â Sections 95a and 95b and the main baulk section 62 after removal of the overlying late ploughing level and patches of clay though to represent funerary deposits. The turf line of one or more turf lines lay to either side of it. It appeared to curve slightly (Illus 9.164). Note 9.9.23: Dating the cultivation beds

Illus 25.11Â The gully between Beds 2 and 3 on Sections 95a and 95b (extract from Illus 9.177)

Without the evidence from the pollen samples the cultivation bed might have been attributed to CaN-2c, the cultivation phase at Calanais Leobag preceding CaN-2d and more precisely it might have been supposed that creation of the cultivation beds started at the same date as formation

Notes \ 1099

Calanais Survey and Excavation, 1979-88

of the 35cm level in Leobag column CN3, where Cerealia pollen first appeared. This dated approximately to between 3900 and 3600 BC. It might also have been supposed that use of the cultivation bed ended at the CaN-2c to 2d transition at between 3490 and 3020 cal BC. The pollen records of columns D2-3 in DIV and D4-5 in DI both started off with relatively high birch and hazel percentages after which a clearance phase followed. This marked the end of zone CaN-2d, dated to between 2980 and 2500 cal BC. Although it was assigned overall to zone CaN-2d, both agriculture and cattle breeding were practised during CaN-3ai in column D6-7, coinciding with a thick turf line. This must have formed the pre-cultivation bed ground surface and where it survived it received cereal pollen from neighbouring cultivation. In columns D2-3 in DIV and D4-5 in DI the first Cereal-type pollen appeared at the beginning of Phase D-2 (zone CaN-3ai) together with Compositae pollen of the Liguliflorae type and Plantago lanceolata. Subsequently grasses increased firmly and pollen of the Trifolium-type and Trifolium repens appeared in the samples, after which the indicators for agriculture declined, marking the end of CaN-3ai at some time between 2750 and 2300 cal BC. Agriculture was temporarily abundant during phase 2 in column D6-7 while cattle-breeding was still practised. Heather increased strongly in phase D-3 (CaN3aii) in columns D2-3 in DIV and D4-5 in DI Possibly cattle breeding was practised because no decline was registered in the Plantago lanceolata, the Ranunculaceae and the Trifolium curves. Grasses showed a temporary decline during this zone. In column D1 the lowest levels belonged near the start of zone CaN-3aii, a stage during which both agriculture and cattle-breeding were practised. It was followed by a stage during which agricultural activities declined and a heather spread was registered in the samples. These three stages were also clearly registered in the CN-3 column zone CaN-3a. The increase in heather marked the transition to zone CaN-3b. At the start of phase-3 in column D6-7 there was a sharp decline in the birch values. Cereal-type

pollen reappeared in the samples and grasses increased sharply. Heather pollen declined, possibly as a consequence of increased agricultural activity around the site. Grasses recovered again in Phase D-4 (CaN-3b), in columns D2-3 in DIV and D4-5 in DI, together with cereal. Heather declined, possibly as a consequence of increased agricultural activities. Together with the birch, hazel, alder and pine declined. Pine fell below the 1% level which coincided with the regional pine decline. This decline in the pine pollen marked the beginning of zone CaN-3c, dated to between 1800 and 1200 cal BC in Leobag column CN-3. During phase-4 in column D6-7 (in CaN-3b) the cereal curve increased but the Compositae lig did not reach the high values of Phase 1 in column D6-7. Heather values recovered while grasses declined to around the 40% level. In all of the columns there was a correlation between the occurrence of turf lines and an increase in heather pollen. The fact that the heather curve showed two successive maxima in D4-5 but the curve in column D2-3 had only one peak was probably due to local influences. In Column D6-7 the turf lines seemed to correlate also with temporary declines in the cereal pollen curve (except in the lowest turf line which immediately preceded CaN-3ai and contained the first cereal pollen from Area D. This demonstrated temporary declines in agricultural activity around the site during periods of turf formation. Note 9.10.3a: Sections 68 and 93 of Ditch 921 On Illus 9.175 the drawings of Sections A-B and C-D have had their horizontal scale adjusted by a factor of 0.85 to bring their width to that of orthogonal cuts across the ditch at the same points. A health warning has to be sounded on Sections 68 and 93. The primary records relating to them are very confusing. ‘North-facing’ and ‘south-facing’ appear to have described the direction the person drawing the sections was facing. Indeed I have a recollection of what I thought was an abstract discussion of what such terms meant when those Notes \ 1100

Calanais Survey and Excavation, 1979-88

sections were recorded. The triangulation co-ordinates for the section points make little sense. Both sections appear to have been ‘flipped’ east to west. The pit on Section 93 was given a context number, 385 the same as that of a pit some metres away, and not related to Ditch 921. It should have been numbered 395 Nevertheless I am confident that the depictions in Illus 9.175 are correct. It seems pretty clear that the fills on Section 68 are different from those on Section 93. I think that that may be because Section 68 was in a different feature, although the section documents more than simply the trough hypothesised above.

by Plan 11 is a fairly good qualitative match with reality; but the detail in its western part must be treated with caution. Note 11.9.7 Stake hole 644 There is a conflict between the recorded position of stake hole 644 on the digitised versions of Plan 10 (Illus 11.17) and Plan 6 (Illus 11.13). A sketch plan of 23 May 1980 (Illus 11.19) resolves the issue; the stake hole was in the position shown on Plan 10. 25.12 Notes for Chapter 12: Area H

Note 9.10.3b Pit 385 and 395 On the digitised version of Section 93 the pit was labelled 385. This number however referred to a posthole shown on Plan 55 (DC38187A) near the cairn. Probably the label should have been 395 (see section E-F on Plan 55). The confusion in labelling probably arose because a line of small features between 385 on Plan 55 and 385 on Section 93 was interpreted at one stage as possibly forming a line running NNW-SSE. Subsequently this interpretation was abandoned. 25.10 Notes for Chapter 10: Area E There are no Notes for Chapter 10 25.11 Notes for Chapter 11: Area F Note 11.1.3: Contour plan ambiguities The spot heights on another plan, Plan 10, suggest that the centre of FV, the east-most trench, was 1.79m below datum and the centre of FVIII, the north-western trench, was 2.38m below datum. That is a height difference of about 0.59m between a point in the avenue and a point on the slope. This correlates well with the height difference of about 1.0 to 0.5m suggested by Plan 11. A note on page 6 of the Day Book says the lowest marshy point was 1.73m below site datum and two low wet spots were at 2.43 and 2.53m - height differences of 0.7 and 0.8m. The photograph on Illus 11.1 suggests that the general impression given

Note 12.2.3: Labelling error On Plan 3 (Illus 12.13) the largest area of chamber fill, marked ‘720?’, was wrongly labelled 721 on the digitised plan. 721 was undisturbed subsoil in the chamber. Section 14 (Illus 12.15) and a sketch plan in the Day Book p.4 (Illus 12.9) suggest that it should have been labelled 720. Note 12.2.7: Cereal context label error I distinctly remember that I found grain in the capping 708 of the cairn or the top of dark layer immediately below. That memory was of an event which was exciting at the time and thus registered firmly. Also it has been refreshed fairly frequently over the years during the various attempts to write up Calanais. The cereal grains, 3 barley and one wheat (Triticum) grain, were recorded in the macroplant report as from samples CSS 752-773) and selected for radiocarbon dating. The context number for the cereal grains in the macroplant report was given as 764, but that must have been misread and must truly have been a carelessly written 704 (general label for the cairn outside the chamber wall) because 764 was a fill of primary Ring stone pit 775. It was called small find 75/81 in the original Finds book where it was assigned to layer 728 and dated 6 May 1981. But Layer 728 had not been reached at that time. The top layer of the cairn was removed on 5th May (Plan 2) and the second layer on 5th/6th May. The latter plan shows soil layer Notes \ 1101

Calanais Survey and Excavation, 1979-88

712 (immediately under 708) amongst the cairn stones. This proves the point. In earlier versions of the Radiocarbon date list the context of the grain had not been resolved satisfactorily. That has been remedied in the list of dates in Chapter 23. Note 12.3.1 Pit 710 and Slot 735 The lower fill of the pit 710 was reinterpreted as the fill of the underlying slot 735 (Day Book Page 3) and Plan 5 has been renumbered accordingly. Note 12.5.12: Interpreting layer 771 Four possible explanations for the origin of Layer 771 under the chamber wall were considered. 1. It was an incipient litter layer which formed on the green clay under the cairn. This was the preferred interpretation on-site and also during post-excavation analysis, because it was visually similar to some of the weaker old ground surfaces elsewhere on Area H. 2. It consisted of fine humic material which had seeped down from grey clay layer 770, or if the latter was itself the result of down-washing of a fine soil fraction, had seeped down earlier than formation of that layer. Indeed the pollen from 771 was quite like that from clay 770. The main support for this idea is the fact that there was fine silt on some of the basal layer of stones under the chamber. 3. The layer was the remains of vegetation ritually scattered on the green clay. This explanation was conceived during post-excavation many years ago. However, the relatively high proportion of tree pollen in the pollen assemblage does not suggest â&#x20AC;&#x2DC;floral offeringsâ&#x20AC;&#x2122;, and even though the pollen might represent a mixture of the general pollen in the air at the time plus pollen from offerings that possibility was therefore not preferred to the others (See Chapter 21: Pollen]. 4. The layer consisted of airborne dust which had blown over the exposed clay; but if so it must have settled after a few of the chamber wall stones had been lain down. This seems to be slightly special pleading.

Kubiena box samples were taken for pollen analysis. Box 1 incorporated layer 771 under the chamber wall and the green clay of the platform. The descriptions were as follows: 0-2cm green humic gritty clay and 2-6cm green slightly humic clay. This largely resolves the issue. Layer 771 under the chamber was a humic version of the green clay, about 2cm thick. It was thus almost certainly an incipient litter layer forming on an exposed old ground surface. The grit disseminated in it (not noted in the slightly humic green clay below) does not reflect a component of downwash but activities which took place as the litter layer formed. That is the preferred explanation. However, layer 771 should survive on both sides of Area H so if there is excavation here again at some time in the future soil micromorphology and other techniques can be used to test the possibilities considered here, alongside other explanations. It would also be interesting to look for drops of tephra (volcanic glass) in this and other layers under the cairn given the possibility that a period of poor tree-growth in Ireland around 2354-2345 BC might have been due to the Hekla 4 volcanic eruption in Iceland Baillie 1999, 205). Note 12.6.2: Ambiguity in record for relationship between layers 749 and 758 There is some confusion in the records for this area, with what seem to be alternative drawings of the stratigraphy at the east end of the section under the massive kerb slab (Illus 12.89 64-5). Clay 749, clean olive green subsoil is recorded on its context sheet as lying over turf line 758. I can find no evidence for this on plans or sections; on Section 85-6 (Illus 12.80) its relationship to the turf line is ambiguous. Also on the section under the kerb the layer below the turf line under the platform was labelled 750 (Illus 12.80 64-65). That label there introduces a stratigraphic anomaly. In fact this layer was probably 777, the light grey layer under the turf line on the main sections (Illus 12.34, 12.36). But green clay 760 was later than some parts of mixed clay 716 and the large stones near the base of Ring Stone 42, which formed a group overlying the standing stone pit fill (Section 85-86). Notes \ 1102

Calanais Survey and Excavation, 1979-88

Note 12.7.3: Depression 746 ambiguities There is some doubt about the identification of layers in the depression 746. The layers most likely to cause confusions have been renumbered in this final narrative. What was labelled 750 has become 1005 and what was labelled 751 has become 1009. There were also some discrepancies in the records relating to these layers. The original label 740 for the patch of grey green material in the upper left-hand corner of plan 6 was wrong (according to the context list, the number was for modern upcast in chamber) / olive green clay). A label 745 on Plan 7 was an original or a digitising error for 743 (according to the context list, the number 745 related to an earth clod in modern upcast 740 in chamber). The identification of the main fill (now labelled 1005) as 750 may have been stratigraphically correct if the identification of the turf line 1009 as 751 was correct. But it is also possible that the turf line was earlier, perhaps equivalent to 776 which formed the lowest turf line on Section 13 (Illus 12.101), and if so the underlying very variable soil 743 may have been the equivalent of 789 and 792, the early plough soil. Much of the depression should survive immediately north and west of the edge of Area H and it may prove possible in any future excavation to relate it to the strata around Ring stone 41. Note 12.8.4: correlating H and B plans Registration between the plans of H and B was poor and numerous slight adjustments had to be made to bring them together. Some details have been added from a photograph (Illus 12.103, 12.104). Errors in the composite plan near Stone 42 are probably of the order of 0.1 to 0.2m. Note 12.8.5 Wrong assignation of Sample 230 charcoal to green clay 767 Work on the excavation narrative for Area H to incorporate information from the Sheridan pottery catalogue led to the realisation on 21 No-

vember 2011 that the radiocarbon-dated charcoal sample ascribed to Context 767 in the McCullagh charcoal report ((AA-24969 4095+/-45) was almost certainly from context 768. Clay 767 was clean green clay in the pit for Ring stone 42. The radiocarbon-dated charcoal sample ascribed to 767 was therefore taken to provide a terminus post quem for erection of Ring stone 42 (Ashmore 1999a, 2000). The original 1985 Rod McCullagh charcoal report (like the pottery bag for Find 230 - see below) did not include sub-Area numbers for Sample 230. If it had it might have distinguished between HI (where 767 was) and HII (where 768 was). The original Finds Book has Find 229 as from 768 and containing pot and carbonised hazelnut, immediately followed by Find 230 (with the contents shown by ditto marks). However, on 20/10/97 Andy Heald annotated the Finds Book entry as follows “wrong context. Should be 767 check on sample bag” Checking other records I find: —— The 1995 version of the Context list has Sample 230 as from both 767 and 768. —— An early typed version of the Finds list (alas undated) has no pot from 767, and both charcoal samples 229 and 230 from 768. —— Mel Johnson’s draft pottery catalogue (the earliest file version is dated 1995) has no pot from 767. Alison Sheridan has checked (22-23/11/2011) the finds bags and reports that the old bags for find 230 have context 768 on them. The pot in Find 230 was as follows. Cat

Comment

7 spalls and 7 fragmentary spalls 1007-1020 from a fairly thick-walled probably E/MN corky pot. Fairly similar pottery was in Find 81.229 from 768, which was a secondary chamber wall capping.

Notes \ 1103

Calanais Survey and Excavation, 1979-88

Cat 18_26 27 28 29

Comment

charcoal sample 230 the charcoal sample bag was mislabelled. This will have happened before its receipt by Rod McCullagh.

A featureless from a fine Beaker or less likely a particularly fine E/MN non-corky pot

Note 12.9.2: Section 34 ambiguities

9 sherds from a large fine E/MN corky pot (ASH 8)

Spall possibly from a very fine Beaker A sherd probably from a fine Beaker.

There is no hint in the context records or the Day Book that clay 767 contained any charcoal or pot, although that is purely negative evidence because finds were often not listed on context sheets even when present. It is from this evidence extremely probable that when Find 230 was divided up into Pot 230 and

There were some apparent discrepancies between Section 34 and the plan on Illus 12.121. The plan is composite. Overlay 18 showing slot 795 / 1001 has been shown on Plan 9, which included the complex slot under the chamber wall and a patch of 752/760 in bright green. Probably the discrepancies are due to the fact that the lower part of Section 34 was not drawn on a vertical plane. However there is a discrepancy between the projected line of the western part 785B under the cairn stones on plan and the record in the section. The composite plan seems to fit well enough with Section 33.

25.13 Notes for Chapter 13: Test Pits G and J Note 13.2.1 Test Pit G1 Table 25.5 Details of the soils in Test Pit G1 Context Layer Depth 1301

1302

Op¹.

Colour Description

0-2cm

Litter layer with decomposed root mat, turf-like. Rare presence of worms noted

5 YR 2-21cm 2.5/1 (19-24cm) Black

Peat with clay. Rare rounded stones weathered 2cm- 6cm. Approx. 5% unstained quartz grains. Crumbly when dry. Uncompacted, springy. Many roots both fleshy and fibrous. Indistinct, clear (boundary definable to 6cm), irregular lower boundary to:-

1303

21-38cm

5 YR 3/2

Brown fibrous peat with very little, if any, included mineral matter. The peat is redder at the top, becoming browner downwards. Many fine fibrous roots. Wetter than Op¹, softer but not crumbly. Sharp flat boundary to:-

1304

Omh 38-41cm

5 YR 2.5/1

Black amorphous peat. No stones, sticky, greasy; fine fibrous roots common. Abrupt irregular lower horizon to:-

Notes \ 1104

Calanais Survey and Excavation, 1979-88

Context Layer Depth

Colour Description

1305

5 YR 3/2

1306

1307

A2/ B1

41cm-?

Brown coarse sandy clay. Many fine and very fine fleshy and fibrous roots. 10% coarse sand grains, unstained. 20% indurated green gneisses 2cm-6cm. Some boulders from the layer below cut the A1, making it discontinuous. Very irregular boundary to:-

Bouldery, sandy clay (NB not boulder clay). 90% boulders with clay and sandy clay between. Boulders of gneiss, rounded, angular, blocky - tabular. Root mats on stones, charcoal between stones. Occasionally the matrix material has a slightly red-brown tinge. Layer A2/B1 contained abundant pottery which is described in Chapter 18: Pottery. Abrupt irregular lower boundary to:-

90cm

Light green olive sandy clay (colour not represented on Munsell chart pages available). Soft, very wet, no roots and some stones. Base not seen [pit became waterlogged] but becoming rockier, suggesting this as B2/C with clay forming down into jointing, clay inwards being hydrated.

B2C

Note 13.3.1 Test Pit G2 Table 25.5Â Details of the soils in Test Pit G2 Context Layer Depth 1321

0 - 0.02m

1322

0.02 - 0.09m

1323

1324

1325

0.09 - 0.28m

Colour

Description

7.5 YR

Layer 2. Black peat, very rooty stone-free with some quartz grains and charcoal. Springy, very rooty, and fleshy and woody roots to 3 mm. No mottle, no charcoal and no structure. 5cm irregular boundary to:-

10YR 2/1

Litter layer under grass.

Layer 3. Black peat with less quartz grains than above; 1-3mm fleshy roots. 5% very disintegrated crumbly stones, 2 - 6 cm, not possible to record angularity. Worms. <1cm flat to irregular boundary to:-

0.28 - 0.39m

10YR 3/2

Very dark greyish brown very humic fine sandy with no mottle and with 30% 2 - 6cm stones and 6 to 20cm sized ones on top surface. Stones are sub-angular, weathered, and rotten when green. Very frequent fissures and rootmats; fine irregular vertical cracks. No structure, no mottle, charcoal in top surface as normal. Many very fine fleshy (woody) roots. Abrupt irregular lower boundary to:-

0.39 - 0.52m

10YR 3/2 to 10YR 4/2

Very dark greyish to greyish brown (lighter than above) slightly sandy clay. Very wet and sticky with some 1-2mm fleshy roots. 20% 2-6cm stones. Sharp irregular boundary to:Notes \ 1105

Calanais Survey and Excavation, 1979-88

Context Layer Depth 1326

0.52 - 0.55m

1327

0.55 - 0.59m

1328

0.59 - 0.69m

1329

0.69m downwards

Colour 10YR 3/3

Description

Rooty mixed sand with a 10% humic clay fraction Mainly uncemented but slight cementing in places. Lenses of green. Discontinuous 0-5mm root mat.

Light green

Colour not on Munsell chart. Fine sandy clay with some clay bands and very fine fibrous roots; a slightly platy structure with very variable cementation. Less than 10% stones. Abrupt lower boundary to:-

Green sandy clay with staining. 20% 2-6cm 10YR 5/6 yellowish brown rotted stones. No roots. Sticky irregular peds. Base flooded and not seen

Note 13.4.1 Test Pit J Test Pit J was occasionally referred to as G3. Its NE side measured 2.17m, its SW side 2.07m, its NE side 2.15m and its SE side 2.11m. It was dug to a depth of 0.76m. Only the main profile is reported here; the labels of the other consisted entirely of cross-references to this main profile. Table 25.6Â The main profile in Test Pit J Context Layer

Depth

1332

2 - 8cm

1331

1333 1334

0 -2 cm

8 - 11cm

11 - 14cm

Colour

Description

Litter layer; medium brown wet under grass

Very wet medium brown peat, fibrous with fleshy roots Slightly dark fibrous peat Dark brown fibrous peat with grits and dried peat lumps. Regeneration layer.

1335

14 - 27cm

1336

27 - 48cm

1337

48 - 49cm

Orange band 0-2cm above it. Dark grey brown amorphous peat. Sharp boundary to:-

49 - 55cm

Layer 1. Dark brown very rich humus rich clay with red brown fine roots; a few fleshy roots and abundant very fine fibrous roots and no mottle. Cemented, imbricated non-greasy, <1 cm rotted stones along with quartz grits and some charcoal. Sharp irregular and occasionally abrupt boundary to:-

1338

Orange brown fibrous peat Medium brown peat becoming orange brown downwards with a 20mm orange band at bottom. Mixed disturbed peat with grits and dried peat lumps.

Notes \ 1106

Calanais Survey and Excavation, 1979-88

Context Layer

Depth

1339

55 - 60cm

1340

60 - 68cm

1341

68-70cm

Layer 4. Uncemented green subsoil

1342

70 downwards

Green clay loam; with a much cemented top surface.

70 cm

Colour

Description

Layer 2. Medium brown slightly sandy silty clay with a slight yellow green tinge and diffuse indistinct dark brown mottles. Cemented but not as much as above. It contained 20% 2-6cm stones and 10% <1cm green stones all rotted but mineralogy retained. Abrupt lower horizon to:-

Layer 3. Uncemented although very compact green brown silty clay with some white rotted stones less than 1cm in diameter at the top. Vertical fine fibrous roots very common

2mm root mat.

Note 13.4.2 Pollen Samples 2036 and 2060 The records for Pollen Sample 2060 conflict (Tables 25.7 to 25.9). Table 25.7Â 1981 samples from GII and J = GIII according to the soil sample list Soil Sample

Test pit

Context

Description of sample and reason for sample

2046

GII

605

For soil and pollen analysis [This is an Area F context number]

2047

GII

662

For soil and pollen analysis [This is an Area F context number]

2055

GII

layer 3

For soil and pollen analysis

2058

GII

layer 4

Charcoal peat and fungus sub sample for soil and pollen analysis

2060

GII

layer 3

NW profile sample

2062

GII

layer 1

SE profile sample

2036

GIII

ly4

Sample for soil and pollen analysis

2037

GIII

ly5

Sample for soil and pollen analysis

2044

GIII

Charcoal peat and fungus sub sample for soil and pollen analysis Notes \ 1107

Calanais Survey and Excavation, 1979-88

Table 25.8 Details according to Bohncke (Chapter 21: Palaeoenvironment) Site Sample Feature Field interpretation G

2060

Lay-3

2036

Lay-4

81 horizon NW profile

B1g horizon in the pit.

Table 25.9 Details according to AOC (Chapter 22 Macroplant) Sample Box 2036 2060

Description

CSS Pit G3, Layer 4, SE 2023-2048 profile

CSS G3, layer 3, B?, NW 2049-2068 profile

If Soil Sample 2060 is from G2 Layer 3 as both Bohncke and the original Soil Sample list have it, and the sample numbering sequence in Table 25.7 suggests (Tables 25.7 and 25.8), the sample is from black peat. The overlying layer was black peat with some quartz grains and charcoal. The assignation to just after the CaN-2d to CaN-3a is slightly surprising given the much later date of the basal peat on the other side of the stone setting and at the Calanais Fields project where peat grew, in round terms, from 500 BC to 800 AD and from 250 BC to after 1700 AD. However, Pit G2 was on what appeared from air photographs CUP RA84 and RA85 of 1955 to have been cut into a squarish pre-modern cultivation area. Peat and other material may have been dug up elsewhere and dumped on this area and modified in the 19th century AD. If this sample is actually from G3 (=J) as the AOC Macroplant Bag record has it, it was from a very compact green brown silty clay. But that would mean an inversion because the pollen in Sample 2036 from J Layer 4 is assigned to the transition of zone CaN-3a to CaN-3b. That too would be somewhat surprising. The records for Soil Sample 2036 have a lesser problem. Pollen sample 2036 was taken from G3 (=J) layer 4 according to the bag, but there

is no layer 4 named in the J=G3 primary record although a layer 3 is specified. The layer in J under layer 3 was described as uncemented green subsoil and the information provided to Dr Bohncke described the sample as from a B1g horizon (although that label did not appear in the primary soil record). A ‘B’ horizon is illuviated subsoil in which fine material (in this case clay) has derived from organic-rich higher layers. A B1g layer is the gleyed form. The two descriptions could thus be complementary. If that is right it seems likely that the pollen in sample 2036 washed down from above along with the clay fraction forming this layer. 25.14 to 25.17 Notes for Chapters 14 (Area S) to 17 (The lithic assemblage) There are no Notes for Chapters 14 to 17 25.18 Notes for Chapter 18: The Pottery Assemblage Note 18.2.1 Number of sherds Note that the Catalogue has 1110 entries, of which one (Cat. No. 613) is void. The ‘official’ overall total of sherds and fragments is, however, taken as 1110 since at least one of the individual catalogue numbers covers more than one piece. The actual total of pieces is not significantly greater than 1110. Note 18.7.6: List of Beaker sherds identified as ‘miscellaneous (as opposed to ‘early, international-style’) fine Beaker’ Note 18.7.6 List of Beaker sherds identified as miscellaneous (as opposed to ‘early’, international-style’) fine Beaker The first table lists those sherds counted in the ‘conservative’ estimate of Minimum. Number of Individual (MNI) vessels – i.e. those where the identification as Beaker is based on diagnostic features such as rims, bases and decoration. The second table lists other sherds that count as ‘probable/possibles’. The entries are sorted by Area and, within that, by Cat. No. Entries grouped together indicate where the Notes \ 1108

Calanais Survey and Excavation, 1979-88

sherds are definitely/ probably/possibly from the same pot. Sherds that had been identified by Audrey Henshall as definite or possible fine Beaker are indicated in bold. An asterisk beside wall thickness indicates that one surface spalled off. i) Sherds included in the conservative MNI estimate Area

Cat. No. SF No.

Context

Shd Wt. Fabric Th (g)

ASH No.

Summary description

677

80.192 149

Poss rim, gently squared off, 5.0* 76 Unc upright or minimally everted; top slopes inwards

BIII

488

80.030 121

BIV

480–483 80.119 134

7.3 NC

BIV

409

81.399 863

7.2 45a

410

81.447 100

7.3 45b

Base-wall junction (base flat); est base diam c 80, so small pot.

4 sherds (of which 2 are refitted) from lower body of small, thin fine pot – Th mostly c 5.5; diam at this point c 100 mm. One sherd is from immediately above flat base.

409 = rim (pointed, rounded on I, minimally everted) from small Beaker; ERD c 100 mm. Deeply-incised horizontal grooves.

410 = neck sherd; diam at this point also c 100 mm. Deep incised horizontal grooves

413 = small neck sherd, poss from this pot or else same pot as Cat. No. 414 (ASH 46), As Henshall suggested; although grooves are more similar to those on ASH 45ab

413

80.096 315

6.5 46

BIVWX 450

81.199 859

6.6 49a

Rim, gently pointed, with low, rounded sub-rim cordon

BIVSX

451

81.612 141

5.7* 49b

Neck sherd with low, gently ridged cordon on E

BVSX

469–470 81.444 141

9.2* 56

2 sherds. Flat base with interior omphalos; EBD c80, so small pot

Notes \ 1109

Calanais Survey and Excavation, 1979-88

Area

BVSX

Cat. No. SF No.

869

414

415

416

457â&#x20AC;&#x201C;8

459

Context

81.137 141

80.175 315

80.157 315

80.097 315

80.159 332

Shd Wt. Fabric Th (g)

80.171 318.3 4

ASH No.

Summary description

Markedly curving neck sherd from small, thin, fine Beaker with everted rim; sherd too small to 4.5 76 crb estimate diameter. Not from ASH 39 Internat Beaker even though shares same small find no.

Rounded, slightly everted rim from thin, fine Beaker; too small to estimate diameter. E has fairly shallow horizontal incised lines. See above regarding whether Cat. No. 413 may be from the same pot

7.1 46

6.1 46

8.5 46

Curving neck sherd from thin, fine Beaker with splaying neck, decorated on E with closely-set, shallow, roughly horizontal incised lines (grooves). Probably NOT same pot as Cat. No. 414)

Body sherd from fairly thin, fine pot but not the same pots as the others listed under ASH 46.Deco: shallow incised lines on E, running in 2 directions: horizontal and diagonal, the latter not reaching as far as the horizontal line

7.1 50

2 sherds. Rimsherd rounded & everted, with prominent narrow horizontal sub-rim cordon. ERD 140 mm. Body sherd from same pot

8.4 51

Prob from neck of med to lg fine Beaker (diam at this point c 200). Low, fairly broad cordon which would have been a subrim cordon

Notes \ 1110

Calanais Survey and Excavation, 1979-88

Area

Cat. No. SF No.

Context

Shd Wt. Fabric Th (g)

ASH No.

Summary description 460 = 4 refitted sherds from curving, splaying neck of small, thin, fine Beaker; diam at this point poss 110â&#x20AC;&#x201C;130. Friable.

460

80.087 315

5.8 52a

BIVSX

461

81.614 141

5.9 52b

461 = curving neck sherd, prob from same pot as ASH 52a

10.5 57

Flat base from fairly fine, fairly small Beaker; EBD c 80; wall splays gently

10.6 58

2 refitted sherds from base and lower wall of small, fine, thinwalled Beaker; gentle pedestal, above which wall kinks out

471

80.078 315

472

80.180 315 + + 10 315 80.038

536

80.111 321.2A 2

5.6 65

Small, heavily abraded sherd from small, fine pot. Very gentle 7.0 76 crb low ridge may be a cordon, so sherd will be from neck

896

80.102 315

918â&#x20AC;&#x201C;9

81.047 315

932

543

998

81.254 360

81.083 735

81.121 739

Prob belly sherd from small, thin, fine pot; diam at this point c 150 mm. Decoration: fairly random stab impressions, some roughly circular, some irregularly-shaped. Fabric 1a but with few or no mica flecks

2 spalls, prob from neck of one or 6.5 76 crb two fine Beakers. Both may have cordons

Either gentle carination or (less 7.0 76 crb likely) neck and edge of rim, or base. Fabric 2 but soft

Small abraded body sherd with dot-stab impressions. Not identical to ASH 67 but could possibly have come from the same pot. (Treated here as if from a separate pot.)

8.2 68

Small abraded spall, either from flat base of small fine Beaker or 5.7* 76 crb E/MN rim flange; seems most likely to be Beaker base

Notes \ 1111

Calanais Survey and Excavation, 1979-88

Area

HII

Cat. No. SF No.

1004

Context

81.155 748

81.229 768

Shd Wt. Fabric Th (g) 1

ASH No.

Summary description

Small abraded sherd, either from neck and gentle carination or 6.6* 76 crb else from flat base and bottom of wall. Virtually inclusion-free but one angular frag of quartz/ite

5.9 NC

Rim, gently squared off, with top sloping very slightly towards I; too small to estimate rim diam

ii) Other sherds likely to be from fine Beakers: Area

Cat. No. SF No.

361

Context

80.144 120

Shd Wt. Fabric Th (g) 1

ASH No.

6.3 NC

Summary description Tiny sherd from thbin fine pot, prob. Beaker

Curving neck sherd from fine Beaker; diam at this point c 150 mm. Very similar to Cat. No. 744. Fabric 1 but minimally corky

BIII

747

81.676 9999

76 Unc, 5.9 ?fine Bkr

BIII

771

80.145 142

5.6 76 crb

BIII

814â&#x20AC;&#x201C;6

80.165 112

4.7 76 crb 3 refitted, prob neck sherd

BIII

828

80.055 121

BIV

383

81.184 814

5.5 NC

Small sherd from thin fine pot, 76 unc most probably Beaker.Paler than 6.9 ?EN most Beaker and E/MN pottery but not necessarily burnt

BIVWX 736

81.454 885

76 crb

Neck sherd; too small to estimate diam.

Small, heavily abraded body sherd, possibly from fine Beaker Body sherd, thin, fine pot, prob Beaker

Notes \ 1112

Calanais Survey and Excavation, 1979-88

Area

Cat. No. SF No.

344–5

Context

81.396 813

Shd Wt. Fabric Th (g) 4

ASH No.

Summary description 2 abraded spalls from fine pot, probably Beaker

5.2* NC

Belly sherd from small, fine, thin-walled pot; diameter at this 6.0 76 Unc point 110. Fabric 3 but slightly sandy

693

81.524 141

694

81.413 160

Body sherd; uncertain whether 5.7 76 unc a surface had spalled off. Rather gritty for a fine Beaker

76 Unc, 6.0 ?fine Bkr

76 Unc, 6.1 ?fine Bkr

Small body sherd from thin, fine pot

7.6 NC

Very small, heavily-abraded body sherd from thin, fine pot, prob Beaker. Fabric 2 but very few corky voids

Belly sherd from small, thin, fine 6.6 76 crb pot. Diam at this point c 150. Fine but gritty fabric

BVSX

744

745

862

540

81.162 813

81.448 812

81.430 874

81.191 141

BVSX

870

81.383 141

BVSX

881–3

81.409 812.1 1

BVSX

884–94

748

81.432 812.1 7

80.121 315

Neck sherd from small pot; diam at this point c 160 mm. Very similar to Cat. No. 747. Can’t rule out possibility that it’s from a very fine E/MN pot

Small abraded body sherd from fairly thin, fairly fine pot; can’t 8.1 76 crb rule out possibility that it’s E/ MN. Fabric 2 but not abundant mica

3 but fairly 7.5* 76 crb Prob from neck soft 3

5 small sherds + 6 frags from thin, fine pot; could be fine or 6.6 76 crb dom Beaker (but can’t rule out possibility that it’s E/MN)

76 Unc Small neck sherd from thin fine 7.2 ?fine Beaker Bkr Notes \ 1113

Calanais Survey and Excavation, 1979-88

Area

Cat. No. SF No.

Context

Shd Wt. Fabric Th (g)

ASH No.

Summary description

897

80.151 315

Small body sherd from small 6.7 76 crb thin fine pot; sherd too small to estimate diameter

898

80.173 315

6.3 76 crb

922

81.249 320

6.7 76 crb

939

81.237 369

5.9 76 crb

Small body sherd from fairly 7.9 76 crb thin, fairly fine pot; could well be Beaker

Body sherd from small, thin, fine pot; diam at this point c 150 mm

Very small sherd from thin fine pot

Poss from belly of small thin fine pot; diam at this point c 130

953

81.282 374

954

81.276 377

7.7 76 crb Small abraded body sherd

955

81.285 385

1(?)

6.3 76 crb

975

81.042 340

3.9* 76 crb

Small belly sherd from small thin fine pot, most prob Beaker

Spall, prob from neck or belly of fine Beaker

984

81.033 1108

Belly sherd from small thin fine pot; est siam at this point c 6.0 76 crb 100 mm. Fabric 2 but mica not abundant

639

80.005 1306

6.5 79

155

81.154 746

4.3 11

Body sherd from thin fine pot, most prob Beaker but canâ&#x20AC;&#x2122;t rule out possibility of E/MN date

Body sherd from thin, fine pot that could be Beaker rather than E/MN

992

81.062 707

Small spall, probably from flat 5.5* 76 crb base; too small to estimate diameter

HII

81.229 768

4.1* NC

Featureless spall from fine, thinwalled pot, poss Beaker

HII

81.229 768

3.6* NC

Spall from very fine, prob thinwalled pot

Notes \ 1114

Calanais Survey and Excavation, 1979-88

Note 18.7.7: List of Beaker sherds identified as ‘domestic Beaker’ The first table lists those sherds counted in the ‘conservative’ estimate of Minimum Number of Individual (MNI) vessels – i.e. those where the identification as Beaker is based on diagnostic features such as rims, bases and decoration. The second table lists other sherds that count

as ‘probable/possibles’. The entries are sorted by Area and, within that, by Cat. No.

Entries grouped together indicate where the sherds are definitely/ probably/possibly from the same pot. Sherds that had been identified by Audrey Henshall as definite or possible fine Beaker are indicated in bold. An asterisk beside wall thickness indicates that one surface spalled off.

i) Sherds included in the conservative MNI estimate of ‘domestic’ Beakers Area

HII

ConCat. No. SF No. text 541–2

672

1031

81.147 733

81.351 730

81.009 708

Shd Wt. Fabric Th (g) 1

ASH No.

5.5* 67

Summary description 2 small, abraded body spalls with same kind of dot-stab impressed decoration seen on ASH 65 and 66, but from a different pot

Gentle carination plus adjacent parts of neck and upper belly of very thin, very fine pot, prob Beaker. Soft; may well be burnt.

6.2 79

Small sherd from base-wall junction, prob small pot but sherd too 7.6 76 crb small to estimate diameter. Wall splays slightly

405

81.443 100

10.5 44a

406

81.198 837

10.1 44b

405 = Neck sherd, broken along ring joint to create ‘false rim’; diam at this point c 140 mm. Horizontal incised lines on E. 406 = belly sherd, poss as large as c 230 mm diam at this point. Could nevertheless possibly be from same pot as Cat. No. 405 if pot had had a funnel-like neck and globular belly (cf. ASH 48). Not much mica in inclusions

Notes \ 1115

Calanais Survey and Excavation, 1979-88

Area

Cat. No. SF No.

Context

Shd Wt. Fabric Th (g)

ASH No.

BIII

554

80.042 117

9.4* 72

555

81.396 813

9.2* 72

556

BIV

437-475 80.119 134

BVSX BIV

484 520

81.178 810

81.387 141 81.449 877

11.5 72

12.4 59.a

13 11

4 2

11.5 59b 9.8 NC

407

81.620 814

10.1 44c

537

80.061 117

9.9 66

538

81.397 810

9.6 66

BVSX

539

141

10.4 66

Summary description 554 + 555 = body sherds; 556 = neck sherd, from large pot with oval jabbed impressions, similar to those on Cat. Nos. 549–53. Could possibly be from same pot as 549–53. Diameter at neck c 200 mm

473–5 =Base + 2 body sherds from fairly large, flat-based dom Beaker; too small for EBD. Wall prob. splays widely. Fabric 4 but inclusions not very abundant.

484 = 2 refitted sherds from base, prob of same pot as ASH 59a; EBD c 110

Curving, splaying neck, prob from domestic Beaker; diam at this point c 190 mm

Probable base plate from large ‘domestic’ Beaker, but not necessarily the same pot as ASH 44ab. Not much mica in inclusions

537 = belly, near base of small ‘domestic’ Beaker; diam at this point c 120

538 = sherd from further up belly of same pot

539 = from further up belly than 537 & 538; diam at this point c 170. Small circular stab impressions, in irregular rows

Notes \ 1116

Calanais Survey and Excavation, 1979-88

Area

Cat. No. SF No.

Context

Shd Wt. Fabric Th (g)

ASH No.

BVWX

422

81.195 859

11.4 48

BVWX

423-424 81.379 859

11.4 48

425

80.79

426-427 80.86

428-429 80.88

430

80.93

859

315

431

80.120 315

438

80.186 326

432-437 80.123 326

439-449 80.186 326

8.5 48

8.2* 48

9.5 48

11.4 48

8.8 48

8.6 48

Sherds from rim, neck, carination and belly of large jar of unusual shape, with funnel-like neck and globular belly, as described in 18.4.6., Entry 3. ERD 190 mm; diam at carination possibly 260 mm. Decoration: 3+ deep, wide horizontal grooves under rim, plus impressions of marine shell edge, with wavy edge: 2 concentric lines on rim top; 2 lines immediately below rim and above grooves; and as horizontal lines below grooves.

10.9 48

Summary description

10.2 48

857-861 81.606 837

5.8* 48

BVSX

466

81.194 141

15.5 53.b

467

81.044 344

15.4 54

BVSX

485–7

81.387 141

12.1 NC

466 =Rim, rounded & everted, from large, thick-walled domestic Beaker; ERD 230–250 mm, with deep horizontal groove below rim on E, creating cordon-like feature. Definitely from same pot as ASH 54 and poss from same pot as ASH 53a, even though sub-rim groove is narrower and I is different colour.

467 = rim from same pot as ASH 53b and possibly same pot as ASH 53a; ERD 230–250 mm

Probably a base sherd

Notes \ 1117

Calanais Survey and Excavation, 1979-88

ii) Other sherds likely to be from ‘domestic’ Beakers (but see qualifications in some of the entries below)

Area

Cat. No. SF No.

Shd Con-text Wt. Fabric Th (g)

BNW

365

80.185

9999

7.4*

359

80.144

120

8.8

BIINE

579

80.029

112

10.8

BIVWX 602

81.380

859

11.1

Sherds possibly from large domestic Beaker, although cannot rule out later date.

602 = prob from neck of large pot; diam at this point c 210 mm. Hard, non-gritty, laminar fabric. Harder than other Calanais Beaker pottery, although from stratigraphy and sherd shape, may be dom Beaker.

603–604 = neck sherd; from same pot as 602 and probably also same pot as 607–9

BIVWX 603-4

BVSX

81.386

607-609 81.141

859

141

c15 3

BVSX

81.906

141

BVSX

81.384

141

81.616

160

758

11.0

10.4

ASH Summary description No.

11.2

c12 3

9.0

Body sherd, poss dom Beaker

Poss from neck, poss domestic Beaker Probably lower belly sherd; diam at this point c 180 mm

607–9 =3 refitted sherds from gently cuving neck of large pot of hard, non-gritty fabric. Diameter at this point c 220 mm

76 Unc Body or base sherd, poss poss Beaker; can’t rule out possidom bility that it’s of later date Bkr

Notes \ 1118

Calanais Survey and Excavation, 1979-88

Area

Cat. No. SF No.

Shd Con-text Wt. Fabric Th (g)

ASH Summary description No.

76 crb

879

81.373

812

8.1

BVSX

610

81.384

141

10.1* 79

BVSX

895

81.615

812.1

5.7

76 crb

Belly sherd from small, fairly thin-walled pot with fine-gritty fabric; est diam at this point c 160. Most likely to be dom Beaker but can’t wholly rule out possibility that it’s E/MN Featureless body sherd

Spall, prob from a flat base; could be either fine or dom Beaker

BVWX

611

81.416

160.3

8.2

Belly sherd from small, fairly thin-walled fine-gritty pot; diam at this point c 170 mm. Slightly grittier than the grittiest definite domestic Beaker from Calanais, but may still be Beaker

BVWX

612

81.416

160.3

9.2

Featureless sherd from pot with fairly hard, gritty fabric

BVWX

698

81.438

160.3

8.0

76 Small abraded belly sherd, Unc poss from thin dom Beaker

8.1

Upright, sharply squaredoff rimsherd from small, fairly thin-walled pot with hard, gritty fabric; ERD poss c 150 mm. Grittier than Beakers of comparable thickness but stratigraphy suggests that it could be mid-2nd millennium or older – in which case, Beaker is the most likely

4.1*

76 unc

595

706

81.278

80.181

376

315

Spalls from gently-curving neck of fairly large cooking pot

Notes \ 1119

Calanais Survey and Excavation, 1979-88

Area

DI DI

HII

Cat. No. SF No.

765 927

671

81.255 81.242

81.351

Shd Con-text Wt. Fabric Th (g)

ASH Summary description No.

376

76 Unc (?dom Bkr)

333

730

6 2

Belly sherd; diam at this point c 200 mm

10.5

76 10.0* crb

Belly sherd, prob from fairly small pot with hard fabric; I spalled off; could possibly be dom Beaker

9.4*

From fairly fine pot, possibly from carination area; possible shallow incised V on exterior. Note: cannot entirely rule out the possibility that this is a post-Early Bronze Age pot

Table of Misc Beakers by context Area

Ctxt

Dom No

Fine No

Dom wt (g)

Fine wt (g)

120

BI BI

BIII BIII BIII BIII BIII

BIINE BIV BIV BIV BIV

BIVSX BIVSX BIVSX BNW BV BV

100 149 112 117 121 142

9999 112 134 814 863 877 141 859 885

9999 117 141

1 0 0 1 0 0 0 1 1 0 0 1 0 2 0 1 1 0

1 1 1 0 2 1 1 0 1 1 1 0 2 1 1 0 0 1

15 0 0 2 0 0 0 8

30 0 0

11 0

31 0 2 4 0

6 1 1 0

10 1 2 0

16 1 7 0 3 2 1 0 0 4 Notes \ 1120

Calanais Survey and Excavation, 1979-88

Area

Ctxt

Dom No

Fine No

Dom wt (g)

Fine wt (g)

810

BV BV BV BV BV BV

BVSX BVSX

BVWX BVWX DI DI DI DI DI DI DI DI DI DI DI

160 812 813 814 837 874 141

812.1 160.3 859 315

318.3 320 326 332 333 360 369 374 376

HI HI HI

2 0 9 1 3 2

11 2 0 3 0 1 0 0 0 2

0 0 1 4 2 0 0

12 1 1 0 1 0 1 1 1 0

11 14 0

90 3

11 14 47 65 0

57 0 2 0 0 0

2 7 0 0 1

18 8 0 0

35 4 1 0 5 0 1 3 1 0

321.2

336

DII DV

385

377

DI DI

340 344

1108

1108 & 1112 1306 707 735 739

0 1 0 2 0 1 0 1 0 1

1 0 1 0 1 0 1 1 1 1

0 6 0

22 0

19 0

14 0 6

1 0 1 0 1 0 1 1 2 1 Notes \ 1121

Calanais Survey and Excavation, 1979-88

Area

Ctxt

Dom No

Fine No

Dom wt (g)

Fine wt (g)

748

HII HII HII HII

746 708 730 733 768

1 0 1 0 0

25.19 Notes for Chapter 19: Soils There are no notes for Chapter 19 apart from a reference to Technical Note 13.4.2 25.20 Notes for Chapter 20: The Vegetation Survey Note 20.1 Authors The records do not make it clear who produced the report on the vegetation. My memory is not clear on the matter. It is my spotty recall however that Marion Wade-Evans had considerable expertise and that Alan Fairweather helped her. It may have been Alan who produced the final typescript (preserved in the archive and to be found in digital form in Calanais Resources). A vegetation survey was conducted for HS years later and is presumably available from it. 25.21 Notes for Chapter 21: Palaeoenvironment Note 21.4.5: Section 105 recording The details of the section in Illus 21.14 differ from those on the original section drawing. They include information from photographs taken just before sampling (Films PO1 and PO2). The differences reflect cutting-back of the section for sampling. 25.22 Notes for Chapter 22: Macroplant Note 22.4.1 Errorst In the original report Sample 391 was numbered

1 1 1 1 3

1 0 4 0 0

1 1 2 1 2

139 which according to the context record is a mistake Sample 255 may actually have been 2255 as suggested by the context list. ‘Contexts’ 240 and 241 (marked as ‘not found’) do not exist. Nor are samples 240 and 241 in the context list. Finds 240 and 241 are pot. I have therefore removed those rows. Sample 672 does not exist in Context or Sample list. Small find 672 is listed as quartz. Sample 3688, said to be from context 369 and from McCullagh’s 1983 work, is labelled wrongly. Samples from context 369 listed in the context and sample lists are: 2567; —— 673/81 —— 678/81 —— 688/81 —— 693/81

charcoal sample - seen 23/10/95; charcoal sample - seen 23/10/95; charcoal sample - seen 23/10/95; charcoal sample - seen 23/10/95

I suppose the likeliest answer is that the identification was from sample 688. Sample 2060 is said to be from GIII but is more likely to be from GII (see Technical Note 13.4.2) Sample 2243 is from 736 not 733 according to the sample list; note that the McCullagh charcoal list has this right despite it being wrong in the original version of this document (corrected). Sample 2256 is not from 759 but from 795 according to the sample list (corrected). 2052 is said to be from D812; but 800-series numbers are from Area B. The full context list has sample 2052 from D360 (corrected). Small find 54 is a carbonised Pomoideae sp. Notes \ 1122

Calanais Survey and Excavation, 1979-88

charcoal twig from Area B 123 not 132 (corrected) According to the original finds book and the context list a piece of barley (75/81 barley (2.20/1.25)) was found in H728. According to the original finds book and the context list a piece of carbonised hazelnut 108/81 (1.18, 2.43/1.05) was found in H732. This report does not seem to include all of the McCullagh table rows (a copy of that table can be found in the digital archive). Uncontexted samples: Table 25.11 Uncontexted samples Sample Other

Notes Vol Sieved

Uncon- Site B, 2595 1 texted S ext, 3.5 Yes …. 141??? Uncon- Site B, 2595 2 texted S ext, 3.5 Yes …. 141??? Uncon- Site B, 2595 3 texted S ext, 5 …. 141???

>4mm char 0.04 0.08

Yes

CallanBH770/ ish Soil Find? n/a BP2 Samples 1x tin of CAL “peat 80/ samBL/001 ples for PA”

n/a

The full sample list has sample 2595 from context 141. It is credible that 3 sample bags were taken from this charcoal-rich soil [PJA] 25.23 Notes for Chapter 23: Radiocarbon Note 23.2.3 Bayesian modelling Several simple Bayesian models were developed.

After further stratigraphic analysis they were abandoned. The first listing is for a model using all of the dates; but it had few stratigraphic controls and was rapidly abandoned. Please note that the Phasing nomenclature here is different from that adopted for final publication. Listing 1: Script for the 6 Date Group model Sequence “Calanais all dates” { Boundary “Start 1”; Phase “1” { R_Date “SUERC-11588” 6295 35; R_Date “SUERC-11989” 6245 35; Interval “Span 1”; !calculates interval between !Start 1 and End 1 }; Boundary “End 1”; Interval “Interval 1 to 2”; Boundary “Start 2”; Phase “2” { R_Date “SUERC-11589” 4880 35; R_Date “SUERC-11601” 4760 35; Interval “Span 2”; }; Boundary “End 2”; Interval “Interval 2 to 3”; Boundary “Start 3”; Phase “3” { R_Date “SUERC-11600” 4515 35; R_Date “SUERC-11608” 4510 35; R_Date “SUERC-11596” 4495 35; R_Date “SUERC-11597” 4495 35; R_Date “SUERC-11607” 4490 35; R_Date “SUERC-11599” 4475 35; R_Date “SUERC-11612” 4475 35; R_Date “SUERC-11592” 4465 35; R_Date “SUERC-11606” 4455 35; R_Date “SUERC-11611” 4450 35; R_Date “SUERC-11618” 4450 35; R_Date “SUERC-11617” 4425 35; R_Date “SUERC-11616” 4430 35; Notes \ 1123

Calanais Survey and Excavation, 1979-88

R_Date “SUERC-11598” 4390 35; R_Date “AA-24965” 4385 50; Interval “Span 3”; }; Boundary “End 3”; Interval “Interval 3 to 4”; Boundary “Start 4”; Phase “4” { R_Date “AA-24966” 4210 50; R_Date “AA-24960” 4205 50; R_Date “AA-24970” 4205 45; R_Date “AA-24964” 4185 45; R_Date “AA-24959” 4140 45; R_Date “AA-24963” 4115 45; R_Date “AA-24969” 4095 45; R_Date “AA-24958” 4065 45; R_Date “AA-24961” 4055 50; R_Date “AA-24967” 4050 45; R_Date “SUERC-11590” 3965 35; R_Date “SUERC-11591” 3915 35; Interval “Span 4”; }; Boundary “End 4”; Interval “Interval 4 to 5”; Boundary “Start 5”; Phase “5” { R_Date “AA-24956” 3580 45; R_Date “AA-24968” 3575 45; R_Date “AA-24962” 3555 50; R_Date “AA-24957” 3495 45; Interval “Span 5”; }; Boundary “End 5”; Interval “Interval 5 to 6”; Boundary “Start 6”; Phase “6” { R_Date “SUERC-11610” 3220 35; R_Date “SUERC-11609” 3215 35; R_Date “SUERC-11602” 3195 35; Interval “Span 6”; }; Boundary “End 6”; }; Phase Group 4 model (30 Jan 07)

The second listing is for a model based only on contexts in the area of the Ring with stratigraphic controls, excluding material thought to be obviously residual. Again the phasing is different from that finally adopted. In this model AA-24969 was assigned to the pit of Ring stone 42. That was wrong and the sample actually came from secondary layers of the cairn. Listing 2: Script for Phase Group 4 model Sequence “Group 4” { Boundary “Start Phase 4a”; Phase “4a” { R_Date “AA-24970” 4205 45; Interval “Span 1”; !calculates interval between !Start 1 and End 1 }; Boundary “End 4a”; Interval “Interval 1 to 2”; Boundary “Start 4b”; Phase “4b” { R_Date “AA-24969” 4095 45; Interval “Span 2”; }; Boundary “End 4b”; Interval “Interval 4b to 4c”; Boundary “Start 4c”; Phase “4c” { R_Date “AA-24959” 4140 45; Interval “Span 3”; }; Boundary “End 4c”; Interval “Interval 4c to 4d”; Boundary “Start 4d”; Phase “4d” { R_Date “AA-24964” 4185 45; Interval “Span 4”; }; Boundary “End 4d”; Interval “Interval 4d to 4e”; Boundary “Start 4e”; Notes \ 1124

Calanais Survey and Excavation, 1979-88

Phase “4e” { R_Date “AA-24963” 4115 45; Interval “Span 5”; }; Boundary “End 4e”; Interval “Interval 4e to 4g”; Boundary “Start 4g”; Phase “4g” { R_Date “AA-24967” 4050 55; Interval “Span 6”; }; Boundary “End 4g”; }; This model and other similar models were abandoned in 2009 when further examination of the primary excavation records confirmed that a Beaker sherd was indeed securely stratified in the primary cairn. In 2011 and early 2012 identification of a few probable Beaker sherds in precairn contexts provided more support for dating the cairn to after 2500 BC. That increased the likelihood that all of the dated charcoal, except perhaps a fragile heather twig, was significantly residual to such an extent as to preclude Bayesian modelling based on the proposition that samples could be chronologically ordered by their stratigraphic relationships. 25.24 Notes for Chapter 24: Discussion and conclusions. The data set used here is derived from the 2005 Historic Scotland database and is called ‘C14dates 2005 plus.mdb’,. It includes the publically available C14dates 2005.mdb housed in NMRS and the data in the splendid lists published in and for the years between 2005 and 2011 in D&ES by Alison Sheridan and her collaborators. It does not include ages obtained by HS since it ceased publication of its lists in D&ES (Ashmore 2005). The latter are now available through NMRS Note 24.2.4 Ancient genes Claims of a human presence in the Western Isles and neighbouring areas from between 13,000 and

11,000 years BC have been made by Oppenheimer (2007, 129-31, Figures 3.3 to 3.5; 134-5, Figures 3.6a to 3.6d). These are based on the modern presence of gene groups which changed around that time, approximately the warm period before the Younger Dryas. In general the date of arrival of people in areas where a particular gene variation now occurs need not be the same as the date the variation first occurred. The genes might have arrived much later in genetically heterogeneous populations. In other words, a single ancient gene variation in a modern population provides a terminus post quem for the arrival of their ancestors, a marker-date some time, perhaps a long time, before the people immigrated. That is useless afor regional studies. But Oppenheimer’s analysis is more convincing (with the strong qualifications described below) when it is based on genetic variations which seem to have arisen locally, and he cites a localised gene variant which he claims arose in the Highlands before the Younger Dryas cold period (Oppenheimer 2007, 173-4 Fig 4.7). But it is important to remember that if the ranges of prey animals shifted northward with regional warming then dependent human populations may have moved after them en masse, taking a local gene variant with them and precluding it from the subsequent gene pool of the area in which it first occurred. That means that the gene variation may not have originated where it is found today and by extension its dating does not apply except as a terminus post quem in the area of its present distribution. Nevertheless there is some published archaeological evidence to support Oppenheimer’s dating of human activity in Highland Scotland to the period 13,000 to 11,000 BC (Chapter 24 Discussion 24.2.4). Overall it seems sensible to treat Oppenheimer’s claim that people who contributed to the modern local gene pool were settled in the Highlands of Scotland before the Younger Dryas as a fair working hypothesis. However many of Oppenheimer’s other conclusions seems to be based on genetic changes which took place elsewhere so it is far from clear when exactly those people arrived. There is another serious problem. Oppenheimer provides age estimates Notes \ 1125

Calanais Survey and Excavation, 1979-88

for c. 65 particular gene changes, giving the size of the standard deviations (or the standard errors) attached to them in various Notes. Generally the errors are around half the associated age estimates. In Illus 25.# his estimates for ages under c. 12,000 BC have been put in order of age (from left to right). The standard deviation associated with each age is shown by the red and green lines. Statistically, only 2/3 of true ages will actually lie in the blue-coloured area although alas it is uncertain which estimates are seriously wrong. But for instance an age estimate of 4000 BC (an age of c. 6000 before present) has a 1/6 chance of actually lying after 1000 BC. Most of his ages are in fact not significantly different from many of their neighbours, in the sense used by statisticians.

Illus 25. 13 The 1 sigma errors associated with Oppenheimer’s age estimates. In fact data can be ordered for any given lineage by adding information about which gene changes preceded others. So probably the situation created by the size of the errors on individual estimates is not as bad as Illus 24.# suggests. Nevertheless it is far worse than his text implies. Further, and presumably devastatingly for his date estimates, in 2012 it was reported that the best estimate for human mutation rates is half that previously assumed (Brahic 2012, 36). Presumably Oppenheimer’s chronological estimates will have to be reviewed and roughly doubled. I have therefore not mentioned his conclusions in the main text.

Note 24.4.3 Forager and Farmer interaction: the arguments offered in Thomas 2005 In what follows the arguments put forward by Thomas (2005, 125-6) are in italics. My comments are in regular type. 1. Britain and Ireland were not isolated in the Late Mesolithic. The presence of domesticated animals in late Mesolithic communities in Ireland undermines any suggestion that they were. Two dates for pre-farming cattle bones from Ferriter’s cave in Ireland have been discussed by Woodman and McCarthy (2003, 32-6). They suggested that cattle might have been prestige gifts or economic supplements to communities which, at least in terms of lithics, were remarkably insular (Woodman and McCarthy 2003, 36). They have since documented problems with the dates (Sheridan 2009) but they suggest those are not bad enough to invalidate all of them. In abstract terms, the bovines in Ireland in the 5th millennium BC may suggest at any one time either a succession of short distance links or a single long distance link. A large skin boat could undoubtedly carry cattle (Bowen 1972, 36). However, the proposition that cattle imports are evidence for a substantial amount of cross-cultural knowledge is otherwise weak. As Thomas himself pointed out people could adopt single traits of a culture while ignoring the rest. 2. Indigenous people will have been aware of the contents of the Neolithic package long before 4000 BC. The adoption of domesticates and novel forms of material culture cannot be explained by the arrival of boat loads of continental people. I have no quarrel with the first proposition, although the direct evidence is at present restricted to Ireland. But the latter argument seems to me a non-sequitur; no new evidence is adduced in support and there is little dated evidence for novel forms of material culture amongst basically indigenous cultures. Indeed Woodman and McCarthy have pointed to a lack of lithic evidence for contact between the continent and Ireland in this period (2003, 36). 3. It is unfeasible that the rapid spread of the Neolithic ‘package’ was triggered only by the arrival of a few boatloads of agriculturalists from NW Europe. Notes \ 1126

Calanais Survey and Excavation, 1979-88

For Thomas’ argument to succeed it has to be demonstrated that it was more difficult to borrow from (a few) people close to hand than from more distant people. Had he modified his argument to say merely that the rapid spread of the Neolithic ‘package’ would have been quicker if indigenous populations adopted farming from close neighbours it would have been much easier to agree with him. 4. There are only two ways to explain the rapidity of the transition: either a massive influx of farmers or its sudden and widespread adoption by indigenous groups. Any assertion that all but extreme explanations should be excluded is intellectually unacceptable. Sheridan has documented the proposition that the introduction of farming need not have been a unitary phenomenon, nor need it have happened abruptly around 4000 BC (Sheridan 2009, 1). That said, even the strongest proponents of the idea that farming was introduced to Britain by small groups of colonists accept that indigenous people may have picked up the idea of farming from their new neighbours (Sheridan 2009; Pailler and Sheridan 2009). 5. In the absence of any single donor population on the continent the only realistic possibility is that indigenous people in Britain, Ireland and Scandinavia adopted farming over a period of two centuries. This should be read in the context of Thomas’ earlier acceptance that the material cultures of the first farmers in Britain showed similarities to those in neighbouring parts of the continent (Thomas 2005, 118; Whittle 1977). The material culture of the indigenous people seems to have been fairly uniform even though some variations in the use of lithic resources may reflect differences (Wickham-Jones 1994, 97) and it is envisaged that they maintained widespread social networks, although that is very largely a presumption based on ethnographic analogies and the lack of variation in lithic assemblages. Turning Thomas’ argument on its head, if indigenous cultures were homogenous around 4000 BC, the subsequent regionalisation of Britain can most easily be explained by (local) contacts with groups of incoming farmers from different parts of

the continent (Pailler and Sheridan 2009, 47-8). 6. Given a long period of interaction between insular foragers and continental farmers, the likeliest explanation for insular change is that the practises of the farmers were good tools for creating and maintaining group identities [and thus highly attractive to insular foragers]. This is a sophisticated argument best seen against a background of an apparent general uniformity in insular forager material cultures. But that evidence is based on not much more than lithics. The claim that foragers could not maintain group identities in ways as attractive as those offered by farming is somewhat audacious. 7. The differences between the early farming material cultures of Britain and Ireland reflect different adoption pathways determined by cultural differences in the pre-farming period. It would be equally easy to argue that the differences between the material cultures of the farmers in Britain and Ireland owed much both to differences in environments and the different origins of farming groups colonising different regions. In other words the differences in both periods may have owed much to environmental and social factors with many inputs, amongst which those of incoming farmers may have played a large part. 7 and 8) The British indigenes adopted pastoral lifestyles because they were used to managing large prey animals. The Irish indigenes, lacking that heritage, adopted more sedentary settlement patterns based on houses. The large British halls were not houses. Thereafter the shift in Britain to sedentary patterns was gradual. Equally one could argue that incoming farming groups learnt from indigenous cultures. The possible parallel cited by Thomas, quoting Richards (2003, 33), of the Norse settlement of Greenland can be used in the opposite way that Thomas did (2005, 117), to suggest adoption of insular subsistence practices by incomers rather than changes amongst the indigenous people. Even the strongest proponents of the idea that farming was introduced to Britain by small groups of colonists accept that indigenous people may have picked up the idea of farming from their new neighbours and spread it to other groups (Sheridan 2009; Pailler and Sheridan 2009). But the barNotes \ 1127

Calanais Survey and Excavation, 1979-88

riers to adoption of cereal cultivation in Scotland should not be underestimated. Ester Boserup in her ‘The conditions of agricultural growth’ showed that the introduction of farming and its intensification are normally dependent on population densities rather than the opposite (1995, 11-14). She showed that most forms of farming are much more laborious than foraging. The lowest input systems involve secondary forest in which fire does most of the work. Bush clearance systems are more labour-intensive. In more intensive systems the need for frequent weeding and grass root clearance makes for even harder work (Boserup 1995, 15, 29-31). That makes the adoption of intensive farming techniques by foragers unlikely in the absence of population pressures (the same is not true of non-intensive cereal planting and some modern foragers do not distinguish between gathering and planting of crops (Zvelebil 1998, 6)). Other important factors may have been in play in the period when farming began in Britain and Ireland. There is a fairly strong possibility that incomers living with domestic animals of Near-Eastern origin brought highly infectious diseases with them, or foragers visiting farmers brought sickness back. Diamond has suggested that the rise of agriculture (and the keeping of domesticated animals in close proximity) in the Near East launched the evolution of crowd-infectious diseases. “Infectious diseases, like smallpox, measles and flu arose as specialized germs of humans, derived by mutations of very similar ancestral germs that had infected animals” (Diamond 2005, 92). He listed the animal pathogens which leapt to humans (ibid, 207). There are many recorded examples of high death rates amongst hunter-gatherer groups through disease after contact with farmers (ibid, 87 Fig 4.1). For example “… in the winter of 1902 a dysentery epidemic brought by a sailor … killed 51 out of the 56 Sadlermiut Eskimos, a very isolated band of people …” (ibid, 204). It is possible that epidemic diseases weakened the ability of indigenous foragers to defend their territories from incoming farmers. It is even possible that some Mesolithic populations were drastically reduced by diseases such as poxes before farmers settled the areas they had inhabited (Holtsby et al. 2012, 2078).That might allow groups of farmers to expand

as fast as or faster than new techniques could be spread by groups of indigenes. Inter-group violence may have been a counter-acting factor. In modern hunter-gatherer groups about 10% of deaths are from violence, which usually takes the form of raiding. There seems little doubt that small isolated groups of incomers would have been vulnerable, and that might be taken as an argument against the likelihood of a successful proliferation of groups of farmers. Note 24.4.4 Pottery from contexts inside the Ring with Phase 3 charcoal This is also referred to from Chapter 23 Radiocarbon 23.5.6. The 13 Phase 3 dates came from 9 contexts, of which 4 were in Area D, 3 in Area H, one in Area S and one under the passage wall in Area BIVWX. Of these, contexts 369 and 398 also produced later radiocarbon-dated charcoal. The following contexts produced dates but no pottery: D398 green clay under cairn, B883 greygreen clay under passage wall, H772 dark grey/ green fill of slot 773 under cairn and H778 gritty very humic middle fill of slot 773. But the lowest fill of this slot produced two probable Beaker sherds (Cat 671 and 672). Five contexts produced potsherds, predominantly Early/Middle Neolithic, although indeterminate sherds were fairly common and some Beaker sherds were also found (Table 25.12). Context 352 was black sandy clay south of the cairn. Context 360 was greasy orange/brown clay at the base of the cairn near the outer face of the chamber wall. It included a fine Beaker sherd (Cat 932). Context 369 was a layer of dark greasy clay rich in charcoal fragments which survived just south of the cairn in a shallow hollow beneath plough soil 315. It included a sherd from a small thin fine pot, probably a fine Beaker (Cat 939). Context 746 was a light grey-brown slightly sandy charcoal-rich clay, with a variable texture from friable to greasy. It immediately underlay the locally uppermost turf line, although technically unrelated to more widely spread layers during excavation it was thought to be contemporaneous with turf line 771, possibly Notes \ 1128

Calanais Survey and Excavation, 1979-88

equivalent to a turf line on the green clay platform under the cairn. It included a sherd possibly from a fine Beaker (Cat 155) and a small sherd probably from a domestic Beaker (Cat 523). Plough soil 1213 on Area S produced numerous small sherds (Cat 1049-1103) all from one E/MN corky pot.

The most likely implication of this pattern is that clay containing Phase 3 charcoal and E/MN pottery was dumped in the area before the cairn was built but after Beakers had reached the area. One sherd group (Cat 759-761 could have been from Domestic Beaker or very much later. Stratigraphically the latter is highly unlikely.

Table 25.12Â Pottery from contexts inside the Ring with Phase 3 charcoal Cat

Area Ctxt

352

116

352

114_115 DI

352

Shd Nos

Date

Pot type Comments

1 2

E/MN

Corky

E/MN

Corky

928_931 DI

352

3& 1F & Indet; crbs

620

360

Corky

Indet

Rim or rim flange from thin fine pot. From abraded relatively thick pot.

Heavily bunt and abraded spalls from a conceivably E/MN pot.

Prob E/ NonMN corky

Belly sherd from large pot with virtually all of exterior spalled off.

Chalco/ Fine EBA Beaker

From thin fine Beaker; Area given as DV on original finds bag.

Non?E/MN corky/ Belly sherd from large pot Heb Inc

621

360

932

360

7+ 2F + E/MN sev crbs

Corky

From thin, fine pot.

30_38

369

39_43

369

4+ 1F

E/MN

Corky

From fairly fine pot.

117

369

E/MN

Corky

118

369

E/MN

119

369

E/MN

Noncorky

From fairly fine pot; much of interior spalled off

222

369

E/MN

Noncorky

339

369

E/MN

Heb Inc Narrow incised herringbone lines on exterior

Corky

Featureless body sherd Much of interior spalled off. Rim rounded, upright.

Notes \ 1129

Calanais Survey and Excavation, 1979-88

Cat

Area Ctxt

Shd Nos

712

369

713

369

759_761 DI

Date

Pot type Comments

E/ MN or Chalco/ EBA

Noncorky or dom Beaker

Prob E/ NonMN corky

Small abraded spall

?Chalco/ EBA or postEBA â&#x20AC;&#x201C; poss 1st millennium AD?

Dom Beaker or much later pot type

May well be from same pot as Cat. 752â&#x20AC;&#x201C;757 form B 141, interpreted as a plough soil. Chalco/EBA Domestic Beaker or much later pot type, poss 1st millennium AD

Indet

Featureless body sherds, burnt; soft and friable.

Indet;

Indet

Small abraded spall.

369

762_764 DI

369

935

369

2& 1F & crbs

936_938 DI

369

Indet

939

369

Prob Chalco

Prob fine Beaker

940_944 DI

369

Prob E/ Corky MN

143_154 HI

746

10 + 2F + E/MN crbs

155

746

? ? fine Chalco/ Beaker EBA

522

746

E/MN

523

746

Prob Prob Chalco/ Beaker EBA

31+ 24 F 1213 E/MN plus crbs

10491103

Small sherd from cooking pot

Corky

Heb Inc

Corky

Three tiny heavily abraded and quite possibly burnt fragments. From small thin fine pot, probably a fine Beaker. . Body sherds (including several spalls) from pot with laminar, friable fabric. Body sherd from thin, fine pot, possibly fine Beaker. Rim flange from large, thin, fine Hebridean Incised pot. Small sherd probably from a domestic Beaker. All probably from a single medium-sized cooking pot of corky fabric; a Notes \ 1130

Calanais Survey and Excavation, 1979-88

Note 24.4.5a The duration of farming in pollen zone 2c and the duration of the subsequent period lacking cereal pollen The part of pollen zone 2c including cereal pollen was represented by 5 cm of peat, and judging by that and the peat column radiocarbon dates and interpolated dates this period probably lasted between 385 and 585 years before farming stopped locally at some date between 3490 and 3020 cal BC. Similarly the subsequent Zone 2d during which pollen did not include cereals was represented by 4 cm of peat and the simplest interpretation is that farming ceased locally for between 430 and 520 years. However, each analysed block of peat represented between 80 and 120 years of growth and it is not certain where within each block the cereal pollen grains were preserved. Therefore the lowest estimates of duration should probably be reduced by about a century; cereals were grown for between 285 and 585 years and the subsequent period without evidence for cereal growing lasted between 330 and 520 years. But even though the calculation of these durations is beset with uncertainties both the period during which peat was grown and the subsequent cereal-pollen-free period seem to have lasted for many generations. It is also possible that the lack of cereal pollen in the peat during Sub-zone 2d was caused by a local thickening of the birch woodland canopy (Chapter 21 Palaeoenvironment 21.3.6). Note 24.4.5b various possibilities relating to the introduction of farming to the Calanais area Listed below are some possibilities and melanges of the possibilities relating to the introduction of farming in the area around Calanais. 1. Small groups of farmers may have settled around inner Loch Roag from neighbouring communities at some date between 3875 and 3605 BC and stopped cereal cultivation after a few centuries with reintroduction of farming only many generations later

a. because they came to concentrate on foraging; b. because they moved the focus of their farming; c. because they died off locally. 2. Farmers may have settled from much further away, leapfrogging intervening communities, with the various sub-possibilities of Possibility 1. 3. Sea-farers may have introduced the idea of growing crops. a. Planting and harvesting them to help provide a secure way-point in their travels north and south, with more or less permanent settlement following. b. Sea farers may, without themselves settling, have introduced the ideas and seeds which allowed indigenous peoples to broaden their subsistence strategies by including farming. 4. Farming may have been introduced locally by indigenous people. a. Indigenous people may have adopted cereal growing and other practises as part of a chain of changes amongst indigenous groups. b. Indigenous people may have visited neighbouring farming communities, and brought back agricultural techniques, and domestic animals, and possibly also spouses who supported new practices such as pot-making. 5. Large groups of farmers from Scotland, Ireland or north-west England may have settled. Possibility 1 The possibility of immigration by farmers is preferred largely because Schulting and Richards have demonstrated a fairly abrupt change of diet in western Scotland around 3800 BC from a reliance on marine sources to a reliance on terrestrial sources of food (Schulting and Richards 2002). That contrasts with what happened in some other Notes \ 1131

Calanais Survey and Excavation, 1979-88

parts of Europe (Lightfoot et al 2011, 73-4, 83). The domination of domestic species in early Neolithic assemblages of plant remains in Atlantic Scotland tends to support the idea that there was a fairly drastic change although in some areas assemblages of plant remains suggest that a reliance on wild food resources continued for many centuries (Bishop et al 2009, 72). Admittedly, for everyone who interprets the evidence as suggesting a change in population there is someone who believes that it demonstrates no more than a change in diet. If farming had thriven in mainland Scotland from about 3800 cal BC it is credible that resource pressures had grown and that people sought new land rather than improving the productivity of their areas of origin (Boserup 1995). That would fit well with the idea that small groups of farmers moved from time to time to previously unexploited areas. The likelihood that agricultural technology was lost at least locally after some 10 to 25 generations of farming is based on the evidence from Calanais Leobag described above in Chapter 24.4.4. Possibility 1a is that the local farmers at some date between 3490 and 3020 BC took up an â&#x20AC;&#x2DC;indigenousâ&#x20AC;&#x2122; life style based largely on hunting and gathering, only for farming to be reintroduced 12 and 20 generations later at a date between 3000 and 2500 cal BC. Option 1b differs from the first in that it places less weight on the gap in the local cereal record; for cereal pollen does not disperse very widely and it is possible that abandonment of farming close to Leobag was balanced by increased activity not far away. Option 1c is based on an idea that the local community was successful for many generations but remained small and eventually succumbed to some accident. Possibility 2 The second possibility is based on the ideas resuscitated by Sheridan (2009). She showed that people from farming communities reached at least as far north as the Irish Channel in the centuries before 4000 BC and it is a formal possibility that their forays, migratory or intermittent, reached

even further north. The extension of this idea to fit the evidence from Calanais is that subsequently other groups did something similar. Possibility 3 The third possibility, that sea-farers initiated farming (either settling people or coming to an arrangement with indigenous groups) to help supply their voyages, rests on a very loose analogy with Greek pioneers in the Black Sea area, and it is included only with hesitation. It is interesting because it allows for the possibility that those responsible for agriculture at Calanais regarded the area mainly as a place to supply voyagers rather than as a new home. Possibility 4 The fourth possibilities, two versions of adoption of farming by local hunter-gatherers, probably have a lower likelihood because if Boserupâ&#x20AC;&#x2122;s observations (1995), admittedly largely developed in very different environments, are accepted there would have been little in it for them. Factors including the diminishing size of the Great Western Island may have concentrated pre-existing populations and influenced them to take up new lifestyles with effects similar to those in areas near Doggerland (Spinney 2008, 43). However size of landmass alone may have been less important to foragers than length and nature of the coastline and the length and variety of habitats may even have increased as sea level rose if the coast became more indented (Illus 24.3, 24.4). Possibility 5 The last item in the list, a large influx of farmers over a short period, is meant merely to widen the options. There is little reason for supposing it reflects the reality of the introduction of farming to the Calanais area. Few chambered tombs are known from the area and more might have been expected if there had been a large population of farmers in the middle third of the 4th millennium BC. Other aspects Four possibilities can be mixed with the above suggestions.They are not in any particular order of preference. Notes \ 1132

Calanais Survey and Excavation, 1979-88

1. Farmers may rapidly have displaced indigenous foragers through hostile action or disease.

3. The volume of packing stones in the pits might have released about the same amount of clay.

2. Farmers and foragers may have coexisted in largely separate groups.

4. The total volume of clay displaced might thus have been about 3 to 6.5 cubic metres.

3. Farmers and foragers may have lived in a social grouping combining aspects of both traditions.

The Monolith pit as a clay source: 1. The stone had a cross-section area of about a third of a square metre

4. Foragers may have displaced farmers through hostile action. There is no direct archaeological evidence from around Calanais for or against these supplementary options. The best prospect for finding relevant evidence, given the current paucity of plough disturbance in the area, is exploration of the shallow waters of inner East Loch Roag. Note 24.6.2 Cereals at Machrie North. At Machrie North pollen analysis of a nearby column produced cereal pollen with an estimated date of 5375 BP (Robinson and Dickson 1993, 117). The uncorrected errors of the adjacent dates were respectively 85 and 130 C14 years and they have to be corrected by multiplying them by 1.4 (Ashmore et al 2000); the error attached to the interpolated estimate is the square root of the sum of the squared (corrected) errors, +/-217. If this is applied the date of the first appearance of cereal pollen calibrates to between c. 4700 and 3700 cal BC. Note 24.11.2 Soil and clay volumes The Ring stone pits as a clay source: 1. There were 13 Ring stone pits. The cross-sectional areas of the Ring stones on the Glasgow University plan varied between 0.16 and 0.36 sq m, and their total plan area was about 3.25 sq m. 2. If the Ring stones were set into the ground to a depth of between about 0.5 and about 1m, then the volume of stone in the ground, and thus the volume of clay displaced would have been between approximately 1.5 and 3.25 cubic metres.

2. If it was buried to (at least) a metre then the volume of stone in the ground, and thus the volume of clay displaced would have been between about 0.3 cubic metres 3. The volume of packing stones in the pits might have released about the same amount of clay. 4. The total volume of clay displaced might thus have been about 0.4 to 0.6 cubic metres Soils between lowest and middle turf line on Area H (soil 777 in HII above turf line 766 and below turf line 758, soil 756 in HI above turf line 766 and below turf line 751): 1. Soil 777 averages between 0.03 and 0.04m thick on Section 33a and 34a 2. Soil 756 also between 0.03 and 0.04m thick on Section 13 3. Area inside Ring 130 sq m. 4. If this soil spread over the entire Ring its volume would be 130 x 0.03 = 3.9 cubic metres to 130 x 0.04 = 5.2 cubic metres. The volume of soils 777 and 756 if they did spread over the whole area of the Ring would have been about the same as that released by digging the 13 Ring stone holes, erecting the stones and putting packers and clay back in the pits. Clay 760 under the cairn on Area H (the chambered cairn had an area of roughly 27 sq m. The underlying green clay platform was only slightly larger):

Notes \ 1133

Calanais Survey and Excavation, 1979-88

1. The green clay platform seen on Sections 33A and 34A was on average 0.1 to 0.15m deep. 2. The upper green clay under the kerb slabs on Section 33D was probably green clay 760 rather than 750. It was about 0.1m deep where not truncated by the slot under the largest kerb stone 3. The volume of the green clay platform was 27 x 0.1 = 2.7 cubic metres to 27 x 0.15 = 4 cubic metres. The volume of the green clay platform would have been less than that released by digging the 13 Ring stone holes, erecting the stones and putting packers and clay back in the pits. But there would not have been enough clay solely from the more local pits. Clay 750 to its north of the cairn on Area H and clay 320 on Area D 1. Clay 750 was found on less than half of the area of Area H. Its maximum recorded depth (on Section 13) was about 0.06m and on Section 33D (under the kerb slabs) where it underlay clay 760 it was about 0.05m deep. Generally its depth was probably nearer on average to 0.04m over less than half of the area north of the cairn. 2. So say its average over the whole area north of the cairn was 0.01 to 0.03m. 3. Clay 320 on Area D was found on three quarters of the area south of the cairn (and not at all recognisably under the cairn there).It did not appear on any sections unless it was the same as 950 which was 0.05 to 0.01m thick according to the Kubiena box column of Section 105 in the Palaeoenvironment report 4. Area B had various spreads of green clay but there was too much destruction there for an estimate of its original volume. 5. Area S was dug only to a shallow depth and did not reach the relevant stratigraphic levels. 6. Suppose there was a fairly even spread of clay

about 0.04m to 0.06m thick in the area excluding the cairn. That area was 130 sq m minus 27 sq m, = 103 sq m. 7. The volume of clay would have been 103 x 0.04 = 4.1 cubic metres to 103 x 0.06 = 6.2 cubic metres. That is at the upper end of the estimate for the volume of clay which might have been released from the pits. The volume of all clays of 320, 750 and 760 type 1. If clays of 320, 750 and 760 type at this approximate level covered the whole Ring area they would have had a volume between (2.7 to 4 + 4.1 to 6.2 cubic metres) or between 6.8 and 10.2 cubic metres. 2. The Ring pits would not have provided enough green clay for this volume of clay. 24.12.8 Northern and western Scottish chambered cairns. Various regional groups of chambered cairns have been recognised in Scotland (Henshall 1963; 1972). It is perfectly clear that there are overall differences between different parts of Scotland. Chambered cairns varied considerably in size and architectural form and indeed there were many areas in which chambered cairns were not built at all (Ashmore 1996, 57; Barclay 1997, 137-9). There was also considerable variation amongst chambered cairns within very small geographical areas. For instance the chambered cairn at Midhowe, Rousay, 32.5 m long with a chamber 23.4 m long by 2 to 2.2 m wide may have had a different function from Kierfea Hill, Rousay, 9.5 m across with a chamber less than 4 m long (Davidson and Henshall 1989, 146-7; 130-1). In most areas at least two popular chamber types occur. For instance, in Orkney there are at (at least) Orkney-Cromarty and Maes Howe style structures, but the former in particular includes many variants (Davidson and Henshall 1989). In Caithness where chambers morphologically similar to the Calanais example are known, three broad types can be identified: Notes \ 1134

Calanais Survey and Excavation, 1979-88

1. single chambers without subdividing slabs, 2. bipartite chambers with a large baggy end compartment 3. long chambers with subdividing slabs, including Henshall’s Yarrows, Camster and Assery types (Davidson & Henshall 1991, 21). In plan, there is a basic subdivision between baggy bipartite chambers and neat rectangular chambers with subdivisions. But hybrids and variants do occur, and in three dimensions the differences between the two main types appear little more significant than variations within the types. The cairns themselves show abundant evidence of extension. They fall into four basic types, round, long, horned and double-horned. There is no correlation between chamber types and cairn types (ibid. 56 Figure 17). Overall the impression is of farmers picking from a bagful of ideas within traditions which did not include some possibilities (such as the horizontal stone sills used to divide up Clyde cairns in western Scotland). In west Sutherland there are clusters on the limestone areas of the straths around Inchnadamph, but most cairns lie on the east coast and in the straths leading inland from it (Henshall and Ritchie 1995, 15-17). Many of the chambers are of types reminiscent of those in Caithness, including single chambers and bipartite chambers with a large baggy end compartment. The baggy-ended Skelpick type of chamber is also relatively common. In the harsh lands of western Inverness and Wester Ross only one chambered cairn is known on the western mainland (Henshall and Ritchie 2001, 29). Overall, if there were links between Lewis and the mainland north of Skye they seem likely to have been with the extreme northwest. In Southwest Scotland there is again evidence for two different types of chambered cairn if one ignores long cairns, the Bargrennan type in Dumfries and Galloway and the Clyde type, which has a wider distribution. The Bargrennan type had small chambers and long passages, usually in a small round cairn. The Clyde type had long chambers opening directly through a facade, often in elongated cairns. Many chambered cairns of

the Clyde type have several chambers in a single mound; each may originally have been surrounded by a small cairn. As others have noted the classification into Bargrennan and Clyde cairns probably oversimplifies a complex relationship between various styles of chamber and cairn. Two cairns were preceded by wooden mortuary enclosures (Henshall 1972; Cummings 2002, 129-130; 139-142). The chambered cairns in the Western Isles are even more heterogeneous than those of the northern isles (Müller 1988, 19-21). Most of them are in North Uist, in the middle island of the Long Isle. Only the use of massive peristaliths, paralleled in Ireland, seems to be a local particularity in a Scottish context; otherwise the characteristics of the chambered cairn are shared with those of the mainland from south-west to north-west. Within this variety, the chamber of the cairn at Calanais is very different from the rest, having a plan reminiscent of (but not identical to) those of some chambers in Caithness and Orkney. There is not enough dating evidence to show whether these differences have any chronological significance. The evidence from most areas could be taken to suggest at least two overlapping traditions or cultural predilections. This suggests a long and complex history, very little of which has been elucidated. On the other hand it has been contended that modifications to monuments took place with an apparent broad continuity of purpose in many parts of Britain and Europe; the funerary traditions at single sites changed significantly but they continued to be used for burials. For instance Bradley argued that Neolithic tumuli at Carnac, Brittany were enlarged and passage graves built in their sides to expand one kind of monument and assimilate another; and that some long barrows on the river terraces of England were converted to round mounds so that ‘their basic affinities were altered from a well established local form of mortuary monument to a tradition of round barrows with quite different symbolic and geographical references’ (Bradley 1993, 100). Another example is the chambered cairn at Achnachreebeag in Argyll which started off with an ‘exotic’ small closed chamber, comparable to some examples in Brittany (Pailler and Sheridan 2009, 35), and was enlarged to support a small Notes \ 1135

Calanais Survey and Excavation, 1979-88

passage grave of a kind found fairly widely in the Hebrides and less frequently in Argyll (Ritchie 1970, 32-37). As Pailler and Sheridan note the Breton-style pottery did not come from the early Breton-style chamber but from the small passage grave, its position perhaps reflecting manipulation of ancestral material as at Calanais. Manipulation of ancestral material may have taken other forms. Skeletal material could have been stored elsewhere (for instance in a midden, grave or cave) before deposition. Understanding what the chambers in cairns meant to those who built and used them is a problem because the function of chambered cairns may have changed over time, and there is no proof that they were all built primarily to take burial deposits. Instead some of the chambers may have been built for â&#x20AC;&#x2DC;hiddenâ&#x20AC;&#x2122; ceremonies and only subordinately or subsequently have been used for storing skeletal material.

Note 24.12.9 Dating Scottish chambered tombs. The earliest C14 ages for human bone in chambered tombs come from N Ayrshire, Argyll and Bute, Caithness and Orkney. They suggest that burials were being made throughout most of the distribution area of chambered tombs from some date between c. 3600 and c.3350 cal BC. This chronological distribution to some extent reflects patterns of investigation and preservation conditions, and is based on a very small data set. Also, if remains were cleared out from time to time the dated bones may not have been the earliest put into the chambers. But the evidence currently available suggests that the practise of putting skeletal material in massive chambers with an access passage started at much the same time throughout western and northern Scotland, around 3500 BC (Table 25.13, Illus 25.22).

Table 25.13Â C14-dates for human bone with weaknesses less than 20 from chambered cairns in Scotland Site

Council

Reference

Description

Context type

Code

Date cal BC

Age BP

+/-

Clachaig, Arran

North Ayrshire

Schulting, R

A human adult mandible

chamber

GrA-25616

3660 to 3380

4800

Tulloch of Assery A

Highland

Sharples 1986

Disarticulated human bone

chamber

GU-1338

3950 to 3300

4800

110

Torlin

North Ayrshire

Schulting, R

Human adult cranium

chamber

GrA-25644

3650 to 3380

4775

Torlin

North Ayrshire

Schulting, R

Human adult

chamber

GrA-25646

3650 to 3380

4770

Crarae

Argyll & Bute

Scott 1961; Schulting Human phalanx chamber 2002, 160;

OxA-7662

3640 to 3370

4735

Holm of Papa West- Orkney ray North

Schulting, R

Human subadult phalange

GrA-25636

3640 to 3370

4715

Clachaig, Arran

Sheridan 2006, 205

Unburnt human chamber cranium

UB-6898

3640 to 3370

4708

Sheridan 2006, 204

Human mandible

UB-7010

3630 to 3360

4695

North Ayrshire

Rattar East Highland

chamber

Notes \ 1136

Calanais Survey and Excavation, 1979-88

Council

Reference

Description

Context type

Code

Date cal BC

Age BP

+/-

Holm of Papa West- Orkney ray North

Schulting, R

Human adult right femur

chamber

GrA-25638

3630 to 3360

4690

Tulach an tâ&#x20AC;&#x2122; Highland Sionnaich

Sharples 1986

Human bone.

chamber

GU-1334

3700 to 3050

4685

110

Point of Cott, Westray

Orkney

Barber 1997b, 59

Human bone chamber from Skeleton E

UtC-1660

3630 to 3360

4680

Point of Cott, Westray

Orkney

Barber 1997b, 59

Human bone chamber from Skeleton B

UtC-1658

3630 to 3360

4680

Clachaig, Arran

North Ayrshire

Schulting, R

A human adult cranium

chamber

GrA-25617

3630 to 3360

4670

Haylie House

North Ayrshire

Schulting, R

Human adult cranium

chamber

GrA-25643

3630 to 3350

4665

Embo

Highland

Sheridan 2006, 205

Human adult right talus

chamber

UB-6879

3520 to 3350

4645

Holm of Papa West- Orkney ray North

Schulting, R

Human adult right femur

chamber

GrA-25637

3620 to 3340

4640

Embo

Highland

Sheridan 2006, 205

Human left femur

chamber

UB-6877

3520 to 3350

4633

Point of Cott, Westray

Orkney

Barber 1997b, 59

Human bone chamber from Skeleton D

UtC-1659

3520 to 3100

4600

Quanterness

Orkney

Renfrew et al 1976; 1979

Human bone.

chamber

Q-1363

3650 to 2850

4540

155

Point of Cott, Westray

Orkney

Barber 1997b

Human bone from an infant burial

chamber

AA-11697

3370 to 2930

4505

Embo

Highland

Sheridan 2006, 205

Human adult left calcaneum

chamber

UB-6878

3330 to 2920

4433

Site

Holm of Papa West- Orkney ray North

Davidson and HenHuman bone shall 1989, 120-2

chamber

GU-2068

3340 to 2910

4430

Rattar East Highland

Sheridan 2006, 204

Human mandible

chamber

UB-7011

3330 to 2920

4427

Point of Cott, Westray

Orkney

Barber 1997b, 59

Human bone from Skeleton C.

chamber

GU-2936

3330 to 2880

4390

Point of Cott, Westray

Orkney

Barber 1997b, 59

Human bone chamber from Skeleton I.

GU-2940

3260 to 2880

4360

Notes \ 1137

Calanais Survey and Excavation, 1979-88

Context type

Code

Date cal BC

Age BP

+/-

chamber

GrA-771

3350 to 2700

4340

Human bone chamber from Skeleton F,

UtC-1661

3090 to 2700

4300

Mean of 3 ages

Site

Council

Reference

Description

Embo

Highland

Henshall Mandible of and Ritchie human infant 1995

Point of Cott, Westray

Orkney

Barber 1997b, 59

Quanterness

Orkney

Renfrew et al 1976; 1979

Human leg bones from a single skeleton Pta-1626 chamber 4300+/- 110, Q-1479 4170 +/-110, SRR754 4360+/-110;

Quoyness, Sanday

Orkney

Renfrew et al 1976; 1979

A single human chamber bone

SRR-753

3350 to 2500

4270

110

Point of Cott, Westray

Orkney

Barber 1997b, 59

Human bone chamber from Skeleton A

GU-2934

3100 to 2500

4250

Quoyness, Sanday

Orkney

Renfrew et al 1976; 1979

Human collagen chamber from a tibia

SRR-752

3100 to 2450

4190

110

Quanterness

Orkney

Renfrew et al 1976; 1979

Human bone from the chamber of a chambered cairn.

Q-1451

3050 to 2200

4110

140

Embo

Highland

Sheridan 2006, 204

Human adult? chamber male right femur

UB-6876

2620 to 2470

4023

Embo

Highland

Henshall Human infant and Ritchie skull 1995

GrA-770

2900 to 2300

4010

chamber

Quanterness

Orkney

Renfrew et al 1976; 1979

Human bone from a single skeleton Pta-1606 4130 +/-60, Q1480 chamber (3905+/- 70 BP) and SRR755 (3870+/- 60 BP).

Clachaig, Arran

North Ayrshire

Sheridan 2006, 205

Unburnt human chamber cranium

Mean of 3 ages

UB-6897

4300

3970

2570 to 2300

3949

Notes \ 1138

Calanais Survey and Excavation, 1979-88

Context type

Code

Date cal BC

Age BP

+/-

chamber

GrA-772

2340 to 1890

3720

Human bone from an infant burial in the cairn.

cairn

AA-11698

3650 to 3000

4585

Proudfoot, E

Bone apatite from cremated bone

chamber or 2140 to SUERC-2723 fill 1910

3645

Angus

Proudfoot, E

Bone apatite from cremated human bone

chamber or 2030 to SUERC-2722 fill 1770

3575

Orkney

Sheridan 2005

Human skull

fill

UB-6421

4515

Site

Council

Reference

Description

Embo

Highland

Henshall Two adult huand Ritchie man vertebrae 1995

Point of Cott, Westray

Orkney

Barber 1997b

Fordhouse Barrow, House of Dun

Angus

Fordhouse Barrow, House of Dun Knowe of Rowiegar

Holm of Papa West- Orkney ray North

Davidson and HenHuman bone. shall 1989, 120-1

fill

GU-2067

3340 to 2890

4395

Achnacree- Argyll & beag Bute

Sheridan 2004 CHECK

Human cremated bones

blocking

GrA-26543

2190 to 1920BC

3660

Cuween

Sheridan 2005, 182

Human left femur

passage

UB-6422

3668

Tulach an Highland t’Sionnaich

Sheridan 2005, 183

Cremated human bone

forecourt

GrA-28611

3705

The Ord North, Lairg

Sharples 1981; Sher- Cremated huidan 2005, man bone 183

roof collapse

GrA-28614

3360

Orkney

Highland

In Table 25.13 all ages with a weakness of 20 or more have been excluded (Ashmore 2004b). Excluded ages include many of those measured during the ‘heroic age’ of radiocarbon dating, most of those from mixed bones and a few with technical problems. Two sets of multiple dates from single burials at Quanterness are each represented by an average. A new set of dates for Quanterness will soon be published (Sheridan pers. comm.). Thanks to the recent publications of Schulting and of Sheridan considerably more radiocarbon-dated evidence is available than in 2004 (Ashmore 2004, 130). But the total number of dated burials is still very small compared to the to-

2140 to 1970

tal number of chambered cairns in Scotland. That means that many styles of chambered cairn are not represented at all, or are represented by a single date. Therefore Illus 25.22 is likely to reflect large sampling biases; it bundles together what may have been distinct traditions and provides only a provisional overall picture losing all the nuances of different regional trajectories. Nevertheless, it is a start. Given the likely date of the chambered cairn at Calanais the later dates for burials are most relevant. There are a few mid to late 3rd millennium dates for what seem to be ‘normal’ inhumations in chambered tombs. Those later than 4000 BP Notes \ 1139

Calanais Survey and Excavation, 1979-88

Illus 25.14 Human bone ages from chambered cairns (Table 25.13) (c. 2580 to 2470 cal BC) come from Clachaig in Arran, Quanterness in Orkney and Embo in eastern Highland (Table 25.13). All of these examples came from chambered cairns with much earlier burials. However a chambered cairn at Balnuaran of Clava dated to after 2000 cal BC judging by charcoal on the old ground surface beneath it (see Note 24.12.8) and Jones and Thomas (2010) have made a case for the building of small chambered cairns along the Atlantic and Irish Sea coasts of Britain at some date after 2500 BC. The chambered cairn at Calanais, built after 2500 BC, is thus very late in the Scottish tradition but it was built before Clava cairns and probably before the date when deposition of ‘normal’ burials in the chambers of chambered cairns ceased elsewhere. All later radiocarbon ages associated with burials in chambered cairns come from remains found in secondary positions (apart from the earliest, two cremations from Fordhouse in Angus which came from a chamber floor and may not be secondary in the sense used here). Of the rest four were from inhumed bone and three from cremated bone (Table 25.13).

Note 24.12.9b The date of the Balnuaran of Clava cairns Within the broad tradition of building chambered cairns some areas did continue the practice longer than others. In round terms the NE chambered cairn at Balnuaran of Clava seems to have been built between 1900 and 1600 BC. A piece of hazel charcoal under the core cairn of the NE passage grave at Clava has been dated (AA-25234) to between 1920 and 1680 cal BC and another (AA25233) to between 2010 and 1740 cal BC, while a piece of charcoal from below the Ring cairn (AA-21256) to between2200 and 1740 cal BC. The dates may, like others from Balnuaran, have been older than their contexts, although the charcoal seems most likely to have come from ground clearance immediately prior to construction. These dates are supported by several others from contexts in the cairns (Bradley 2000, 115). The chambers continued in use for some purposes until 1050 to 810 cal BC, judging by a date (AA-21254) from a piece of charcoal in the chamber of the SW cairn (Bradley 2000, 115) Notes \ 1140

Calanais Survey and Excavation, 1979-88

24.13.4 Tormore, Arran If the Stage 2 enclosure was originally oval a loose analogy can be made with a house-like structure found at Tormore, Arran (Barber (ed) 1997, 7-21). But the area enclosed by the bank of the Tormore was considerably greater at about 9m by 12m and it was a complete oval, so the analogy is not close. The Tormore structure had two main phases with several episodes of activity in each. Beaker pottery was found in the first phase bank which also produced a radiocarbon age (GU-1176 3485+/- 60 corrected to +/-110) from charcoal, which calibrates to 2150 to 1500 cal BC. That is later than the preferred date for the Stage 2 enclosure at Calanais. A fine barbed and tanged arrowhead was found in the final bank and turf horizon 141 of this phase (Barber (ed) 1997, 24), recalling the arrowheads at Calanais. The second phase produced later pottery of types not found at Calanais and later radiocarbon dates (Barber (ed) 1997, 21). It must be noted, however, that the dating evidence for the first phase structure at Tormore could all have been residual; the pottery and dated charcoal came from the bank and the arrowhead came from soil which could have been in part imported. Note 24.14.5: Radiocarbon-dated burials in Scotland before 2000 BP

Illus 25.15Â Radiocarbon-dated burials in Scotland before 2000 BP Illustration 25.24 was prepared by counting the number of inhumations and cremations per uncalibrated century BP and plotting that in Ex-

cel. The Cal BC scale was added from a plot of the calibration curve because there is no simple one-dimensional correspondence between ages and calibrated dates. This technique was adopted to avoid claimed problems with summing calibrated date distributions using OxCal. It has its own problems, but they are probably insignificant compared to biases in the data set. The main source of bias is a probably disproportionate number of dates from human material associated with artefacts as opposed to burials without them. The other factor which must be born in mind is that radiocarbon-dated burials represent a miniscule proportion of the past populations of Scotland. Nevertheless it probably gives a better idea of at least the relative distribution in time of Scottish cremations and inhumations than guesswork. The graph shows that cremation and inhumation rites had been used in Scotland from at least the mid 4th millennium BC, and although dated inhumations outnumber date cremations for the next two millennia that may reflect sampling bias. Certainly many of the Caithness tombs excavated in the 19th century contained both burnt and unburnt human bone; none of this 19th century material has been scientifically dated (Davidson & Henshall 1991, 60-62). Inhumations Most early Scottish radiocarbon-dated inhumations came from chambered tombs. A major change in burial practises started slightly before 2400 BC. More and more burials were put in short cists, sturdily constructed of slabs and long enough for a crouched inhumation. A few burials in timber coffins are also known, as are some multiple burials and some cremations. Often the bodies were accompanied by Beakers. Beakers were also used for domestic purposes in Britain and ranged in size from fine wares holding a third of a litre to large pots holding 7 litres, suggesting that a wide range of uses is represented. The Beakers accompanying burials were selected from the low to medium end of this range (Gibson 2002, 91). Early Beakers As discussed by Sheridan in Chapter 18 The PotNotes \ 1141

Calanais Survey and Excavation, 1979-88

tery Assemblage 18.7.4 the first radiocarbon-dated Beaker burials in Britain probably date to the third quarter of the third millennium and may represent incomers. In Chapter 24 Discussion and Conclusions 24.12.6 and 24.12.7 I suggest that that does not preclude the possibility that there were Beakers in use a generation or so before those Beakers were deposited with burials. Beakers have been found from Portugal to Hungary and from Britain to North Africa (Desideri & Besse 2010, 157). They were probably first made in northern Iberia during the first half of the 3rd millennium BC (ibid quoting Giulaine 1998) although even in Iberia the first Beakers in some areas such as the southern Meseta are dated only to the second half of the millennium (Garrido-Pena 2007, 192). Early types, decorated in an International style, are found widely along the Atlantic coast from Portugal to Brittany from about 2600 cal BC (Cottiaux et al 2007, 154). Beakers reached the Paris Basin, for instance, perhaps c. 2600-2500 BC (Cottiaux et al 2007, 157) or only after c. 2500 BC (ibid 155). Modern research focuses on regional variation in the interaction of Beaker-users with 3rd millennium communities (Desideri & Besse 2010, 158). For instance, Desideri and Besse have shown through comparison of the non-metric characteristics of teeth from burials in various parts of the Continent that Beaker-using people from the south (the Mediterranean coast) were buried in western Switzerland. Their conclusion is that Beaker-users settled amongst local indigenous communities and made both populational and ideological changes (Desideri & Besse 2010, 171). That seems to be a generally accepted model for what happened around the middle of the 3rd millennium in other parts of southern Europe; Beaker potters were members of a sprinkling of incomers bringing intriguing novelties including metal working into local cultures and the consequence was a fusion of the practices of indigenes and settlers (see also Chapter 24: Discussion and Conclusions; Jones and Thomas 2010). Despite the increasing evidence Europe-wide for the movement of individuals (see Sheridan, Chapter 18) it is not clear exactly how far the first occurrence of Beakers in north-western Scotland

reflects continental incomers or the adoption of continental fashions by those here already. Burials with Beakers occurred in the northwest of Scotland as early as anywhere else in Britain. A Beaker burial at Sorisdale, Coll has been dated to between 2470 and 2230 cal BC (OxA-14722 387932, refining the previously measured BM1413 3884+/-46; Sheridan 2005, 183). None need imply true dates before c. 2400 cal BC. There are in fact two earlier radiocarbon dated features with Beaker sherds in them, Pit 50 at Dunragit, Dumfries & Galloway and a midden at Holm of Papa Westray, Orkney. The Dunragit pit had a complex fill and contained Grooved Ware sherds as well as Beaker ones. It produced four dates from individual pieces of charcoal, SUERC-2109 (4175+/-45) from oak and SUERC-2107 (4150+/-35), SUERC-2104 (4085+/35) and SUERC-2106 (4055+/-35) from hazel (Thomas 2004). These ages calibrate to 2890 to 2600, 2880 to 2600, 2870 to 2490 and 2860 to 2460 Cal BC. A report, with details of the Beaker sherds, is in preparation. A limpet midden (context IV.1) and a pocket of limpet shells at Holm of Papa Westray North produced three sherds from a coarse Beaker with impressed cardial decoration (Henshall in Ritchie 2009, 36). Context IV.1 also produced two dates, each from a red deer astralagus, OxA-17782 (4294+/-37) and OxA-17781 (4075+/-30) (Ashmore in Ritchie 2009, 60-61). These ages calibrate to 2870 to 2570 and 2860 to 2490 cal BC. Nearby structures contained Grooved Ware sherds, providing a probably coincidental analogy with the occurrence of both pottery types in Pit 50 at Dunragit. The latest date from Pit 50 at Dunragit is significantly different from the others demonstrating residuality. Even if only the latest date is taken into account it would be surprising (given current dating models) if the Beaker sherds from the pit at Dunragit did date to before 2490 cal BC and it is natural to suppose that the bit of charcoal which provided the latest date was also residual. The two dates from Holm of Papa Westray North were also significantly different from one another suggesting that even the later one, also falling before 2490 cal BC, was from bone deposited before the Beaker sherds. Neither site, therefore, provides incontroNotes \ 1142

Calanais Survey and Excavation, 1979-88

vertible evidence for Beakers in Scotland before c. 2500 cal BC. Nevertheless, given the Continental evidence for early Beakers on the Atlantic coast it is possible that Beaker-using people reached Scotland several generations before the practise of grave and then cist inhumation became prevalent. That provides a background for the early Beaker sherd in the cairn at Calanais, as well as making it conceivable (although it must again be emphasised inadequately supported by current evidence and highly speculative) that the Holm of Papa Westray North Beaker and possibly, pending its publication, even the Dunragit Beaker, both dated to before 2500 cal BC. Cremations Although most dated cremations belong after about 2200 cal BC (Illus 25.23) some were much earlier. A cremation at Midtown of Pitglassie in Aberdeenshire has been dated to between 3940 and 3660 cal BC (GrA-34772 4995+/-35; Sheridan 2007, 220). Almost as early is another in a mortuary enclosure at Pencraig Hill, East Lothian where a piece of cremated human bone was dated to between 3920 and 3630 cal BC (SUERC-7910 4940+/-50 BP; Lelong & MacGregor 2008, 41). This date was supported by several similar dates from charcoal. The third place with an early cremation is Moleigh (also known as Cleigh) in Argyll where a cremated bone from a large cist has been dated three times (GrA-24858 4945+/-40, GrA-26158 4920+/-40 and GrA-28741 4855+/-45 Sheridan 2004; Sheridan 2005, 182-3 Note 1). These three ages form a satisfactory group (agreement 108%) and if there are no technical reasons for not doing so can be combined to suggest a date range at 2 sigma of 3715 to 3640 cal BC. Many radiocarbon-dated 2nd millennium BC cremations were from urns, often highly decorated. No such urns have been identified at Calanais which suggests that it had few far-flung connections after about 2000 BC. That said, the small kerb cairn at Olcote, Breasclete, c. 1.5 km away, produced many 2nd millennium BC sherds including a small bucket urn accompanying a cremation burial and discrete deposits of sherds in similar fabrics to the urn. The kerb cairn was

broadly dated to the early to mid 2nd millennium BC. Johnson has identified a widespread tradition of plain domestic and funerary urns in the Western Isles in the second millennium BC ( Johnson in Neighbour 2005, 31-3) and their absence from the main setting at Calanais is as notable as the absence of collared and other types of urn common on mainland Scotland. It hints at a diminution in its local importance. 24.14.6 Crathes pit alignment and Upper Largie double pit alignments The pit alignment at Crathes Warrenfield, Aberdeenshire has been interpreted on the basis of radiocarbon dating as originating between the 9th and 7th millennia with further use in the 6th millennium (Murray et al 2009, 16-17). One piece of oak charcoal in a pit of the alignment was dated (SUERC-4031) to the first quarter of the 4th millennium (Murray 2009). It may be that the earlier charcoal was residual; that would bring it more into line with evidence from elsewhere in Scotland. The double pit alignment at Upper Largie in the Kilmartin Glen was discovered amongst a palimpsest of timber structures including a cursus, a pit/post ring complex, a large timber ellipse and many burials. It had been partially destroyed by quarrying; the surviving part was between 11 and 13.76m wide and up to 40m long. The pits were spaced between 3.2 and 5.3m apart from one another (centre to centre). It seemed to be later than the ellipse, which was dated to between c. 1600 and 1400 cal BC (Cooke et al 2010, 174-175, 193, 195, 202). Because of the destruction prior to excavation it is unclear whether it was an avenue as defined by Burl (1993, 4). An increasingly intriguing question is how avenues relate, if at all, to the supposedly or definitely much earlier bank barrows, earthwork cursuses, post-defined cursuses and long enclosures which have been discovered over the past few decades in Scotland. Some of these are truly massive, such as the Eskdalemuir bank barrow (Brophy 2007, 165, Fig 20.6). Indeed the variety of sites to which the label â&#x20AC;&#x2DC;cursusâ&#x20AC;&#x2122; has been attached is great enough that the term may have lost some of its usefulness (Brophy 2007, 158-60). Notes \ 1143

Calanais Survey and Excavation, 1979-88

24.14.7 Cut Hill Stone Row The stone row at Cut Hill, Dartmoor was at least 215m long and included at least 9 large stones; they now lie prone, spaced between 19.5 and 34.5m from their neighbours. Excluding the smallest stone at 1.53m the average height was 2.1m; again omitting the smallest stone at 0.5m wide they were on average 0.9m wide. Dating of peat above and below them suggested that one stone fell during the second quarter of the 4th millennium BC and another one three centuries later. They were probably submerged in peat around the middle of the 3rd millennium BC (Fyfe and Greeves 2010, 55, 59, 62-67). The row was constructed on an open heath in woodland. Its stones were more widely spaced, on average, than those of other long stone rows in Cornwall and Fyfe and Greeves regarded it as a typological outlier (ibid 68). Note 24.15.3: The earliest radiocarbon-dated farming settlement in Scotland The case for early impacts by continental farmers in the west of Scotland has been outlined in Chapter 24: Discussion and Conclusions. But the earliest radiocarbon-dated large timber constructions in Scotland assignable to farmers are at Eweford West and Pencraig Hill in East Lothian (Lelong and MacGregor 2008), where the timber structures can be interpreted as free-standing or as lining long mounds. They appear to date to the first quarter of the 4th millennium. The pots associated with the Eweford structures were mainly carinated bowls (Lelong and MacGregor 2008, 25, 33). More generally these often seem to be a component of early pottery assemblages in northern Britain (Sheridan 2009, 8-11). There is just a possibility that other fine pottery styles may have been used, at least in eastern Scotland where at Dubton near Brechin Impressed Ware sherds formed part of a large assemblage including carinated bowls in pit B233/1 (MacSween in Cameron 2002, 37). Charcoal with the assemblage was radiocarbon-dated to between 3940 and 3659 cal BC (AA-39951; Cameron 2002, 25-7, 37, 68-9). But this date is at least a

few hundred years earlier than the start-date for Impressed Wares of a little before 3300 BC suggested by Gibson (2002, 78) and unless the dating is supported by other examples it is likely that the hazel charcoal used for dating was older than the Impressed Ware pottery. Large timber halls were built in several areas from Aberdeenshire to Stirling from perhaps as early as the second quarter of the 4th millennium BC (Barclay 2003, 73-80). Large timber monuments - enclosure and avenues - appeared at about the same time (Ashmore 2007, 249). The long structures at Eweford were later than the first long mounds built in north-western Europe. Amongst others creating them, from a date before the middle of the 5th millennium BC, the Funnel-neck Beaker or TRB cultural grouping arose in the lands facing eastern Britain, in northern Germany and Denmark, while the Cerny culture communities of north-west France lived in areas facing southern England (Midgeley 2004, 120-121). The long mounds were set out in fanshaped arrays forming cemeteries, a phenomenon unparalleled in Britain, and no direct connection can be suggested. It has not been demonstrated that the succeeding cultures of these areas built long mounds around 4000 BC. Sheridan (2009, 10-11) noted that the dating evidence then available suggested that settlement by farmers occurred at about the same time in the north and the south of Britain, suggesting multiple points of entry. More recent analysis of English and lowland Scottish radiocarbon dates by Whittle and his colleagues suggests settlement in the southeast of England by small groups of colonisers from the continent slightly before 4000 cal BC with subsequent interaction and fusion with indigenous peoples and an extension to southern Scotland, perhaps through maritime leapfrogging of some intervening areas by around 3800 cal BC. They also envisage the possibility of direct settlement of Aberdeenshire from the continent by that date. They suggest more tentatively that occupation of the southern parts of the Highlands and Islands may have occurred around 3700 BC (Whittle et al 2011, 19-20). It seems probable that interaction between foragers and farmers in northern Britain was highly Notes \ 1144

Calanais Survey and Excavation, 1979-88

variable and took place at various times. It may have taken a few centuries for farming lifestyles to predominate and in some areas of Scotland forager subsistence strategies seem to have remained important even longer (Bishop et al 2009, 84). That said, feedback may have been an important element in the spread of farmers, whether or not descended from foragers in the lands around Britain and Ireland (Thomas 2005). If some groups were successful, and if they maintained links with continental groups, others might follow them fairly swiftly. Estimates of the time span over which the change took place in various parts of Scotland require substantially more radiocarbon-dated evidence if they are to improve. Note 24.18.2 Rejected parallels for the Calanais enclosure in Scotland

are common in Scotland. The database C14dates 2005 plus, archived in NMRS, was scanned for radiocarbon-dated examples. No structures closely similar to the Calanais enclosure were noted. Looser analogies were fairly common but apart from those discussed in 24.16.4 the differences between them and the Calanais enclosure were too great for useful discussion. The embanked round houses at Lairg, Sutherland, for instance, were all much larger (McCullagh & Tipping (eds) 1998). Many other Scottish structures were defined by ditches rather than banks or included substantial post-holes. The Calanais enclosure does thus fit very broadly into a roughly 2nd millennium framework but little more can be said from the current radiocarbon-dated corpus.

Round domestic and funerary enclosures of around the late 3rd and the 2nd millennium BC

Notes \ 1145

Calanais Survey and Excavation, 1979-88

26. Archiving and finds disposal All of the documents from excavation and the main post-excavation documents have been offered for archiving in the National Monuments Record of Scotland. They include the original records and digital documents (including plans and photographs) created during the post-excavation

process. Several superseded versions of digital documents have been included even though they contain misinterpretations and can only be of use to those interested in the process of writing-up. Disposal of the finds to a museum awaits the publication of this report.

Archiving and finds disposal \ 1146

Calanais Survey and Excavation, 1979-88

27. Acknowledgements

Illus 26.1 Calanais, and in the foreground Calanais III, taken from the east by Mike Brooks of Historic Scotland I am very conscious of the support of my colleagues: administrators, architects, inspectors, professional and technical staff in what was then the Ancient Monuments Division of the Property Services Agency. In those days the Ancient Monuments Division was small - ridiculously so by today’s standards. Inspectors turned their hands to many tasks. General civil service archiving policies mean that many records of the background to our work on the monuments have been destroyed. I apologise to those whose help I have not recognised here. The men of the works squad, led by Mr MacPhee, not only supplied us with equipment but also moved and re-erected large stones, rebuilt the cairn and undertook many other arduous tasks. I am most grateful to them for their skill and amiability. Iain MacIvor, the then Chief Inspector, benevolently allowed me to excavate for two seasons. David Breeze, then managing excavation matters, provided much welcomed support. Even their patience had to have limits, however, and a third season was not feasible. The architects, Bill Boal and Ingval Maxwell along with Alan Armstrong managing the works team authorised crucial services and backup. Mr. B. Petersen of the Ancient Monuments Drawing Office kindly translated the note on Worsaae’s sketches from Danish.

Without the skilled supervision of individual trenches by Jean Comrie, Anne-Marie Gibson, Lily Linge, Ian Maté and Peter Strong the excavation would have been far less successful than it was. Mike Brooks, with his usual skill and care, took the better photographs included here. He also kindly let me have copies of Captain F W L Thomas’ photographs which he had recognised in the National Archives of Scotland. The civility of the people of Calanais and Breasclete lightens even further my recollection of those sunny days when we dug amongst the stones, and mitigates any memories of the rainy ones. My thanks go in particular to Mr and Mrs MacArthur of Breasclete who were most pleasant hosts during our stay there, to Mrs MacDonald of Calanais who minded our children and to the staff of Breas-

Illus 26.2 The works squad re-erecting Stone 33a watched by school pupils from Breasclete Acknowledgements \ 1147

Calanais Survey and Excavation, 1979-88

Illus 26.3 Mr MacPhee setting a fallen chamber orthostat on end [Film 81.21.13]

Illus 26.4 Jean Comrie, Anne-Marie Gibson and Ian Maté photographed by Mike Brooks

Illus 23.6 Some of the volunteers and staff in 1980 clete School. The co-operation of Carloway Estate and of Mr Angus Morrison, adjoining proprietor and crofter, was most welcome. Jim Crawford and friends organised a delicious salmon supper for the whole dig, still remembered with pleasure. I hope that the many people who worked on

Illus 26.5 John Linge, Fionna Ashmore and Lily Linge photographed by Mike Brooks [Film Brooks 1-34] the site in 1980 and 1981 both enjoyed themselves and found the experience helpful. We advertised for volunteers in the Stornoway Gazette and the Guardian. Many more applied than we could accept and I once again extend my apologies to those whom we could not fit in. I shall not reveal which of the successful applicants noticed our advert in the Guardian because it was used to wrap a fish supper. They included Mark Angliss, Alexandra Campbell-Stewart, Jane Corrie, Alan Fairweather, Catriona Graham, Dick Grove, Robert Horrocks, Richard Jones, Elaine Lawes, Mary McCann, Brian Matthews, Cathy Payne, Margaret Ponting, David Rowswell, Clive Ruggles, Peter Roberts, Sarah-Jane Sloan, Rhian Sterland, Marion Wade-Evans and Adrienne Walton. Large scale plans of Calanais were kindly made available by Dr Tait of the Department of Geography, University of Glasgow (Tait et al 1978). Their Acknowledgements \ 1148

Calanais Survey and Excavation, 1979-88

accuracy saved us much work and I am grateful to him and his colleagues. Colin Heathcote at the Bradford School of Undergraduate Studies very kindly produced plots from the 1979 resistivity survey in September 1980. Tom Borthwick produced the finds illustrations with his customary skill and accuracy. I am grateful for information and discussion to Steven Briggs, Mike Brooks, Trevor Cowie, Ron Curtis, Diana Murray, Colin Richards, J. N. Graham Ritchie and Clive Ruggles. Marion Wade Evans and Alan Fairweather in addition to working as volunteers kindly completed a vegetation survey of the area. Above all I owe a large debt to Gerald Ponting and Margaret Ponting (now Curtis), whose researches into the early sources for study of Callanish, and other work on the stones and the surrounding landscape, have accumulated a treasure trove of information. It was always a pleasure to exchange discoveries with them and they gave more information on documentary sources than they received. Margaret also worked as a volunteer in both seasons and Gerald’s photography helped to soften the loss of two failed rolls of film negatives. The post-excavation process was long drawn out. That was no fault of the specialists, including Rod McCullagh, Sjoerd Bohncke, Audrey Henshall, Melanie Johnson, Alison Sheridan, Ian Maté, Torben Ballin and Caroline Wickham-Jones, to all of whom I am most grateful. Alison Sheridan in particular worked all hours to produce an exhaustive report on the pottery which significantly changed my conclusions. Without the post-excavation labours of the site supervisors, of Ann Clarke, Melissa Seddon and more recently, Emma Carver and Andrew Heald publication would have been more difficult.

The first set of radiocarbon samples was identified by Rod McCullagh, prepared by Philip Naismith of SURRC and sent by Gordon Cook for dating at the Arizona AMS laboratory. The second set of samples was identified by Robin Inglis, Ann Crone and colleagues at AOC Archaeology; these samples too were prepared by Philip, and Gordon arranged their measurement on the SUERC AMS laboratory. I am grateful to all for their skill and care. The site drawings were turned into vector drawings by Headland Archaeology, with management of the project by Mike Middleton and main editing by Mike Kimber. Some of the photographs were digitised by Historic Scotland photographers David Henry and Chris Hutchison; most were digitised commercially. In 1997 and 2000 Professor David Breeze, by then the Chief Inspector, kindly allowed me time to complete a draft of the accounts of the excavation areas. Peter Yeoman and Richard Strachan managed completion of the writing up project for Historic Scotland after my retirement from Historic Scotland in April 2006. Several people helped during the latter parts of the post-excavation process. Camilla Toulmin introduced me to Ester Boserup’s The conditions of agricultural growth, without which my appreciation of early agriculture would have been even less informed than it is. Alison Sheridan and Alan Saville kindly provided useful off-prints. Research into the Ordnance Survey Name Books benefitted from the welcome help of staff of the National Monuments Record of Scotland. Lastly, (those who know her will say ‘of course’) Fionna Ashmore provided unflagging support throughout this project, along with a vigorous emphasis on completion. I cannot thank her enough.

Acknowledgements \ 1149

Calanais Survey and Excavation, 1979-88

28. Bibliography Alexander, D 1997 Excavation of pits containing decorated Neolithic pottery and early lithic material of possible Mesolithic date at Spurryhillock, Stonehaven, Aberdeenshire. Proc Soc Ant Scot 127, 17–27. An Lanntair 1995 Calanais. Stornoway. Angus, S 1987‘Lost woodlands of the Western Isles’, Hebridean Naturalist1987, 24-30. Ananthaswamy, A 2013 ‘Quantum shadows’ New Scientist No 2898 5 January 2013, 36-9 Armit, I 2003 ‘The Drowners: permanence and transience in the Hebridean Neolithic’, in Armit et al (eds) 2003, 93–100. Armit, I forthcoming. The Excavation of a Neolithic Settlement on Eilean Domhnuill, Loch Olabhat, North Uist. Armit, I, Murphy, E, Nelis, E and Simpson, D (eds) 2003 Neolithic Settlement in Ireland and Western Britain. Oxford Ashmore, P J 1983 ‘Callanish: Making a stone stand again’, in Magnusson M 1983Echoes in Stone. Edinburgh Ashmore, P J 1984 ‘Callanish’, in Breeze D J 1984,Studies in Scottish Antiquity Edinburgh, 1-31. Edinburgh Ashmore, P J 1989. ‘Excavation of a Beaker cist at Dornoch Nursery’, Sutherland’,Proc Soc Antiq Scot 119, 63-72. Ashmore, P J 1995 Calanais: The Standing Stones’. Calanais Ashmore, P J 1996 Neolithic and Bronze Age Scotland. London. Ashmore, P J 1997 ‘Calanais, Western Isles’ D&ES 1997, 116. Ashmore, P J 1997 ‘Radiocarbon Dates from Archaeological Sites in Argyll and Arran’,in J. N. G. Ritchie (ed.) The Archaeology of Argyll, 236-283. Edinburgh. Ashmore, P J 1999a ‘Radiocarbon dating: avoiding errors in dating by avoiding mixed samples’,Antiquity 73 No 279 1999, 124-127. Ashmore, P J 1999b ‘Dating the ring of stones and chambered cairn at Calanais’,Antiquity 73 No 279 March 1999, 128-130. Ashmore, P. J. 1999c ‘Archaeology and Astronomy: a

view from Scotland’ Archaeoastronomy: The Journal of Astronomy in Culture Vol. XIV Number 2 1999, 3-32 Ashmore, P J 2000 ‘Single Entity Dating’,Actes du colloque <<C14 Archéologie>>, 65-71, Lyons Ashmore, P J 2003 ‘Terminology, time and space: labels, radiocarbon chronologies and a ‘Neolithic’ of small worlds’, in Armit et al (eds) 2003, 40-46. Ashmore, P J 2004a ‘Absolute Chronology’, in Shepherd, I. A. G. and Barclay, G. J. (eds) 2004. Scotland in Ancient Europe: The Neolithic and early Bronze Age of Scotland in its European Context. Edinburgh: Society of Antiquaries of Scotland. Ashmore, P J, 2004b. ‘Dating forager communities in Scotland’, in Saville, A. (ed) 2004. Mesolithic Scotland and its Neighbours. 83-94. Society of Antiquaries of Scotland. Edinburgh. Ashmore, P J 2004c ‘A date list (to October 2002) for early foragers in Scotland’, in Saville, A. (ed) Mesolithic Scotland and its Neighbours., 95-158 Society of Antiquaries of Scotland. Edinburgh. Ashmore, P J 2004c ‘Time, Space and the Standing Stones round Calanais’, in Gibson, A. & Sheridan, A. (eds) From Sickles to Circles: Britain and Ireland at the Time of Stonehenge, 64-77. Stroud Ashmore, P J 2007 ‘Radiocarbon dates from the Pict’s Knowe, Holywood and Holm’, in Thomas J (ed.) Place and Memory: Excavations at from the Pict’s Knowe, Holywood and Holm Farm, Dumfries and Galloway, 1994-8, 248-259. Oxford. Ashmore, P J and Wickham-Jones C 2010 ‘Radiocarbon Determinations in context’ in Hardy K and Wickham-Jones C (eds) 2010, Section 4 Ashmore, P J, Cook G T & Harkness D D 2000 ‘A Radiocarbon Database for Scottish Archaeological Samples’ Radiocarbon 42, 1 (2000), 41-48.Atkinson, J A 2002 ‘Excavations at a Mesolithic and Neolithic occupation site at Chapelfield, Cowie, Stirling. Proc Soc Ant Scot 132, 2002, 139-192. Atkinson, R J C 1960 Stonehenge, London. Bailley, M 1999 Exodus to Arthur Batsford. London. Ballantyne, C K 2004 ‘After the Ice: Paraglacial and Postglacial Evolution of the Physical Environment of Scotland, 20,000 to 5000 BP’ in Savile A (ed) 2004, 27-44 Bibliography \ 1150

Calanais Survey and Excavation, 1979-88

Ballin, T B 1999 ‘Bipolar Cores in Southern Norway Classification, Chronology and Geography’ Lithics 20, 1999, 13-22. Ballin, T B 2004 ‘The Mesolithic Period in Southern Norway: Material Culture and Chronology’ In Saville, A (ed.) 2004: “Mesolithic Scotland and its Neighbours: The Early Holocene Prehistory of Scotland, its British and Irish Context, and some Northern European Perspectives”, 413-438.Society of Antiquaries of Scotland, Edinburgh. Ballin, T B 2004 The worked quartz vein at Cnoc Dubh, Isle of Lewis, Western Isles: Presentation and discussion of a small prehistoric quarry. Scottish Archaeological Internet Reports 11 http://www. sair.org.uk/sair11/index.html. Ballin, T B forthcoming a ‘The Lithic Assemblage’. In Shepherd, I & A.: Rosinish, Benbecula, Outer Hebrides. Ballin, T B & Saville, A 2003 ‘An Ahrensburgian-type tanged point from Shieldaig, Wester Ross, Scotland and its implications’ Oxford Journal of Archaeology Vol 22 Issue 2, 115-131. Ballin, T B, Saville, A, Tipping R & Ward T 2010 ‘An Upper Palaeolithic flint and chert assemblage from Howburn Farm, South Lanrkshire, Scotlanf: First results.’ Oxford Journal of Archaeology 29(4) 323-360. Ballin, T B forthcoming ‘The Quartz Assemblage’. In Wilson, G. & Moore, H.: Bayanne, Yell, Shetland. Ballin-Smith, B and Petersen, G 2003 ‘Brodgar Farm’ D&ES New Series 4 2003, 102-3 Barber, J (ed) 1997, The Archaeological Investigation of a Prehistoric Landscape: Excavations on Arran 1978-1981. The Scottish Trust for Archaeological Research. Edinburgh. Barber, J 1997. The excavation of a Stalled Cairn at the Point of Cott, Westray, Orkney. The Scottish Trust for Archaeological Research. Edinburgh. Barclay, G J & Russell-White, C J 1993 ‘Excavations in the ceremonial complex of the fourth to second millennium BC at Balfarg/Balbirnie’, Proc. Soc. Antiq . Scot. 123 (1993), 43-210 Barclay, G J 1992 ‘The Scottish gravels: a neglected resource’ in Fulford M and Nichols E 1992 Developing Landscapes of Lowland Britain: The Archaeology of the British Gravels: A Review, 106-124. Society of Antiquaries of London Occasional Paper 14. London.

Barclay, G J 1996 ‘Neolithic Buildings in Scotland’ in Darvill T and Thomas J (eds) 1996 Neolithic Houses in North-west Europe and Beyond, 61-75. Oxbow Monograph 57. Oxford. Barclay, G J 1997 ‘The Neolithic’ in Edwards and Ralston (eds) 1992 Barclay, G J 2000 ‘Between Orkney and Wessex: the search for the regional Neolithics of Britain’ in Ritchie A (ed) 2000, 275-85 Barclay, G J 2001 ‘Neolithic enclosures in Scotland’ in Darvill T and Thomas J 2001 (eds) Neolithic Enclosures in Atlantic Northwest Europe, 144-154 Barclay, G J 2003 ‘Neolithic settlement in the lowlands of Scotland: a preliminary survey’ in Armit et al 2003 (eds) Neolithic Settlement in Ireland and Western Britain, 7-83. Oxbow Books. Oxford. Barclay, G J 2004 ‘”Four Nations Prehistory”: cores and archetypes in the writing of prehistory’, in Philips, R & Brocklehurst, H (eds) History, nationhood and the Question of Britain. Palgrave. Barclay, G J 2005 ‘The ‘henge’ and ‘hengiform’ in Scotland’ in Cummings V and Pannett A 2005 (eds) Set in Stone: New approaches to Neolithic Monuments in Scotland, 81-94. Oxbow. Oxford. Barclay, G J and Brophy K 2002 ‘Carsie Mains, Perth and Kinross (Blairgowrie parish), rectilinear timber structure; timber ring’, Discovery and Excavation, Scotland, 3, 2002, 91, Barry, J 1808, History of the Orkney Isles, London. Barton, R N E 1999 ‘The Lateglacial or Late and Final Upper Palaeolithic colonisation of Britain’ in Hunter, J & Ralston, I 1999 The Archaeology of Britain, 13-34. Routledge. London and New York Bayliss A, Bronk Ramsey C, van der Plicht J and Whittle A 2007 ‘Bradshaw and Bayes: Towards a Timetable for the Neolithic’ Cambridge Archaeological Journal 17 No 1 (Supplement) February 2007, 1-28 Beek, R van 2011 ‘Diversity Rules. On Late Prehistoric Settlement of the Eastern Netherlands and the Need for Regionally Specific Models’ Proc. Prehist. Soc 77, 2011, 25-47 Bennett, K D, Fossit, J A, Sharp M J & Switsur, V R 1990 ‘Holocene vegetational and environmental history at Loch Lang, South Uist, Western Isles, Scotland’ in New Phytol, 114 (1990), 281-98 Bennett, K D, Boreham, S, Sharp, M J & Switsur, V R 1992 ‘Holocene history of environment, vegetation, and human settlement in Catta Ness, Lunnasting, Bibliography \ 1151

Calanais Survey and Excavation, 1979-88

Shetland’, J Ecol 80: 241-273. Beveridge, E 1911 North Uist, its Archaeology and Topography. Edinburgh. William Brown. Birks, H J B 1989 ‘Holocene isochrone maps and patterns of tree-spreading in the British Isles’, J Biogeogr 16: 503-540. Birks, H J B & Madsen, B J 1979‘Flandrian vegetational history of Little Loch Roag, Isle of Lewis, Scotland’, J Ecol 67, 825-42. Bishop, R R, Church, M J and Rowly-Conwy, P A 2009 ‘Cereals, fruits and nuts in the Scottish Neolithic’ Proc Soc Ant Scot 139, 47-103 Bishop, R R, Church, M J and Rowly-Conwy, P A 2010 ‘Northton, Harris’ Discovery and Excavation in Scotland New Series 11, 2010, 178 Bishop, R R, Church, M J and Rowly-Conwy, P A 2011 ‘Northton, Harris’ Discovery and Excavation in Scotland New Series 12, 2011, 185 Boardman, S 1995 ‘Charcoal and charred plant macrofossils’, in Branigan, K & Foster, P Barra. Archaeological research on Ben Tangaval, 149-157. Sheffield: Sheffield Academic Press. Bohncke, S J P 1988 ‘Vegetation and habitation history of the Callanish area, Isle of Lewis, Scotland’ in Birks H H, Birks H J, Kaland P E and Moe D (eds) The Cultural Landscape – Past Present and Future, 445-461 Cambridge University Press, Cambridge. Bonsall, C, Tolan-Smith, C & Saville, A 1995 ‘Direct dating of Mesolithic antler and bone artifacts from Great Britain: new results for bevelled tools and red deer antler mattocks’, Mesolithic Miscellany 16(1), 2–10. Boserup, E 1993 The conditions of agricultural growth: The economics of agrarian change under population pressure. Earthscan Publications Ltd. London. Bowen, E G 1972 Britain and the Western Seaways. Thames and Hudson. Boyd, J M (ed) 1979 The Natural Environment of the Outer Hebrides. Proc Roy Soc Edinb 77B. Bradley R & Lamdin-Whymark 2009 ‘Pullyhour - a signpost to the past’ Past 61 April 2009, 3-5 Bradley, R 1988 ‘Stone circles and Passage Graves - a Contested Relationship’ in Gibson and Simpson 1988 2-13 Bradley, R 1993 Altering the Earth: The Origins of Monuments in Britain and Continental Europe Society of Antiquaries of Scotland Monograph Series Number 8. Edinburgh.

Bradley, R 1998‘Excavations at Clava’ Current Archaeology 148, 136-142 Bradley, R 2000 The Good Stones: a new investigation of the Clava Cairns. Society of Antiquaries of Scotland Monograph Series Number 17. Edinburgh. Bradley, R 2005 The moon and the bonfire. Society of Antiquaries of Scotland. Edinburgh Bradley, R 2006 ‘Northern Scottish henge project’ Discovery and Excavation in Scotland New Series 7, 2006, 19-20 Bradley, R 2007 ‘Northern Scottish henge project’ Discovery and Excavation in Scotland New Series 8, 2007, 26 Bradley, R 2010 ‘Passage graves, statues and standing stones: megaliths and social identities in prehistoric Scotland and Ireland’ European Megalithic Studies Group www.jungsteinSITE.de Bradley, R, Harding J and Mathews M 1993 ‘The Siting of Prehistoric Rock Art in Galloway, South-west Scotland’ Proceedings of the Prehistoric Society 59, 1993, 269-283 Brahic C 2012 ‘Our true dawn’ New Scientist 2892 24 November 2012, 34-37 Brayshay, B A and Edwards, K J 1996 ‘Late-glacial and Holocene vegetational history of South Uist and Barra’, in Gilbertson, D D, Kent, M E and Grattan, J (eds) The Outer Hebrides. The Last 14,000 Years, 13-26. Sheffield: Sheffield Academic Press. Brayshay, B A et al 2000 ‘Surface pollen-vegetation relationships on the Atlantic seaboard: South Uist, Scotland’, J Biogeogr 27: 359-378. Brennan, J, Briggs, C S & ApSimon, A M 1978`A giant Beaker from Cluntyganny Townland, County Tyrone’ Ulster J Archaeol 41, 33–6. Bronk Ramsey, C 1995 Radiocarbon Calibration and Analysis of Stratigraphy: The OxCal Program Radiocarbon 37(2), 425-430. Bronk Ramsey, C 2005 Oxcal 3.10 Oxford Radiocarbon Accelerator Unit, Research Lab for Archaeology. Oxford. Brophy, K & Barclay, G J in press ‘A rectilinear timber structure and post-ring at Carsie Mains, Meikleour, Perthshire’, TAFAJ Brophy, K 2000 ‘Cursus Monuments and Rivers’ in Ritchie A (ed) Neolithic Orkney in its European context, 59-70. McDonald. Cambridge Brophy, K 2006 ‘Rethinking Scotland’s Neolithic: combining circumstance with context’ Proc. Soc. Antiq. Bibliography \ 1152

Calanais Survey and Excavation, 1979-88

Scot.136, 2006, 7-46 Brophy, K 2007 ‘The cursus monuments of south-west Scotland’ in Thomas J (ed) 2007 Place and Memory, 158-67. Brophy, K & Noble, G 2011 ‘From Beakers to daggers: Bronze Age burial practice at Forteviot’ Past 67 April 2011, 1-3 Brown, A 2007 ‘Dating the onset of cereal cultivation in Britain and Ireland: the evidence from charred cereal grains’, Antiquity 81: 1042-1052. Bruck, J 1995 ‘A place for the dead: the role of human remains in Late Bronze Age Britain’ Proceedings of the Prehistoric Society 61, 1995, 245-277 Buchanan M 2002 Small World: uncovering nature’s hidden neyworks. Weidenfeld & Nicolson. London. Buck, C E, Cavanagh W G & Litton C D 1996Bayesian Approach to Interpreting Archaeological Data. John Wiley and Sons. Chichester. Bunting B T 1969 The Geography of Soil. Hutchison University Library. London. Burl, A 1993 From Carnac to Callanish: The Prehistoric Stone Rows and Avenues of Britain, Ireland and Brittany Yale University Press. New Haven and London. Burl, A 2000 The Stone Circles of Britain, Ireland and Brittany Yale University Press. New Haven and London. Burl, A. 1976, The Stone Circles of the British Isles, London. Callaghan, R & Scarre, C 2009 ‘Simulating the Western Seaways’ Oxford Journal of Archaeology 28, 357-72 Callender, C. 2010 ‘There is no theory of everything’ New Scientist 11 September 2010, 50. Callender, H 1857 ‘Notice of the Stone Circle at Callanish’, Proc. Soc. Antiq. Scot. 2 (1857) 380-4. Cameron, K 2002 ‘The excavation of neolithic pits and Iron Age souterrains at Dubton farm, Brechin, Angus’ TAFAJ 8 2002, 19-76 Campbell, A 2010 ‘Stacaiscal’ Discovery and Excavation in Scotland New Series 11, 2010, 175 Card, N 2007 ‘Ness of Brodgar’ D&ES New Series 8 2007, 147 Card, N 2010 ‘Colour, cups and tiles - recent discoveries at the Ness of Brodgar’ Past 66, November 2010, 1-3 Carte, T 1747 The History of England, London. Case, H J 2004 Bell Beaker and Corded Ware Culture burial associations: a bottom-up rather than top-down approach. In A. M. Gibson and J. A.

Sheridan (eds.), From Sickles to Circles: Britain and Ireland at the Time of Stonehenge. Stroud: Tempus. 201–14. Catling, C 2009 ‘Bluestonehenge’ Current Archaeology December 2009 237, 23-4 Caulfield, S, O’Donnell, R G and Mitchell, P I 1998 ‘14C dating of a Neolithic Field system at Céide Fields, County Mayo, Ireland’ Radiocarbon 40-2 1998, 629-640 Champion, T 1999 ‘The later Bronze Age’ in Hunter J and Ralston I 1999 The Archaeology of Britain, 95-112. Routledge. London and New York. Clarke, D L 1970 Beaker Pottery of Great Britain andIrelandCambridge University Press. Cambridge Clarke, D V 2004 ‘The construction of narratives for Neolithic Scotland’ in Shepherd and Barclay (eds) 2004. Soc Ants Scot. Edinburgh. Clarke, D.L. 1970 Beaker Pottery of Great Britain and Ireland. Cambridge: Cambridge University Press. Close-Brooks, J 1995 Excavation of a cairn at Cnip, Uig, Isle of Lewis. Proceedings of the Society of Antiquaries of Scotland 125, 253–77. Coles, G M 1993 ‘Cnoc an Tursa’ Discovery and Excavation in Scotland 1993, 110-111 Coles, G M and Rees, T 1994 ‘Cnoc an Tursa’ Discovery and Excavation in Scotland 1994, 96 Cook, M, Ellis, C & Sheridan, A2010 ‘Excavations at Upper Largie Quarry, Argyll and Bute, Scotland: New light on the prehistoric ritual landscape of the Kilmartin Glen’Proc. Prehist Soc 76, 165-212 Cooke, J A, Few, R W, Morgan, I G and Ruggles, C L N 1977 ‘Indicated Declinations at the Callanish Megalithic Sites’, J. Hist. Astron., 8 (1977) 113—33. Cooney, G 2003 ‘Rooted or routed?’ in Armit et al 2003, 47-55 Cowie, R 2005 The pottery. In A. Ritchie, Kilellan Farm, Ardnave, Islay: Excavations of a Prehistoric to Early Medieval Site by Colin Burgess and Others 1954–1976, 49–96. Edinburgh: Society of Antiquaries of Scotland. Cox, R A V 1989 ‘Ainmean-Aite: Cruth, Fas is Comharrachadh’ in MacLeod F 1989, Togail Tir: Marking Time 123-33 Crawford, I A 1977 A corbelled Bronze Age burial chamber and Beaker evidence from the Rosinish machair, Benbecula. Proc Soc Antiq Scot 108 (1976–7), 94–107. Crawford, I A 1980 Excavation and Research at CoileBibliography \ 1153

Calanais Survey and Excavation, 1979-88

agan an Udail, N Uist 1980. Unpublished interim report. Crone, B A 1993 ‘Excavation and survey of sub-peat features of Neolithic and Bronze Age date at Bharpa Carinish, North Uist, Scotland’, Proc Prehist Soc 59, 361-382. Crone, B A 1999 ‘Analysis of the carbonised and waterlogged wood’, in Grinter, P & Mills, C M Eilean Domhnuill, Loch Olabhat, North Uist: Neolithic subsistence strategy and environment. Unpubl report for Historic Scotland. Cummings, V 2002 ‘Between Mountains and Sea: a Reconsideration of the Neolithic Monuments of South-west Scotland’ Proc Prehist Soc 68, 125-146 Curtis, G R 1980 ‘Some Geometry associated with the Standing Stones of Callanish’ Hebridean Naturalist 1980, 29-40 Curtis, M R & Curtis, G R 2000, see Ponting & Ponting 2000 The stones around Callanish. Curtis M R & G R. Calanais. Curtis, M R & Curtis, G R 2003 ‘Cnoc Gearraidh Nighean Choinnich, Breasclete’ Discovery and Excavation in Scotland New Series 4, 138 CSA. Edinburgh. Curtis, M R & Curtis, G R 2005 ‘Cnoc Gearraidh Nighean Choinnich, Breasclete’ Discovery and Excavation in Scotland New Series 6, 149 CSA. Edinburgh. Curtis, G R & Curtis M R 2008 Callanish: stones, moon & sacred landscape: 2008Windfarm Submission Curtis, G R & Curtis M R 2009 Callanish: stones, moon & sacred landscape: 2009 extrapolation Curtis M R & G R. Calanais. Dalland, M 1993 The calibration of the radiocarbon dates. In G.J. Barclay and C.J. Russell-White, Excavations in the ceremonial complex of the fourth to second millennium BC at Balfarg/Balbirnie, Glenrothes, Fife, 161. Proceedings of the Society of Antiquaries of Scotland 123, 43–210. Darvill, T 1997 ‘Neolithic landscapes: Identity and definition’ in Topping P (ed) 1997 Neolithic Landscapes, 1-14. Neolithic Studies Group Seminar Papers 2. Oxbow. Oxford. Darvill, T & Thomas, J 2001 (eds) Neolithic Enclosures in Atlantic Northwest Europe. Neolithic Studies Group Seminar Papers 6. Oxbow. Oxford. Davidson, J & Henshall A S 1991 The Chambered Cairns of Caithness Edinburgh University Press,

Edinburgh Desideri, J and Besse, M 2010 ‘Swiss Bell Beaker population dynamics: eastern or southern influences’ Archaeological & Anthropological Sciences vol 2 number 3, 157-173 Dickson, C & Stenhouse, M J 1981 Charcoal identification of radiocarbon samples and the results of radiocarbon assays undertaken upon these samples. In R.J. Mercer, The excavation of a Late Neolithic henge-type enclosure at Balfarg, Markinch, Fife, 144–145. Proceedings of the Society of Antiquaries of Scotland 111, 63–171. Downes, J & Badcock, A 1999 Berneray Causeway; archaeological watching brief and excavations at the Screvan Quarry site and Otternish, North Uist. Unpublished report (ARCUS report 231c). Sheffield: ARCUS. Downes, J & Richards, C 2005 ‘The Dwellings at Barnhouse’ in Richards C (ed) 2005, 57-128. Duns, J 1883. Notes on North Mull. Proceedings of the Society of Antiquaries of Scotland 17 (1882–3), 79–89. Dymond, C W 1877 ‘The Megalithic Antiquities at Stanton Drew’, J. Brit. Archaeol. Assoc., 33 (1877) 297-307. Edmonds, M & Richards, C (eds) 1998 Understanding the Neolithic of North-western Europe. Cruithne Press. Glasgow Edwards, K J 1990 ‘Fire and the Scottish Mesolithic: evidence from microscopic charcoal’ in Vermeesch, P & van Peer, P (eds) Contributions to the Mesolithic in Europe, Leuven, 71-9 Edwards, K J 1996 ‘A Mesolithic of the Western and Northern Isles of Scotland?’, in Pollard, T and Morrison, A (eds) The Early Prehistory of Scotland, 23-38. Edinburgh: Edinburgh University Press. Edwards, K J 2000 ‘Pollen, archaeology, and burdens of proof ’, in Young, R (ed) Mesolithic Lifeways, 67-74. Leicester: Leicester Archaeological Monographs 7. Edwards, K J & Mithen, S 1995 ‘The colonisation of the Hebridean Islands of Western Scotland: evidence from the palynological and archaeological records.’ World Archaeology Vol 26 No 3 Edwards, K J, Mulder, Y, Lomax, T A, Whittington, G and Hirons, K R 2000 ‘Human-environment interactions in prehistoric landscapes: the example of the Outer Hebrides’, in Hooke, D (ed) Landscape, the Richest Historical Record, 13-32. Birmingham: SoBibliography \ 1154

Calanais Survey and Excavation, 1979-88

ciety for Landscape Studies Supplementary Series 1. Edwards, K J & Ralston, I B M (eds) 1997 Scotland: Environment and Archaeology 8000 BC-AD 1000. John Wiley & Sons. Chichester. Edwards, K J, Whittington, G & Hirons, K 1995 ‘The relationship between fire and long-term wet heath development in South Uist, Outer Hebrides, Scotland’, in Thompson, D, Hester, A & Usher, M B (eds), Heaths and Moorland: Cultural landscapes, 240-8. Edinburgh: SNH. Ellice, H 1860 ‘Donations: Some pieces of Charcoal dug up in a Stone Circle near Callernish in the Lewis’ Proc Soc Antiq Scot 3 1857-60, 202 Ellis, C 2000 ‘Upper Largie’ Discovery and Excavation in Scotland New Series 1, 2000, 16 Ellis, C 2002 ‘Upper Largie’ Discovery and Excavation in Scotland New Series 3, 2002, 145 Eogan, G 1984 Excavations at Knowth. Volume 1. Dublin: Royal Irish Academy. Eogan, G. and Roche, H. 1999 Grooved Ware from Brugh na Bóinne and its wider context. In R. Cleal and A. MacSween (eds), Grooved Ware in Britain and Ireland, 98–111. Oxford: Oxbow (Neolithic Studies Group Seminar Papers 3). Evans, J, Chenery, C and Fitzpatrick, A P 2006 Bronze Age childhood migration of individuals near Stonehenge, revealed by strontium and oxygen isotope tooth enamel analysis, Archaeometry 48, 309–22. Ferguson, W 1998 The Identity of the Scottish Nation. Edinburgh. Edinburgh University Press Fettes, D J, Mendum, J R, Smith, D I & Watson, J V 1992 Geology of the Outer Hebrides British Geological Survey HMSO. London. Finlayson, B 2000 ‘Chipped Stone Assemblage’ In Downes, J, & Lamb, R 2000 Prehistoric Houses at Sumburgh in Shetland. Excavations at Sumburgh Airport 1967-74. Oxford. Fitzpatrick, A P 2009 In his hands and in his head: the Amesbury Archer as a metalworker. In P. Clarke (ed), Bronze Age Connections: Cultural Contact in Prehistoric Europe, 176–88. Oxford: Oxbow. Fitzpatrick, A P 2011 . The Amesbury Archer and the Boscombe Bowmen. Early Beaker burials at Boscombe Down, Amesbury, Wiltshire, Great Britain. Salisbury: Wessex Archaeology Report 27. Flitcroft, C, Johnson, M & Coles, G 2000 ‘Calanais Farm’ Discovery and Excavation in Scotland New Series 1, 102-3

Fossitt, J A 1996 ‘Late Quaternary vegetation history of the Western Isles of Scotland’, New Phytol. 132: 171-196. Freeman, P R 1976, ‘A Bayesian Analysis of the Megalithic Yard’, J. Roy. Stat. Soc., 139, Pt. 1 (1976) 20-55. Fuller, S 2003 Kuhn vs Popper. Icon Books Ltd. Cambridge. Fyfe, R M & Greaves, T ‘The date and context of a stone row: Cut Hill, Dartmoor, south-west England’ Antiquity 84 No 323, March 2010, 55-70 Garriba-Pena, R 1997 ‘Bell Beakers in the Southern Meseta of the Iberian Peninsula: socioeconomic context and new data’ Oxford Journal of Archaeology 16(2) 1997, 194-209 Garrow, D & Sturt, F 2011’Grey waters bright with Neolithic argonauts? Maritime connections and the Mesolithic-Neolithic transition within the ‘western seaways’ of Britain, c. 5000-3500 BC’ Antiquity 85 No. 327, 59-72 Garwood, P. 1999 Grooved Ware in southern Britain: chronology and interpretation. In R. Cleal and A. MacSween (eds), Grooved Ware in Britain and Ireland, 145–76. Oxford: Oxbow (Neolithic Studies Group Seminar Papers 3). Gibson, A M 1982Beaker Domestic sites. A Study of the Domestic Pottery of the Late Third and Early Second Millennia B.C. in the British Isles. Oxford: British Archaeological Reports, British Series 107. Gibson, A M 1995 The Neolithic pottery from Allt Chrisal. In K. Branigan & P. Foster, Barra. Archaeological Research on Ben Tangaval, 100–115. Sheffield: Academic Press Gibson, A M 1998 Stonehenge and Timber Circles. Tempus, London Gibson A M 2002 Prehistoric pottery in Britain and Ireland Tempus. Stroud Gibson, A.M 2006. Pottery. In D.D.A. Simpson, E.M. Murphy and R.A. Gregory, Excavations at Northton, Isle of Harris, 90–133. Oxford: British Archaeological Reports (British Series 408).Gibson A & Simpson D 1988 Prehistoric Ritual and Religion Sutton Publishing Ltd. Stroud Gibson, A M 2010 ‘ Dating Balbirnie: recent radiocarbon dates from the stone circle and cairn at balbirnie, Fife, and a review of its place in the overall Balfarg/Balbirnie site sequence’ Proc. Prehist. Soc 140 (2010), 51-77 Gilbertson, D D, Kent, M & Grattan, J 1996 The Outer Bibliography \ 1155

Calanais Survey and Excavation, 1979-88

Hebrides: The last 14,000 Years Sheffield University Press, Sheffield Gilbertson, D D, Schwenninger, J-L, Kemp, R A & Rhodes, E J 1999 ‘Sand-drift and Soil Formation along an Exposed North Atlantic Coastline: 14,000 Years of Diverse Geomorphological, Climatic and Human Impacts’ Journal of Archaeological Science (1999) 26, 439-469 Guilaine, J 1998 Impressions de Riva del Garda: Bell Beakers today: pottery, people, culture, symbols in 3rd millennium Europe. Bull. Soc. Préhist. Fr. 95, 418–419 Green, H S 1980 The Flint Arrowheads of the British Isles. A detailed study of material from England and Wales with comparanda from Scotland and Ireland. Part i-ii. BAR British Series 75. Oxford. Gregory R A, Murphy E M, Church M J, Edwards K J, Guttman E B & Simpson D D A ‘Archaeological evidence for the first Mesolithic occupation of the Western Isles of Scotland’ The Holocene 15,7 (2005) 944-50. Gregory, R A 2003 ‘Northton, Harris’ Discovery and Excavation in Scotland new Series 4, 2003 Grøn, O & Skaarup, J 1993 ‘Mollegabett II – Submerged Mesolithic Site and a ‘boat burial’ from Aero’, Journal of Danish Archaeology 10:38-50. Haggarty, A 1991 ‘Machrie Moor, Arran: recent excavations at two stone circles.’ Proc. Soc. Ant. Scot. 121, 1991, 51-94 Haggarty, A. 1991 Machrie Moor, Arran: recent excavations at two stone circles. Proceedings of the Society of Antiquaries of Scotland 121, 51–94. Hamilton, M, Sharples, N M & Pollard, J 2012 Early Bronze Age settlements at Machair Mheadhanach and Cill Donnain. In M Parker Pearson (ed), From Machair to Mountains: Archaeological Survey and Excavation in South Uist, 199–214. Oxford: Oxbow. Hardy K and Wickham-Jones C (eds) 2010 Mesolithic and later sites around the Inner Sound, Scotland: the work of the Scotland’s First Settlers project 19982004 SAIR 31 Society of Antiquaries of Scotland and Council for British Archaeology Hartwell, B. 1998 The Ballynahatty complex. In A.M. Gibson and D.D.A. Simpson (eds), Prehistoric Ritual and Religion, 32–44. Stroud: Tempus. Hawkes, J 1968 ‘The Proper Study of Mankind’, Antiquity, 42 (1968) 255-62. Hawkins, G 1965 ‘Callanish, a Scottish Stonehenge’,

Science, 147 (1965) 127-30. Hawkins, G 1971, ‘Photogrammetric Survey of Stonehenge and Callanish’, National Geographic Society Research Reports for 1965, 101-8. Henderson, J C (ed) 2000 The Prehistory and Early History of Atlantic Europe. BAR International Series 861. Oxford. Henley, C 2004 ‘Falling of the edge of the Irish Sea: Clettraval and the two-faced Neolithic of the Outer Hebrides’ In V. Cummings & C. Fowler (eds), The Neolithic of the Irish Sea: materiality and traditions of practice. Oxford: Oxbow Henley, C 2005 ‘Choreographed monumentality: recreating the centre of other worlds at the monument complex of Callanish, western Lewis’ in Cummings V & Pannett A 2005 (eds) Set in Stone: New approaches to Neolithic Monuments in Scotland, 95-106. Oxbow. Oxford. Henshall, A S 1963 The Chambered Tombs of ScotlandVol 1Edinburgh University Press, Edinburgh Henshall, A S 1970 Small finds. In J.N.G. Ritchie, Excavation of the chambered cairn at Achnacreebeag, 39–41. Proceedings of the Society of Antiquaries of Scotland 102 (1969–790), 31–55. Henshall, A S 1972 The Chambered Tombs of Scotland, Vol 2. Edinburgh: Edinburgh University Press. Henshall, A S 1993 The Grooved Ware: vessels P41–82. In G.J. Barclay and C.J. Russell-White, Excavations in the ceremonial complex of the fourth to second millennium BC at Balfarg/Balbirnie, Glenrothes, Fife, 94–108. Proceedings of the Society of Antiquaries of Scotland 123, 43–210. Henshall, A S 2009 In Ritchie, A 2009 On the fringe of Neolithic Europe: The excavation of a chambered tomb at Holm of Papa Westray, Orkney. Society of Antiquaries of Scotland. Edinburgh. Henshall, A S and Mercer, R J 1981 Report on the pottery from Balfarg, Fife. In R.J. Mercer, The excavation of a Late Neolithic henge-type enclosure at Balfarg, Markinch, Fife, 128–39. Proceedings of the Society of Antiquaries of Scotland 111, 63–171. Henshall, A S & Ritchie, J N G 1995 The Chambered Cairns of Sutherland Edinburgh University Press, Edinburgh Henshall, A S & Wallace, J C 1963 ‘The excavation of a Chambered Cairn at Embo, Sutherland’ Proceedings of the Society of Antiquaries of Scotland 96 (1962-63), 9-36 Bibliography \ 1156

Calanais Survey and Excavation, 1979-88

Hingley, R 1996 ‘Ancestors and identity in the later prehistory of Atlantic Scotland: the reuse and reinvention of Neolithic monuments and material culture’ World Archaeology 28, 231-243 Hingley, R 1999 ‘The Creation of Later Prehistoric Landscapes and the Context of the Reuse of Neolithic and Earlier Bronze Age Monuments in Britain and Ireland’ in Bevan B (ed) Northern Exposure: interpretative devolution and the Iron Ages in Britain. Leicester Archaeological Monographs 4. Leicester. Hirons, K R 2003 ‘Preliminary investigation of lake sediments from the machairs of the Outer Hebrides’, in Barber, J 2003 Bronze Age farms and Iron Age farm mounds of the Outer Hebrides, 199-205. Edinburgh: SAIR Monog 3. Historic Scotland 1998 Nomination of Neolithic Orkney for Inclusion in the World Heritage List. Historic Scotland. Edinburgh. Hodder I and Orton C 1976 Spatial analysis in archaeology. Cambridge University Press. Cambridge. Holtby, I, Scarre, C, Bentley, R A & Rowley-Conwy, P 2012 ‘Disease, CCR5-δ32 and the European spread of agriculture? A hypothesis’Antiquity 86 Number 331, 207-210. Huang, G T 2008, ‘Essence of Thought’ New Scientist Vol 198 No 2658 31 May 2008, 30-33 Hubbard, C E 1954 Grasses, Penguin Books Limited Innes, C 1860 ‘Notice of the stone circle of Callernish ...’ Proc Soc. Antiq Scot 3 (1860), 110-12 James, H 1867 Stonehenge and Tursachan in the Island of Lewis, Southampton. Jensen, Å 1973 Sten. Mineraler, Ædelsten, Bjergarter, Malme. G.E.C. Gads Forlag, København. Johnson, M 2006a Pottery. In D D A Simpson, E M Murphy & R A Gregory, Excavations at Northton, Isle of Harris, 44–69. Oxford: Oxbow. Johnson, M 2006b The pottery. In T. Neighbour, Excavation of a Bronze Age Kerbed Cairn at Olcote, Breasclete, Near Calanais, Isle of Lewis, 28–33. Scottish Archaeological Internet Reports 13. http:// www.sair.org.uk/sair13 Johnson, M, Flitcroft, C & Verrill, L in prep. ‘Calanais Fields Project: Excavation of a prehistoric sub-peat field system at Calanais, Isle of Lewis, 1999-2000’ Jones, A M 2005 The Grooved Ware from Barnhouse. In C Richards (ed), Dwelling Among the Monuments: an Examination of the Neolithic Village of Barnhouse, Maeshowe Passage Grave and Surrounding

Monuments at Stenness, Orkney, 261–281. Cambridge: McDonald Institute (Monograph Series) Jones, A M & Thomas, C 2010 ‘Bosiliack and a Reconsideration of Entrance Graves’ Proc. Prehist. Soc.76, 271-296 Keble-Martin, W 1965 The Concise British Flora in Colour, Ebury Press, Michael Joseph Limited Keiller, A & Piggott, S 1936 ‘The West Kennet Avenue…‘, Antiquity, 10 (1936) 417-27. Knappett, C, Evans, T & Rivers, R 2008 ‘Modelling maritime interaction in the Aegean Bronze Age’ Antiquity 82, 1009-1024 Knappett, C, Rivers, R & Evans, T 2011 ‘The Theran eruption and Minoan palatial collapse: new interpretations gained from modelling the maritime network’ Antiquity 85, 1008-1023 Lambeck, K 1995a ‘Glacial isostasy and water depths in the Late Devensian and Holocene on the Scottish Shelf west of the Outer Hebrides’ Journal of Quaternary Science 1995, 10(1), 83-86 Lambeck, K 1995b ’Late Devensian and Holocene shorelines of the British Isles and North Sea from models of glacio-hydro-isostatic rebound.’ Journal of the Geological Society, London, Vol 152, 1995, 437-448 Lane, A 2007. Ceramic and Cultural Change in the Hebrides AD 500–1300. Cardiff: University of Cardiff (Cardiff Studies in Archaeology, Specialist Report 29). http://www.cardiff.ac.uk/hisar/resources/ CSA¬_29_A_Lane.pdf Lanting, J N and van der Waals, J D 1972 British Beakers as seen from the Continent: a review article. Helinium 12: 20–46. Lawson, J 2001 ‘Cramond, Edinburgh’ D&ES New Series 2, 2001, 124 Lelong, O & MacGregor, G 2007 The Lands of Ancient Lothian: Interpreting the Archaeology of the A1. Edinburgh: Society of Antiquaries of Scotland. Lemercier, O 2012 ‘Interpreting the Beaker Phenomenon in Mediterranean France: an Iron Age analogy’ Antiquity 86 Number 331, 131-144 Lewis, A L 1883 ‘On the relation of Stone Circles to Outlying Stones.. .‘, J. Roy. Archaeol. Assoc., 12 (1863) 176-91. Lewis, A L 1909, ‘Some Stone Circles in Ireland’, J. Roy. Archaeol. Assoc., 39 (1909) 517-29. Lockyer, N 1906 Stonehenge and Other British Stone Monuments Astronomically Considered, London. Bibliography \ 1157

Calanais Survey and Excavation, 1979-88

Logan, J 1831 The Scottish Gael II, 332, fig page 397. Lomax, T A and Edwards, K J 2000 ‘Pollen and related studies of human impact at Loch Bharaphat’, in Harding, D W and Dixon, T N (eds) Dun Bharaphat, Cnip. An Iron Age settlement in West Lewis, 110-14. Edinburgh: Dept. of Archaeology Calanais Research Series 2. Longworth, I.H. and Cleal, R. 1999 Grooved Ware gazetteer. In R. Cleal and A. MacSween (eds), Grooved Ware in Britain and Ireland, 177–206. Oxford: Oxbow (Neolithic Studies Group Seminar Papers 3). Lyne, E. 2008 Built according to plan: two enigmatic structures on the M3. Seanda 3, 23–5. MacAlpine, N 1929 MacAlpine’s Pronouncing Gaelic Dictionary New Large Type Edition. MacLaren & Sons. Glasgow. (The dictionary was first published in 1832 and reissued numerous times) MacCulloch, J 1819, A Description of the Western Islands of Scotland, Edinburgh. MacCulloch, J 1824, The Highlands and Western Isles of Scotland, Edinburgh. McCullagh, R P J 1983‘Callanish; the charcoal’. unpublished report for Historic Scotland. Macdonald D 1978 Lewis: A History of the Island Edinburgh Gordon Wright Macfarlane, W 1907 Geographical Collections, Edinburgh. MacIver, I F 1989 ‘A 17th Century ‘Prose Map’’ in MacLeod F 1989, 23-32 Mackenzie, C 1792 ‘An Account of Some Remains of Antiquity in the Island of Lewis’, Archaeol. Scot. 1 (1792),182-92. Maclagan, C 1879, unpublished plan in NMRS, RCD 113110. MacLean, G & MacLeod, F 1989 ‘Captain Otter & Captain Thomas’ in MacLeod F 1989, 117-122 MacLeod, F (ed) 1989 Togail Tir: Marking Time Acair. Stornoway. MacSween, A 1992 ‘Orcadian Grooved Ware’ in Sharples N and Sheridan A (eds) Vessels for the Ancestors: essays on the Neolithic of Britain and Ireland 259-271. Edinburgh. McVean, D A & Ratcliffe, D V 1962 Plant communities of the Scottish Highlands. Edinburgh: HMSO. Magee, B 1973 Popper. Fontana. London. Maitland, W 1757 History and Antiquities of Scotland, I, Edinburgh.

Mannion, A & Moseley, S P forthcoming ‘Pollen and diatom diagrams from Loch Scolpaig and Balelone Farm, North Uist’, in Barber, J forthcoming Bronze Age farms and Iron Age farm mounds of the Outer Hebrides. Edinburgh: STAR. Marshall, P 2009 Barpa Langais Radiocarbon Dating. Unpublished report for ARCUS, Sheffield. Martin, M 1703 A Description of the Western Islands Of Scotland (circa 1695) London May, A M 1953 ‘Neolithic Habitation Sites, Stone Circles and Alignments at Beaghmore, Co. Tyrone’, J. Roy. Soc. Antiq. Ir., 83 (1953) 174-97. Megaw, J V S & Simpson, D D A 1979 Introduction to British Prehistory Leicester University Press Leicester Mellars, P 1976 ‘Fire ecology, animal populations and man: a study of some ecological relationships in prehistory’, Proc Preh Soc 42: 15-45. Mellars, P 2004 ‘Mesolithic Scotland, Coastal Occupation, and the Role of the Oronsay Middens’ in Saville A (ed) 2004 Mesolithic Scotland and its Neighbours, 167-184 Melton, N D 2008 ‘West Voe: a Mesolithic-Neolithic transition site in Shetland’, in Noble, G et al (eds) Scottish Odysseys. The archaeology of islands, 23236. Stroud: Tempus. Mercer, R J & Rideout, J 1987 ‘The excavation of a cistgroup at Upper Largie, Kilmartin, central Argyll’, Glasgow Archaeol. J. 14 (1987), 25-38 Mercer, R J 1981, ‘The excavation of a late Neolithic henge-type enclosure at Balfarg, Markinch, Fife, Scotland’ Proc. Soc. Antiq . Scot. 111 (1981), 63-171 Michelet, C 1994 (trans. Dickie S) Applewood Orion. London Midgeley, M S 2004 ‘Settlement in north-west Europe: an overview’ in Shepherd & Barclay (eds) 2004, 115-124 Mills, C M, Armit, I, Edwards, K J, Grinter, P and Mulder, Y 2003 ‘Neolithic land-use and environmental degradation: a study from the Western Isles of Scotland’, Antiquity 76: 886-895. Mitchell, M E C 1934 A new analysis of the early bronze age Beaker pottery of Scotland. Proceedings of the Society of Antiquaries of Scotland 68 (1933–4), 132–89. Mithen, S (ed) 2000 Hunter-gatherer landscape archaeology: the Southern Hebrides Mesolithic Project 1988-98. Vols 1-2 McDonald Institute for ArchaeBibliography \ 1158

Calanais Survey and Excavation, 1979-88

ological Research. Cambridge Moir, G 1978 ‘Standing Stones and Astronomy in the Outer Hebrides’, Northumberland Archaeological Group News, 2, 1 (1978) 13—23. Műller, J 1988 The chambered cairns of the Northern and Western Isles: Architectural Structure, Information Transfer and Locational Processes Department of Archaeology University of Edinburgh Occasional Paper 16 Edinburgh Murray, H 2004 ‘Crathes Warrenfield’ Discovery and Excavation in Scotland New Series Vol 5 2004, 155 Murray, H, Murray, C & Fraser, S 2009 ‘Crathes Warrenfield’ British Archaeology 107 July August 2009, 13-17 Needham, S P 2005 Transforming Beaker culture in north-west Europe: processes of fusion and fission. Proceedings of the Prehistoric Society 71, 171–217. Neighbour, T 2005 Excavation of a Bronze Age Kerbed Cairn at Olcote, Breasclete, Near Calanais, Isle of Lewis. Scottish Archaeological Internet Reports 13. Soc Ants Scot. Edinburghhttp://www.sair.org. uk/sair13/index.html Neighbour, T 2005 Excavation of a Bronze Age Kerbed Cairn at Olcote, Breasclete, Near Calanais, Isle of Lewis. Scottish Archaeological Internet Reports 13. http://www.sair.org.uk/sair13 Newell, P J 1988 ‘A buried wall in peatland by Sheshader, Isle of Lewis’, Proc Soc Antiq Scot 118: 79-93. NMRS The National Monuments Record of Scotland, Edinburgh. Noble, G 2009 ‘Forteviot’ Current Archaeology 231, 12-19 Norenzayan A 2012 ‘The idea that launched a thousand civilisations’ New Scientist 231 No 2856, 42-4 O’Kelly, C 1971 Illustrated Guide to New Grange, Wexford. Oppenheimer, S 2007 The Origins of the British. Robinson Ltd, London OSA 1796 First Statistical Account..., Edinburgh. Pailler, Y & Sheridan, A 2009 ‘Everything you always wanted to know about … la néolithisation de la Grande-Bretagne et de l’Irelande’ Bulletin de la Société préhistorique française 106, No 1, 25-56 Parker Pearson, M 2012 Settlement, agriculture and society in South Uist before the Clearances, In M Parker Pearson, M (ed) From Machair to Mountains: Archaeological Survey and Excavation in South Uist, 401–425. Oxford: Oxbow.

Parker Pearson, M & Ramilisonina 1998 ‘Stonehenge for the ancestors: the stones pass on the message’ Antiquity 72 (1998), 308-26 Parker Pearson, M, Chamberlain, A, Jay, M, Marshall, P, Pollard, J, Richards, C, Thomas, J, Tilley, C & Welham, K 2009 ‘Who was buried at Stonehenge?’ Antiquity 83 (2009),23-39 Parker Pearson, M, Cleal, R, Marshall, P, Needham, S, Pollard, J, Richards, C, Ruggles, C, Sheridan, A, Thomas, J, Tilley, C, Welham, K, Chamberlain, A, Chenery, C, Evans, J, Knüsel, C, Linford, N, Martin, L, Montgomery, J, Payne, A & Richards, M 2007 ‘The Age of Stonehenge’ Antiquity 81 (2007), 617-639). Parker Pearson, M, Sharples, N and Symonds, J 2004 South Uist: Archaeology and History of a Hebridean Island. Stroud: Tempus. Peacock J D, Austin, W E N, Selby I, Graham D K, Harland R & Wilkinson I P 1992 ‘Late Devensian and Flandrian palaeoenvironmental changes on the Scottish continental shelf west of the Outer Hebrides’ Journal of Quaternary Science (1992) 7 (2), 145-61 Phemister, J 1948 British Regional Geology, Scotland: The Northern Highlands. HMSO Piggott, S 1948 ‘Excavations at Cairnpapple Hill, West Lothian, 1947-48’ Proceedings of the Society of Antiquaries of Scotland 82 1947-1948, 68-123 Piggott, S 1954 Neolithic Cultures of the British Isles, Cambridge. Piggott, S 1973 ‘A Note on Climatic Deterioration in the First Millennium BC in Britain’, Scottish Archaeol. Forum, 4 (1973) 109-113. Piggott, S 1975 The Druids, Book Club Associates. London. Pilcher, J R 1969 ‘Archaeology, Palaeoecology, and C14 Dating of the Beaghmore Stone Circle Site’, Ulster J. Archaeol., 32 (1969) 73—91. Pitt Rivers 1885 (n.p.), Public Record Office WORK 39/15 118-129 Pollard T 1997 ‘Excavation of a Neolithic settlement and ritual complex at Beckton Farm, Lockerbie, Dumfries & Galloway’, Proc. Soc. Antiq . Scot. 127 (1997), 69-121 Pollard, J & Ruggles, C 2001 ‘Shifting Perceptions: Spatial Order, Cosmology, and Patterns of Deposition at Stonehenge.’ Cambridge Archaeological Journal 11:1 (2001), 69-90 Bibliography \ 1159

Calanais Survey and Excavation, 1979-88

Ponting, G & Ponting, M 1979 Callanish: the Documentary Record. Callanish. G & M Ponting. Callanish, Isle of Lewis.. Ponting, G & Ponting, M 1981 Callanish, the documentary record. Part Two: the Minor Sites. G & M Ponting. Callanish, Isle of Lewis. Ponting, G & Ponting, M 1981 Decoding the Callanish Complex - some initial results in Ruggles & Whittle 1981 Ponting, G & Ponting, M 1982 Decoding the Callanish Complex - a progress report in Heggie 1982 Ponting, G & Ponting, M 1984a New light on the stones of Callanish. Ponting, G & Ponting, M 1984b The stones around Callanish. G & M Ponting. Callanish Ponting, G & Ponting, M 2000 The stones around Callanish.Curtis M R & G R. Calanais. Ponting, M 1988 Megalithic Callanish in Ruggles 1988, 423-41 Ponting, M and Ponting, G 1984. Dalmore. Current Archaeology 8, 230–5. Popper K 1935 Logik der Forschung Verlag von Julius Springer. Vienna. Popper K 1976 Unended Quest: an Intellectual Biography.Fontana/Collins. London Popper, K 2002 The Logic of Scientific Discovery Routledge Classics. London and New York. Pratchett, T 1987, Equal Rites. Victor Gollancz. London. Ralston, I B M 1982 A timber hall at Balbridie Farm. Aberdeen University Review 168, 238–49. RCAHMS 1928 Inventory of Outer Hebrides etc. Edinburgh: Royal Commission on the Ancient and Historical Monuments of Scotland. RCAHMS 1988 Argyll. An Inventory of the Monuments. Volume 6: Mid Argyll and Cowal. Edinburgh: Royal Commission on the Ancient and Historical Monuments of Scotland. Renfrew, C 1979 Investigations in Orkney. Society of Antiquaries. London. Renfrew C & Bahn P 2008 Archaeology: Theories, Methods and Practice Thames and Hudson. London. Renfrew, C, Harkness, D D & Switsor, R 1976 ‘Quanterness, Radiocarbon and the Orkney Cairns’, Antiquity, 50 (1976) 194-204. Richards, C & Wright, J 2006 ‘Clach an Trushal D&ES New Series 7, 2006, 17 CSA Edinburgh Richards, C 1993, ‘Monumental choreography, archi-

tecture and spatial representation in later Neolithic Orkney’ in Tilley C (ed) 1993, Interpretative Archaeology, 143-81. Berg. Oxford. Richards, C 1996, Monuments as Landscape: creating the centre of the world in late Neolithic Orkney’. World Archaeology 28, 190-208 Richards, C (ed) 2005 Dwelling among the monuments: the Neolithic village of Barnhouse, Maeshowe passage grave and surrounding monuments at Stenness, Orkney McDonald Institute Monographs. Cambridge Richards, C 2006 ‘Calanais Meadows’ D&ES New Series 7, 2006, 182 CSA Edinburgh Richards, C 2006 ‘Na Dromannan’ D&ES New Series 7, 2006, 184 CSA Edinburgh Richards, J 1990 The Stonehenge Environs Project English Heritage. London. Ritchie, A 2009 On the fringe of Neolithic Europe: The excavation of a chambered tomb at Holm of Papa Westray, Orkney. Society of Antiquaries of Scotland. Edinburgh. Ritchie, J 1920 ‘The Stone Circle at Broomend of Crichie’, Proc. Soc. Antiq. Scot., 65 (1920) 154-71. Ritchie, J N G 1970 Excavation of the chambered cairn at Achnacreebeag. Proc Soc Antiq Scot 102 (1969–70), 31–55. Ritchie, J N G 1976 ‘The Stones of Stenness, Orkney’ Proceedings of the Society of Antiquaries of Scotland 107 (1975-6) 1-60 Ritchie, J N G & Crawford, J 1978 ‘Excavations at Sorisdale and Killunaig, Coll.’ Proceedings of the Society of Antiquaries of Scotland 109 (1977-8)75-84 Ritchie, J N G. and Stevenson, J 1982 Cists at Traigh Bhan, Islay, Argyll. Proceedings of the Society of Antiquaries of Scotland 112, 550–9. Ritchie, J N G and Welfare, H 1983 Excavations at Ardnave, Islay. Proceedings of the Society of Antiquaries of Scotland 113, 302–66. Ritchie, W 1985 ‘Inter-tidal peats and sub-tidal organic deposits and sea level changes in the Uists, Outer Hebrides’, Scott J Geol 21, 161-176. Ritchie, W 1972 ‘The Evolution of Coastal Sand Dunes’, Scot. Geog. Mag., 38 (1972) 19-35. Ritchie, W 1979 Machair development and chronology in the Uists and adjacent islands., in Boyd, J M (ed) 1979, 107.122. Robson, M 1989 ‘The Living Voice’ in MacLeod, F 1989, 97-104 Bibliography \ 1160

Calanais Survey and Excavation, 1979-88

Rosenblum B and Kuttner F 2006 Quantum Enigma: Physics Encounters Consciousness. Gerald Duckworth & Co. London. Ruggles, C L N & Whittle, A W R 1981 (eds) ‘Archaeoastronomy and Society in Britain during the period 4000-1500 BC. Oxford. BAR Ruggles, C L N 1988 Records in Stone: Papers in memory of Alexander Thom. Cambridge. CUP. Ruggles, C L N 1999 Astronomy in Prehistoric Britain and Ireland Yale University Press Saville, A 2003 ‘Indications of regionalisation in Mesolithic Scotland’. In L. Larsson et al. (eds), Mesolithic on the Move: Papers Presented at the Sixth International Conference on the Mesolithic in Europe, Stockholm 2000, 340–50. Oxford: Oxbow Books. Saville, A, Hardy, K, Miket, R & Ballin, T B 2012 An Corran, Staffin, Skye: A rockshelter with Mesolithic and later occupation. SAIR 51, 2012 Schulting, R J & Richards, M P 2002 ‘The wet, wild and the domesticated: the Meolithic- Neolithic transition on the West coast of Scotland’ European Journal of Archaeology 5, 147-89. Schulting, R J, Sheridan, JA, Crozier, R & Murphy, E 2010. ‘Revisiting Quanterness: new AMS dates and stable isotope data from an Orcadian chamber tomb’. Proceedings of the Society of Antiquaries of Scotland 140, 1–50. Schulting, R J, Bronk Ramsey, C, Reimer, P, Eogan, G, Cleary, K, Cooney, G &Sheridan, JA forthcoming. Dating Knowth. In G Eogan & K Cleary, Excavations at Knowth. Volume 6. Scott, J G 1964 The chambered cairn at Beacharra, Kintyre, Argyll. Proc Prehist Soc 30, 134-58. Scott, J G 1989 ‘The stone circles at Temple Wood, Kilmartin, Argyll’ Glasgow Archaeological Journal 15, 53-124 Scott, W L 1935 ‘The chambered cairn of Clettraval, North Uist’, Proc Soc Antiq Scot 69, 480-536. Scott, W L 1942 ‘Neolithic Culture of the Hebrides’ Antiquity 16 (1942), 301-6 Scott, W L 1948 The chambered tomb of Unival, North Uist. Proc Soc Antiq Scot 82 (1947–8), 1–48. Scott, W L 1951a ‘Eilean an Tighe; a pottery workshop of the 2nd millennium BC’, Proc Soc Antiq Scot 85, 1-37. Scott, W L 1951b ‘Drift timber in the West’, Antiquity 25, 151-3. Scott, W L 1951c ‘The colonisation of Scotland in the

Second Millennium BC’ Proc Prehist Soc 17, 16-82 Sharples, N M 1983 ‘Dalmore (Uig parish) early bronze age settlement.’ Discovery and Excavation in Scotland1983, 81. Sharples, N M 1992, ‘Aspects of regionalism in the Scottish Neolithic’ in N M Sharples &J A Sheridan (eds) 1992 Vessels for the Ancestors, 322-31 Sharples, N M 2012 ‘The Beaker-period and Early Bronze Age settlement at Sligeanach, Cill Donnain.’ In M Parker Pearson, M (ed) From Machair to Mountains: Archaeological Survey and Excavation in South Uist, 215–258. Oxford: Oxbow. Sharples, N M & Sheridan, J A (eds) 1992, Vessels for the Ancestors.Essays on the Neolithic of Britain and Ireland in Honour of Audrey Henshall. Edinburgh: Edinburgh University Press. Shepherd, I A G 1976 ‘Preliminary results from the Beaker settlement at Rosinish, Benbecula.’ In C. Burgess and R. Miket (eds), Settlement and Economy in the Third and Second Millennia B.C., 209–19. Oxford: British Archaeological Reports (British Series 33). Shepherd, I A G 2012 ‘Is there a Scottish Chalcolithic?’ In M Allen, J P Gardiner & J A Sheridan (eds), Is there a British Chalcolithic? People, Place and Polity in the Latert Third Millennium, 164–171. Oxford: Oxbow (Prehistoric Society Research Paper 4. Shepherd, I A G & Barclay, G J 2004, Scotland in Ancient Europe: The Neolithic and Early Bronze Age of Scotland in their European Context. Soc. Ants. Scot. Edinburgh. Sheridan, J A 1995, Irish Neolithic pottery: the story in 1995’ in Kinnes I & Varndell G (eds) 1995, Unbaked Urns of Rudely Shape, 1-9. Oxbow Monograph 55. Oxford. Sheridan, J A 2000 ‘Achnachreebeag and its French connections: vive the Auld Alliance’ in Henderson J C (ed) 2000, 1-15 Sheridan, J A 2003 ‘French Connections I: spreading the marmites thinly.’ In I. Armit, E. Murphy, E. Nelis and D.D.A. Simpson (eds). Neolithic Settlement in Ireland and Western Britain, 3–17. Oxford: Oxbow. Sheridan, J A 2004a ‘Going round in circles? Understanding the Irish Grooved Ware ‘complex’ in its wider context.’ In H. Roche, E. Grogan, J. Bradley, J. Coles and B. Raftery (eds), From Megaliths to Metals: Essays in Honour of George Eogan, 26–37. Oxford: Oxbow. Bibliography \ 1161

Calanais Survey and Excavation, 1979-88

Sheridan, J A 2004b ‘Scottish Food Vessel chronology revisited.’ In A.M. Gibson and J.A. Sheridan (eds), From Sickles to Circles: Britain and Ireland at the Time of Stonehenge, 243–69. Stroud: Tempus. Sheridan, J A 2004c ‘Neolithic connections along and across the Irish Sea,’ in V. Cummings &C. Fowler (eds.), The Neolithic of the Irish Sea: Materiality and Traditions of Practice, 9-21. Oxford. Sheridan, J A 2005 ‘The National Museums’ of Scotland radiocarbon dating programmes: results obtained during 2004/5.’ Discovery and Excavation in Scotland 6, 182–3. Sheridan, J A 2007a ‘From Picardie to Pickering and Pencraig Hill? New information on the ‘Carinated Bowl Neolithic’ in northern Britain,’ in A.W.R. Whittle & V. Cummings (eds.), Going Over: the Mesolithic–Neolithic Transition in North-West Europe, 441–92. Oxford: Oxford University Press (Proceedings of the British Academy 144). Sheridan, J A 2007b ‘Radiocarbon dates arranged through National Museums of Scotland during 2006/7’ Discovery and Excavation in Scotland New Series 8 2007, 221 Sheridan, J A 2007c ‘Scottish Beaker dates: the good, the bad and the ugly.’ In M. Larsson and M. Parker Pearson (eds), From Stonehenge to the Baltic: Living with Cultural Diversity in the Third Millennium BC, 91–123. Oxford: Archaeopress (British Archaeological Reports International Series 1692). Sheridan, J A 2008a ‘Radiocarbon dates arranged through National Museums of Scotland Archaeology Department during 2007/8.’ Discovery and Excavation in Scotland New Series 9 2008, 202 Sheridan, J A 2008b ‘Upper Largie and Dutch-Scottish connections during the Beaker period.’ In H. Fokkens, B.J. Coles, A.L. van Gijn, J.P. Kleijne, H.H. Ponjee and C.G. Slappendel (eds), Between Foraging and Farming: an Extended Broad Spectrum of Papers Presented to Leendert Louwe Kooijmans, 247–60. Leiden: Leiden University (Analecta Praehistorica Leidensia 40). Sheridan, J A 2009 Pottery from Barpa Langais. Unpublished report for ARCUS, Sheffield. Sheridan, JA 2010 ‘The Neolithisation of Britain and Ireland: the Big Picture.’ in B. Finlayson & G. Warren (eds.), Landscapes in Transition, 89–105. Oxford: Oxbow. (CBRL Supplementary Series.) Sheridan, J A 2012a ‘Contextualising Kilmartin: build-

ing a narrative for developments in western Scotland and beyond, from the Early Neolithic to the Late Bronze Age,’ in A.M. Jones,J. Pollard, M.J. Allen & J. Gardiner (eds.), Image, Memory and Monumentality; archaeological engagements with the material world, 163–183. Oxford: Oxbow/Prehistoric Society (Prehistoric Society Research Paper 5). Sheridan, J A 2012b ‘A Rumsfeld Reality Check: what we know, what we don’t know and what we don’t know we don’t know about the Chalcolithic in Britain and Ireland.’ In M Allen, J P Gardiner & J A Sheridan (eds), Is there a British Chalcolithic? People, Place and Polity in the Later Third Millennium, 40–55. Oxford: Oxbow (Prehistoric Society Research Paper 4. Sheridan, J A 2012c ‘Pottery.’ In N M Sharples, The Beaker-period and Early Bronze Age settlement at Sligeanach, Cill Donnain, 227–236 in M Parker Pearson (ed). Sheridan, J A and Higham, T. 2006 ‘The re-dating of some Scottish specimens by the Oxford Radiocarbon Accelerator Unit (ORAU)’. Discovery and Excavation in Scotland 7, 202–4. Sheridan, J A and Higham, T. 2007 ‘The re-dating of some Scottish specimens by the Oxford Radiocarbon Accelerator Unit (ORAU): results received during 2007.’ Discovery and Excavation in Scotland 8, 225–6. Simpson D D A 1996 ‘Excavation of a kerbed funerary monument at Stoneyfield, Raigmore, Inverness, Highland, 1972–3’ Proc Soc Antiq Scot 126, 1996, 53-86 Simpson, D D A, Murphy, E M & Gregory, R A 2006 Excavations at Northton, Harris British Archaeological Reports (British Series 408). Oxford. Smith, D I & Fettes, D J 1979: The geological framework of the Outer Hebrides. InBoyd, J.M. (ed.) 1979: The Natural Environment of the Outer Hebrides. Proceedings of the Royal Society of Edinburgh, Section B (Biological Sciences) 77. Edinburgh: The Royal Society of Edinburgh. Smith, I F 1965 Windmill Hill and Avebury. Excavations by Alexander Keiller, 1929-35. Oxford. Sokal A & Bricmont J 1998 Intellectual Impostures. Profile Books Ltd. London. Somerville, H B 1912, ‘Astronomical Indications in the Megalithic Monument at Callanish’, J. Brit. Astron. Assoc., 23 (1912) 83. Bibliography \ 1162

Calanais Survey and Excavation, 1979-88

Speak, S & Burgess, C 1999 ‘Meldon Bridge: a centre of the third millennium BC in Peebleshire’ Proceedings of the Society of Antiquaries of Scotland 129, 1-118 Spinney, L 2008 ‘Where are the bodies?’ New Scientist 8 Nov 08, 40-43 Stevens, C J & Fuller, D Q ‘Did Neolithic farming fail? The case for a Bronze Age agricultural revolution in the British Isles’ Antiquity 86 (2012), 707-722 Stuart, J 1860 ‘Note of Incised Marks on one of a Circle of Standing Stones in the Island of Lewis’ Proceedings of the Society of Antiquaries of Scotland 3 (1857-60), 212-4 Stukeley, W 1743, Abury, a Temple of the Druids, with Some Others, Described, London. Tait D 1978 Callanish: A Map of the Standing Stones and Circles at Callanish, Isle of Lewis, with a detailed plan of each site. Glasgow University Department of Geography. Glasgow. Terry, J 1997 ‘Upper Largie’ Discovery and Excavation in Scotland 1997, 19-21 Thom, A 1967 Megalithic Sites in Britain, Oxford. Thom, A 1971 Megalithic Lunar Observatories, Oxford. Thomas, F W L 1860a ‘Notice of Beehive Houses in Uig…‘, Proc. Soc. Antiq. Scot., 3 (1860) 127—34. Thomas, F W L 1860b ‘Remarks on the same stone circle of Callernish’, Proc. Soc. Antiq. Scot., 3 (1860), 112. Thomas, J 1998 ‘Towards a Regional Geography of the Neolithic’ in Edmonds & Richards 1998, 37-60 Thomas, J 2005 ‘Current debates on the Mesolithic-Neolithic transition in Britain and Ireland’ Documenta Praehistorica 31, 113-30 Thomas, J 2007 Place and Memory: Excavations at the Pict’s Knowe, Holywood and Holm far, Dumfries and Galloway, 1994-8. Oxbow Books. Oxford. Thomas, J forthcoming, on excavations at Dunragit Tilley, C 1996 ‘The powers of rocks: topography and monument construction on Bodmin Moor’. World Archaeology 28, 161-76 Tilley, C 1998 ‘Megaliths in Texts’ in Edmonds & Richards 1998, 141-160 Tipping, R 1994 ‘The form and fate of Scotland’s Woodlands’ Proc Soc Antiq Scot 124 (1994), 1-54 Tipping, R 1996 ‘Microscopic charcoal records, inferred human activity and climate change in the mesolithic of northernmost Scotland’, in Pollard, A and Morrison, A (eds) The Early Prehistory of Scotland, 39-61. Edinburgh: Edinburgh University Press.

Tipping, R 2000. Pollen preservation analysis as a necessity in Holocene pollen analysis. In Huntley, J P & Stallibrass, S (eds) Taphonomy and Interpretation. Oxford: Oxbow, 23-34. Tipping, R 2004 ‘Interpretative Issues Concerning the Driving Forces of Vegetation Change in the Early Holocene of the British Isles’ in Savile A (ed) 2004, 45-54 Toland, J 1726, The History of the Druids. University of Glasgow Department of Geography 1978 see Tait 1978 Wainwright G J & Longworth I H 1971 Durrington Walls Excavations 1966-1969. Society of Antiquaries. London. Wainwright, G J and Longworth, I H 1971. The Rinyo-Clacton Culture reconsidered. In G J Wainwright, with I H Longworth, Durrington Walls: Excavations 1966–68, 235–306. London: Society of Antiquaries of London (Research Committee Report 29). Walker, M J C 1984 ‘A pollen diagram from St. Kilda, Outer Hebrides, Scotland’, New Phytol 97: 99-113. Ward, G K & Wilson, S R 1978 ‘Procedures for comparing and combining radiocarbon age determinations: a critique’ Archaeometry 20 (1) 1978 19-31 Watkins, T & Shepherd, I A G 1980. A Beaker burial at Newmill, near Bankfoot, Perthshire. Proceedings of the Society of Antiquaries of Scotland 110 (1978–80), 32–41. Welfare, A 2011 Great Crowns of Stone: The recumbent Stone Circles of Scotland RCAHMS Edinburgh. Whitehouse, H 2000 Arguments and icons: divergent modes of religiosity. Oxford University Press. Oxford. Whittle, A, Healy F & Bayliss A 2011 ‘The domestication of Britain’ British Archaeology119 ( July-August 2011), 14-21 Wickham-Jones, C 1990 Rhum. Mesolithic and Later Sites at Kinloch. Excavations 1984-86. Soc. Antiq. Scot. Monograph Series 7. Edinburgh. Wickham-Jones, C 1994, Scotland’s First Settlers. Batsford. London. Wickham-Jones, C 2004, ‘Structural Evidence in the Scottish Mesolithic’ in Saville A (ed) 2004 Mesolithic Scotland and its Neighbours, 229-242 Wilkins, D A 1984 ‘The Flandrian woods of Lewis (Scotland)’, J Ecol 72: 251-258. Wilson, D 1863 The Archaeology and Prehistoric AnBibliography \ 1163

Calanais Survey and Excavation, 1979-88

nals of Scotland, Woodman, P 2012 ‘Making Yourself at Home on an Island: The First 1000 Years (+?) of the Irish Mesolithic.’ Proc. Prehist. Soc 78, 2012, 1 - 35 Worsaae, J J 1846 (n.p.) Notebooks in National Museum of Denmark, Copenhagen) Young, A 2011 ‘Moving the stones’ British Archaeology 117, 2011, 44-5 Zvelebil, M 1998 ‘What’s in a Name: the Mesolithic, the Neolithic, and Social Change at the Mesolithic-Neolithic Transition’ in Edmonds & Richards 1998, 1-36

Bibliography \ 1164

Calanais Survey and Excavation, 1979-88

Appendix 1 Approaches and biases A short version of this can be found in Chapter 2 Introduction 2.8. It helps readers to understand what authors have written if the latter’s prejudices are transparent. This appendix reveals some of mine. I learnt archaeology in the late 1960s at the University of Wales, Cardiff, largely from Richard Atkinson, Leslie Alcock, Mike Jarrett, Bill Manning, John Evans and some of my contemporaneous students. I was part of an undergraduate and postgraduate group inspired by the prehistoric and early historic archaeology of Britain. At that time culture-historical approaches were being influenced by processual ideas. I studied British archaeology because I was at least as interested in things as in people. Many may find that deplorable, but objects and the connections between them have their own fascinations, and if one is really focussed on people rather than things then history, ethnology, Pleistocene archaeology and a host of social and hard sciences offer a more direct route to their study than the archaeology of the post-glacial period in Britain before the advent of written records. By the time Calanais was excavated in 1980 and 1981 I had come to think that much archaeological theorising was the result of academic fashions, its fabric woven on second-hand looms with a small range of bobbins. In promulgating ideas many archaeologists tended to quote authorities favouring particular schools of thought rather than use original evidence. I had not read Kuhn’s ‘The Structure of Scientific Revolutions”, published in 1962; had I done so I would have realised that much academic archaeology was conducted within the framework of ‘normal science’ described by him (Fuller 2003, 18-22, 216). The problems associated with ‘normal science’ persist. I find no help in most of the forms of interpretive archaeology listed by Renfrew and Bahn (2008, 491); and in particular the extreme forms of relativism seem to me nihilistic and highly unlikely to increase understanding. But I do accept that even

archaeological theorising which goes far beyond the primary evidence has the merit of revealing biases and ignorance, my own as well as the authors’. As Terry Pratchett put it: ‘I was confused and uncertain about all the little details of life. But now’, he brightened up, ‘while I am still confused and uncertain it’s on a much higher plane …’ (Pratchett 1987, 156). But my approaches in 1980 and 1981, when most of the excavation described here took place, probably fell most closely into the processual category and during the writing of this report into the cognitive-processual one as defined by Renfrew and Bahn (2008, 495-6). The greatest influence on my thinking has been Karl Popper’s theories of science, first published in 1935 and subsequently translated into English (e.g. Popper 2002). Before 1980 I absorbed them indirectly through the approaches of people whose use of evidence I admired. Of the books which have since provided me with inspiration I single out Bryan Magee’s accessible ‘Popper’ published in 1973 in a Fontana edition (Magee 1973), Popper’s own ‘Unended Quest (Popper 1976) and Steve Fuller’s well-balanced ‘Kuhn vs. Popper (Fuller 2003). Popper’s lesson was that in practise all non-trivial scientific theories are incomplete and that in principle though one can occasionally show that a theory is wrong one can never show that it is right in all circumstances. He accepted that one can say many worthwhile things even if they are not ‘science’, provided that one does not pretend that they are (Magee 1973, 45-6). I have tried to follow his advice that one should foster a scientific culture providing support to those who challenge currently accepted ideas, although I have found it difficult, despite Popper’s advocacy, to welcome written criticism of my own ideas and equally difficult to criticise in writing the ideas of others (Magee 1973, 39). There is however a profound difficulty in applying Popper’s approach: problems must be manageable. One must be able to formulate a theory in such a way that it can in principle be refuted by Appendices \ 1165

Calanais Survey and Excavation, 1979-88

data. More broadly, problems tend to be selected for testing because they are parts of systems which can be described using traditional mathematical or logical tools. This applies to many other scientific disciplines: “The systems discussed [in textbooks of science and mathematics] are usually ones that are specifically chosen to be amenable to fairly complete analysis, and whose behaviour is therefore necessarily simple.” (Wolfram 2002, 115). That can mean that scientific tests are limited to one or at most a few aspects of complex systems. Another influence has been Bayes’ theorem and the intellectual apparatus which surrounds its use. Some aspects of the Bayesian approach are common sense. It ‘emphasises that the interpretation of data is conditional on the information available and on an individual’s understanding of it at that time’ (Buck et al, 1996, 1). It demands explicit modelling based on clearly stated prior assumptions, which often provides a healthy antidote to fuzzily defined interpretations. I do however have concerns about its application. It is designed to give some rigour to inferential approaches but all too often it seems to apply only a ‘scientific’ gloss to received ideas. To those who accept Popper’s approach to learning more about the world inferential methods are inherently flawed and should not be described as scientific. “To choose a ‘likely’ hypothesis and then test it back on the data that helped to suggest it is clearly to move dangerously in the direction of circular argument” (Ruggles 1999, 77). But Bayes’ theorem can equally well be used in a critical framework, and there I include less rigorous attempts at refutation than those advocated by Popper. In published studies the recognition of competing prior beliefs and assessments of their probability are often highly incomplete. With narrow models the real probability of observing an event within a particular interpretational framework is that estimated multiplied by the probability that the interpretation is correct, plus the probability that the event would have occurred even if the interpretation was wrong multiplied by the probability that it was wrong. More broadly, Bayes’ theorem was defined with billiard balls in mind; life is usually more complex than that. Commonly archaeological evidence allows more than one interpretation,

and often they are not mutually exclusive. For instance, the skeletal material in a chambered tomb may have been deposited fairly evenly over time, each time involving the remains of one or a few freshly dead people, or it may have been added very clumpily, perhaps including the freshly dead and the long-dead. The state and ordering of the excavated remains may reflect processes including manipulation and removal of bones; they may be stratified in a different order from the times of death of the people. They may include only a few founder burials and those of the last few generations to use the tomb. This latter may have been what happened for instance at Point of Cott and Holm of Papa Westray in Orkney (Barber 1979; Ritchie 2009). But given the fact that only very specifically stated theories are open to refutation in the way that Popper advocated, I accept that Bayesian approaches to archaeological questions offer a substantial improvement over traditional ones provided that they follow the formal methods advocated by Buck et al (1996, 355-62); and provided that their purely inferential nature is recognised. I thoroughly agree with Buck and her colleagues’ recommendation that the project team should spend a considerable effort on considering the background to the analysis (Buck et al 358). To their Step 10, Interpretation, I would add one phrase: stand back and think again whether the prior(s) and the new evidence can bear more interpretations than those initially contemplated. Indeed the examples they adduce (e.g. Buck et al 1996, 290) include such considerations, but I think the point is worth re-emphasising because there is a danger that, wooed by the wonders of poorly understood computer programmes which magically take in numbers and output others, critical faculties may be lulled. In earlier drafts of this report on Calanais I attempted Bayesian analysis as advocated by Buck and her colleagues at junctures where it seemed important to distinguish between various possibilities. But my attempts proved unsuccessful (except in making me look more critically at my interpretations of the evidence) for they were usually based on an incomplete set of non-mutually exclusive prior beliefs leading only to fuzzy posterior beliefs. I have removed them in this final version, to avoid Appendices \ 1166

Calanais Survey and Excavation, 1979-88

what would almost certainly have amounted to no more than a spurious sheen, returning to descriptions like ‘possibly’ and ‘probably’. The unhealthy stepchildren of relativism (Sokal & Bricmont 1998) still rule some areas of modern archaeology. In a non-trivial sense I follow those who believe that there is a reality independent of human (or other conscious) observers. This is perhaps all the more fundamental a declaration given some (in my view carelessly phrased) recent statements by brilliant theoreticians. For instance Hawking and Mlodinow appear to claim in their ‘The Grand Design’ that that because science offers many incomplete windows on reality there is no theory-independent reality (Callender 2010, 50). That phrasing of their view will be grist to the mill of relativists who confuse human weakness with some imagined unreality in the fabric of the universe. And some statements by brilliant scientists are unambiguous and, in my view, wrong. One eminent scientists has said that ‘The universe exists because [and only because] we are aware of it’ (Sir Martin Rees quoted by Rosenblum and Kuttner 2006, 192). Of course the fabric of the cosmos can appear to be ambiguous, wavy or particulate at very small scales; the way in which an experiment is set up can produce evidence for one or the other; and it has been argued that that it is not merely a demonstration that the choice of apparatus reveals different aspects of an underlying reality but that a set of potential realities collapses into a single state due to (for an extended and accessible discussion see Rosenblum and Kuttner 2006; see also Ananthaswamy 2013, 39 for the view that, in the light of recent experiments with a control photon determining the way in which a photon is measured, attempting to describe reality at very small scales with concepts like wave and particle is doomed to failure). But I reject the interpretation that a ‘collapsed’ reality is due to observation by a conscious being, although admittedly with no more hard evidence for so doing than Rees had for saying it (and with full recognition that this belief is held by people with far deeper, fresher and agile brains than mine). I would rather say that ‘We conceptualise aspects of the existence of the universe because we want to understand it’, and that we do not yet have

the concepts to express the basic building blocks of reality (Ananthaswamy 2013, 39 ). In archaeology we should focus on the macroscopic. I believe that the universe trundled merrily on its way before observer-style consciousness evolved, and continues to do so now that it contains conscious beings. Consciousness is a mystery only because it is the result of feedback systems involving the interaction of the body (including the mind) with its environment. Understanding of feedback systems, which are characteristically nonlinear, is very imperfect, although new approaches may eventually remedy that (Wolfram 2002). That underlies my belief that the objects and evidence of processes commonly found in archaeological sites should be regarded as the results of past causes and effects (in the simple everyday sense) operating independent of modern observers. But I do accept that all observations and the frameworks within which they are made are affected by (often unconscious) mental models. ‘… there is no such thing as an unprejudiced observation. All observation is an activity with an aim (to find, or to check, some regularity which is at least vaguely conjectured); an activity guided by problems, and by the context of expectations … There is no such thing as passive experience; no passively impressed association of impressed ideas,’ (Popper 1976, 51-2). There is a theory of mind, ascribed in its 2008 form to Karl Friston of University College, London and colleagues, which suggests not only that the brain works in an inductive Bayesian fashion, but that it also works to minimise prediction error. Sensory information is always compared to the brain’s internal expectations, which are updated by information. But this mechanism leads not only to our ability to make sense from partial information such as mumbled speech but also to our tendency to distort or discard sensory information when it conflicts with internal predictions (Huang 2008, 31-3). Put brutally, the relevance of this is that while excavating I and my colleagues ignored information. I do not in this refer only to the forgetting of information which did not fit in the framework which governed how I looked at it. I mean that much sensory information was discarded by pre-conscious parts of the brain because it did not Appendices \ 1167

Calanais Survey and Excavation, 1979-88

fit expectations. I still believe (it was very much part of the approach I learned as an undergraduate) that excavators must try to interpret at all levels in the field, so that interpretations and indeed observations can be challenged while it is still possible to gain new information from the site itself. It is much chancier to try to interpret the site only during the writing-up process because samples, texts, drawings and photographs provide a limited and distorted extract of the original. But interpretation in the field can reinforce a tendency to ignore ‘irrelevant’ information, and a prescription for the best balance between ‘objectivity’ and interpretation during excavation eludes me; the most one can do is adapt excavation techniques to the evidence, remain conscious that nothing exists independent of its environment, sample systematically and record abundantly, carefully and consistently, in the hope that biases can be subverted by reality. Turning to other aspects of interpretation at Calanais, it is vital to understand what a tiny proportion of the activities which took place there left any recognisable trace. Around two and a half millennia passed between the creation of a curving ditch there and the period when peat covered the place. Most deeds will have left no retrievable evidence in the excavation areas; others will have left unrecognisable traces; relatively few will have left evidence which could be retrieved by the techniques we used in 1980 and 1981, within the conceptual frameworks we then employed. Put succinctly, our work captured only a miniscule proportion of what people did. The same applies to processes. We cannot always have discriminated adequately between one and several episodes of ground-working, nor of natural soil developments. We might have done a little better if we had had an on-site laboratory; our capture of information suffered from the lack. But I doubt whether we could have discriminated between for instance a few and many overlapping processes even with a laboratory to hand. Those, then, are some of the underlying reasons why complete certainty about anything other than basic observations is rarely claimed in this report. I would like to say that that includes all simple assertions, such as that an object was found in a

particular layer, however it got there. But defining layers was in part subjective. Worse, the excavation records include some demonstrable errors. Although I think it unacceptable to pretend certainty where it cannot exist, and have tried to flag up all significant problems in the relevant parts of the narratives, I know that over-use of qualifiers runs counter to people’s general desire for clarity. So detailed descriptions of important ambiguities, and also expressions of vague doubts and uncertainties, have been hosted in the Technical notes in Chapter 25, in appendices, and in the Resources element of the archive hosted by NMRS. P J Ashmore (January 2013)

Appendices \ 1168

Calanais Survey and Excavation, 1979-88

Appendix 2 Area B summary of excavation progress, discoveries and remaining archaeology

This appendix is referred to in Chapter 7 Area B 7.20. Appendix 2.1 Trench layout and excavation progress in 1980

Illus Appendix 2.1 Area B from the southeast during laying out [Film 1980.2.4] Illus A2.1 shows Area B, the monolith and the northern part of the Ring before excavation, with stones of the East Row to the left and avenue stones to the right.

Illus Appendix 2.2 Area B as first opened, from the east [Film 1980.2.27]

Area B was initially laid out as three trenches (Illus Appendix 2.2). The north, south and east sides of BI to BIII were parallel to the National Grid and overall they measured 8m north-south. The west side of Area BI and the central baulk ran 10 degrees to the west of national grid north. The width of Area B thus varied from east to west between 5.5m and slightly under 6.5m; but the various extensions discussed below increased those dimensions. Area BI lay immediately to the east of the Ring of Stones. It was designed to allow investigation of the main part of the interior (120) of an enclosure formed of head-sized stones (103), along with its relationship to Ring stones 42 and 43; its south edge was at the modern drain (100). Area BI was extended to the north-west to allow examination of a prone stone (148) and of standing stone 34, between the ring of stones and the east side of the avenue. That area was called BIN. Area BII covered the north-eastern part of the enclosure wall-base and BIII covered its eastern part. A triangular area between Areas BII and

Illus Appendix 2.3 The southern part of Area B looking west along the north side of the East Row on 12 May 1980 [Film 1980.5.2]

Appendices \ 1169

Calanais Survey and Excavation, 1979-88

BIII was called Area BII/III. The trenches were intended to reveal all of the stones related to the enclosure including what seemed to be tumble outside the visible enclosure. The tumble turned out to be more complex than expected and to incorporate an earlier wall-base, while under the wall-bases a bank formed the earliest enclosure. A further strip of ground between the ditch and the East Row was opened by 12 May 1980 revealing lines and clusters of stones south of the ditch (Illus Appendix 2.3). A trench across the line of the East Row to the south of Trench BI, and extending westward towards the cairn, was called Area BIV. It was designed to illuminate the relationship between the modern drain (100) and the adjacent ridge (101) running eastward from Ring stone 43. During the excavation of Area BIV in 1980 a deep slot trench was cut to see how deep the layers were. The ra-

pidity with which this was done subsequently led to recording problems, although it did show that deep pottery-rich deposits survived. A small extension to the south of Trench BIII along the line of the central baulk was called BV in 1980 but it is treated as part of BIII and called BIIISX in this text to avoid confusion with the 1981 extension of BIV to form BV. Illus Appendix 2.4 shows the extent of the trenches at the end of the 1980 season. Stones which were considered to be in-situ parts of the latest enclosure are in black and others in grey; most had been removed by the end of the season. The modern ditch running from BIV to BIII is shown in light ochre. Appendix 2.2 Excavation progress in 1981

Illus Appendix 2.5Â Area B on 6 May 1981 (near the beginning of the 1981 season) from the north-west [Film 1981.3.26]

Illus Appendix 2.4Â Area B in 1980 [NMRS DC38024]

Excavation in 1980 left many unanswered questions. In 1981 work on the enclosure produced more evidence allowing its interpretation as a four-stage structure. Iron-panned stake holes were discovered in the bank and elsewhere. The extent of a palisade trench in the southern part of the enclosure circuit was defined. The area BIN (BI north of the enclosure) was extended to the west to form BINX, at the bottom left hand side of Illus Appendix 2.5. BINX is sometimes referred to in the records as BINW. Area BIN contained the northern half of a bank seemingly continuing the line of the enclosure bank, two large buried stones (148 and 171), and Appendices \ 1170

Calanais Survey and Excavation, 1979-88

Illus Appendix 2.6Â The south part of Area B from the East on 11 May 1981 [Film 1981.6.7]

Illus Appendix 2.7Â A view of Area B from the west showing an excavation drainage ditch on the east side of BIII, not otherwise recorded. Area H is in the foreground [Film 1981.6.21] Appendices \ 1171

Calanais Survey and Excavation, 1979-88

Illus Appendix 2.8 Area B on 17 May from the north before removal of the prone slab in Area BIWX [Film 1981.9.31] part of the pit for Avenue stone 34. Area BIV was widened and extended southward on 7 May 1981 to form BV, to the left on Illus Appendix 2.6, somewhat hidden by East Row stone 30. The sondage dug in BIV in 1980 is

Illus Appendix 2.9 Moving the prone slab from BIWX [Film 1981-10-18]

visible on the photograph. The stratigraphy of the area proved to be complicated with evidence for deposition of burial material and ground-working. The part of BIV north of the modern ditch was excavated only down to a layer corresponding roughly to the period when the Ring was built, revealing remains of a burial and other deposits including much pottery. But because of the differences in stratigraphic depth over the area preRing strata, including a shallow trough running between cultivation beds, were revealed in the southern part of BV. Area BI was extended westward in 1981 to al-

Illus Appendix 2.10 Area B on 22 May from the south-east after the prone slab had been moved [Film 1981.12.35] Appendices \ 1172

Calanais Survey and Excavation, 1979-88

Illus Appendix 2.11 Area B extension BIWX from the east on 25 May 1981 [Film 1981.14.25]

low excavation of the area around the prone slab between Ring Stones 42 and 43 (BIWX) and of the passage (BIVWX). Before excavation much of Area BIWX was covered by the prone slab (122) measuring 1.5 m by 0.8 m and 0.25 m thick. The prone slab was removed and temporarily stored on the baulk between BI and BIII (Illus Appendix 2.9) Discovery of slots corresponding to the base of the slab allowed it to be re-erected at the end of the 1981 season. Extension BIVWX revealed almost none of the original south passage wall, on the left of the drain in Illus Appendix 2.11, survived. The west part of the north passage wall was well built (Illus Appendix 2.12) and covered a shallow but complex succession of earlier features cut

Illus Appendix 2.12 Area B extension BIVWX from the east c. 20 May [Film 1981.11.22]

Illus Appendix 2.14 The final layout of Area B

Illus Appendix 2.13 Area B extension BIVWX on 27 May 1981 from the south showing the north passage wall [Film 1981.16.35]

into the green clay under the cairn. These features linked up with similar ones in Area H allowing correlation of layers below the passage and chamber walls. During the last few days of excavation Area Appendices \ 1173

Calanais Survey and Excavation, 1979-88

B was connected to Area H (north of the chamber) by removing part of the baulk between Area BIWX and Area H. to help confirm the stratigraphic relationship between the chambered cairn and Stone 42 of the Ring (Illus Appendix 2.13). In early 1981 the rest of Area BIII was extended eastward by a metre. A week or so from the end of excavation this extension was itself extended further north. Appendix 2.3 Summary description of the archaeology of Area B This summary focuses on low-level interpretation from earliest to most recent. Key maps show the subarea most relevant to the proximate text. Higher level discussion can be found in Chapter 24: Discussion. Several sub-areas of Area B were not cleaned down to the natural clay (Appendix 1). Where subsoil was reached, mainly in BII, BII/III and BIII, the colour of its surface varied depending on what iron compounds had formed. It was mostly light green but included yellow and orange variants and its texture varied from somewhat sandy to stiff fine clay. In places it contained rotten stones. 2.3.1 Phase 5 Early cultivation On Area B the most convincing evidence for cultivation beds at or near the base of the succession came from BV, the southern part of the area near the Ring. There, a fair case could be made for two or three cultivation beds orientated slightly north of east. They would have been about 1.3m apart centre to centre. In Area BIII, the easternmost part of Area B, there did seem to be ghosts of cultivation beds even in the upper soil developments, again running somewhat north of east at about the same spacing as on BV. It seems probable that their direction had influenced that of later features including the East Row which would have sat along an old cultivation bed. There was probably a subsequent phase of flat

cultivation judging mainly by the evidence from BIWX. Elsewhere later ploughing had helped form a palimpsest of ard marks of various dates. This story of cultivation beds followed by flat ploughing is probably basically right even if some of the detail has been misinterpreted. 2.3.2 Phase 6b Ring stones 42 and 43 Ring stones 42 and 43 were probably put up at much the same time as each other. Their pits cut the same ground surface. The most convincing evidence that their erection took place over a period of less than a few years was the coherence of the green clay layer outside the circuit of the Ring, which seemed to reflect spreading of the spoil from both stone-pits without any intervening layers. No signs of construction ramps were detected. 2.3.3 Phase 7b The green clay platform In BIWX the green clay platform had a boundary slot from Stone 43 northward, cur ving round to the north-west. As on Area H there was polygonal cracking in the clay surface. The cracks suggest that the clay lay open to the sun for a while. On the north side of the passage the green clay was cut by a slot and several small pits or scrapes. These succeeded a fairly large pit, central to the passage of the overlying chambered cairn. It had been backfilled mainly with green clay and stones, but its upper fills also included stacks of turf. The sequence on the south side of the passage had been damaged by Victorian activities and was not fully explored. Appendices \ 1174

Calanais Survey and Excavation, 1979-88

The slot under the passage wall joined up with a slot below the chamber wall in Area H and seemed to form part of a carafe-shaped enclosure. It contained charcoal which was ancient at the time the slot was filled, along with tiny fragments of cremated bone. Its pollen, however, was probably roughly contemporary with its filling which took place after 2500 cal BC judging by probable Beaker sherds in the features under the chambered cairn wall. 2.3.4 Phase 7b The East Row The favoured interpretation is that Stone 30 of the East Row was inserted into a layer which formed before most of the dumping of burial material to the south of the chambered cairn, entrance. Attempts to date East Row stone 31 were frustrated by truncation of the evidence. However the preferred interpretation was that the East Row was started before the chambered tomb was built. 2.3.5 Phase 8 The chambered cairn The chambered cairn had a well-built core cairn (or chamber wall) in Area BIVWX and, a sloppily built (or rebuilt) outer cairn. The evidence from Area B supplies no information about the original outer face of the cairn. Probably what we saw on Area B reflected back-piling of stones after the massive kerb-slab in Area BIWX was first erected. The passage width and direction were not easy to establish because all but one of the stones forming the south side of the passage had been disturbed in early modern times. The best estimate is that it was 0.4m to 0.5m wide. It was at least 0.56m tall above the clay forming its floor. There

was no proof that the floor had ever been paved. Its original length is also a matter of interpretation. It could have been as short as 1.8m or as long as 2.5m. Anyway, it was fairly short and narrow. There is no way of telling whether the passage had a low or a high roof, for with the support of the well-built core cairn and the substantial Ring stones it could have been built up to 2m or more. The part of the passage north wall which had shored up the outer cairn was severely damaged. The entire south wall-face also seems to have been destroyed. There is some evidence for disturbance of the passage in latter half of the 1st millennium AD. But possibly more damage to the passage took place fairly soon after 1857. 2.3.6 Phase 9 The Stage 1 and Stage 2 enclosures The first enclosure was represented by stakeholes in BI forming a small structure of uncertain shape. The second enclosure was roughly 3.6 m across internally, defined by an earthen bank with an entrance to the northeast. The ground at the entrance looked much worn; there were depressions in it with signs of several episodes of silting-up. The southern bank (and just possibly the northern bank also) seems to have incorporated a palisade. In places it looked as if the palisade was doubled up but more generally it looked designed to revet the inner face of a turf wall. The previous ground level survived in the area between Ring stones 42 and 43. There the cut or eroded slope of the west side of the enclosed area was ragged and some of the slabs which originally surrounded the base of Stones 42 and 43 had been removed. As described above it is possible that there were two ridges lining an external pathway to the enAppendices \ 1175

Calanais Survey and Excavation, 1979-88

trance in Area BIII. It should be recognised that there is a possibility that the palisade inside the enclosure bank was earlier than the Ring. But there was a mass of circumstantial evidence that it was later than the Ring and that is the interpretation preferred here. 2.3.7 Phases 11 & 12 The third and fourth stage enclosures An area including the earthen enclosure was ploughed and a somewhat larger thirds t a ge w a l l - b a s e d enclosure was built. There were enigmatic hints of wall-bases radiating from this second-phase enclosure; and two of them may have formed the sides of an eastern entranceway. The fourth stage enclosure was also wall-based. Its northern and eastern parts were neatly laid out. It was pretty well at the modern ground surface, and therefore it may have been tidied up at any time after peat was cleared from the area. The preferred interpretation however is that it was not a substantially Victorian invention because if it had been it should have been more complete. In fact its southern part was grossly disturbed and in that area it was impossible to distinguish between stones of the middle enclosure, the latest enclosure and random stones. There was just enough order to the near-surface stones in BV and BIV south of the ditch to give a little support to Somervilleâ&#x20AC;&#x2122;s depiction of a matching southern enclosure. If there was such an enclosure it had been just as damaged as the south side of the known enclosure and on grounds of economy of hypothesis it is better to suppose that the possible southern enclosure stones were instead stones cleared from the area when it was tilled again at some period in the second millennium BC.

2.3.8 Phases 11 & 12 Burials and ground working The ploughing and ground-working west of Stone 30 in BIV/ BV alternated with deposition of burial material including beaker sherds and pieces of impressed and other wares. There seems also to have been at least one crouched burial with an insular beaker. Some of the surface stones hinted at the existence of rough burial enclosures. 2.3.9 Phase 13 The massive kerb slab At some stage the outer face of the cairn between Ring stones 42 and 43 was faced with a massive slab like that still in situ in Area H. The slab in BIWX had been set up more than once. Its original slot survived. Only a relative date could be established: it was (hardly surprisingly) set up later than the slot bounding the green clay platform. The two secondary slots were really almost more like casts of its base than dug slots and it seems that after it fell it was twice leant up against the collapsed side of the cairn. A superficially similar slab was found in Area BIN flat on its face underground between the Ring and the Avenue; it appeared to be in the position that the ice sheet had left it. 2.3.10 Phase 15 to 17 Peat Although basal peat near Area C beyond the pre-excavation end of the East Row was dated to the 1st millennium BC finds of pottery near the Ring suggest that there was disturbance near the mouth of the passage in the second half of the 1st Appendices \ 1176

Calanais Survey and Excavation, 1979-88

millennium AD. So peat growth may have been highly diachronic. 2.3.11 Phase 18 Victorian activities In 1857 the last of the peat was cleared from an area round the standing stones. In the north-east corner of Area B we found the edge of the cleared area. Beyond it peat had been left to a depth of 0.1 to 0.2 m. That boundary can be seen on air photographs and even, with care, traced on the ground. After the peat was cleared Sir Jamesâ&#x20AC;&#x2122; workmen may have tidied up the walls of the cairn and passage. Glass was found under the large innermost basal slab of the south wall of the passage. But if they did repair the chamber and passage, then others undid their work. Indeed, apart from one basal stone of the south side of the passage and the north core wall the surviving passage stones were in disarray. More surprisingly there were a few modern pits inside the enclosure and at least one in the area south of the entrance to the passage. Most Area BI BII

of the modern pits were very small but one, more or less central to the enclosure, was substantial. These were probably not the work of J J Worsaae in 1846, despite his account of test-pitting, because details in his description do not correspond exactly to what was found, and there was no peat in the backfills, which suggests that the work was done after 1857. Appendix 2.4 Areas remaining unexcavated At the end of excavation none of the areas shown on Illus Appendix 2.13 had been completely dug to subsoil level. Most had in part. The main surviving areas of archaeology were in BIVWX, BIV/ BV and adjacent to the medial baulk. That baulk itself was not excavated and preserves a record of the stratigraphy of the enclosure. No systematic record was made of what remained unexcavated. The table below thus depends largely on memory and the recent experience of revising the narrative for Area B.

Comments on remaining archaeology

Basal levels of the bank, and levels including and below turf line 162 near the Ring survive. Small parts of the basal bank survive?

BII/BIII Small parts of the basal bank survive and possibly basal soils all over the area BIII

All gone, to subsoil except perhaps near the bank and the East Row.

BIV/BV

Levels including and below clay 161 in the north survive; bank 838 in the south and the trough to its north in the middle survive, as do parts of the Pit of East Row stone 30.

BIIISX

BIN

BINX BIVWX BIWX Baulk

Quite possibly fairly complex strata remain.

Pretty well everything was removed down to natural subsoil May be basal soil remaining.

Much of the green clay platform and features in it on the north side survives, and layers of about the same period on the south side.

Some of this area was taken down to subsoil on the last day as clay soil 871 was trowelled away, but some may remain near slot 858 which bounded the cairn, along with the green clay platform. The pit for Ring stone 42 was not emptied and a little of soil 871 may remain between BIWX and Area H.

The medial baulk should preserve a complete record of stratigraphy across the enclosure from north to south. Appendices \ 1177

Calanais Survey and Excavation, 1979-88

Appendix 3 Area B enclosure contexts with finds This appendix is referred to in Chapter 7 Area B 7.10.13. The contexts listed here are not all directly related to the enclosure; topsoil and plough soils

are included and some contexts, like 130.1, are only interpreted as related. The Pottery Catalogue provides a reference to the narrative where the pottery is discussed.

Area

Ctxt

Description and interpretation

Pot Cat No Details

Other

BINX

130.1

A band of clay/soil well north of the enclosure 580_581 wall-base and bank

EBA Food Vessel ASH 75

BINX

130.1

A band of clay/soil well north of the enclosure 727_728 wall-base and bank

E/MN Non-corky

149

Apparent pit in the central area of enclosure. Cut by 129 = modern hole. Cuts 180 = round-bottomed pit.

677

?Chalco/EBA ?fine Beaker

165

Area of mixed yellow clay in the enclosure. Possible floor fill.

781_783

Prob E/MN corky

ASH 76

80.137 quartz BI

81.185 rotten sandstone at junction of this and soil 112

BIII

175

Bank composed of iron impregnated clay.

BII/III

130

Bank fill.

BII/III

130

Bank fill.

063

E/MN Corky ?Heb Inc

130

Bank fill.

055

E/MN Heb Inc?

ASH 11

BII/III

130

Bank fill.

768

Indet non-corky

ASH 76

BII

130

Bank fill. Brown fibrous soil with charcoal in its upper part.

809

?E/MN Non-corky

BINX

803

Black layer, turf lines formed before and after an episode of ground working associated with reduction of the bank of the first enclosure.

other

81.52 not in Pot catalogue.

803

Black layer, turf lines formed before and after an episode of ground working associated with reduction of the bank of the first enclosure.

other

81.53 not in Pot catalogue.

Lost?

54 burnt bone 56 not described, seemingly discarded BINX

Appendices \ 1178

Calanais Survey and Excavation, 1979-88

BINX

803

Black layer, turf line. The simplest interpretation is that the two turf lines formed before and after an episode of ground working associated with reduction of the bank of the first enclosure.

167

Greenish/khaki gritty clay at the western edge 297 of the enclosure. Possible floor fill.

E/MN Heb Inc

ASH 26 80.215 charcoal

167

Greenish/khaki gritty clay at the western edge 318 of the enclosure. Possible floor fill.

E/MN Heb Inc

ASH 36

167

Greenish/khaki gritty clay at the western edge 319 of the enclosure. Possible floor fill.

E/MN or poss. Beaker non-corky

ASH 75

167

Greenish/khaki gritty clay at the western edge 566 of the enclosure. Possible floor fill.

EBA/Food Vessel

ASH 75

BIWX

125

Impressions of slabs around Ring stone 42.

EBA Food Vessel ASH 75

BIWX

125

Impressions of slabs around Ring stone 42.

EBA Food Vessel ASH 75

835

Linear feature. - red brown sandy clay - under 836? = khaki strip which was first seen run295_296 ning parallel to 835 and just to its south

E/MN Heb inc

815

One of the shallow features in the entrance way to the first enclosure. Under 183 = soil fill of enclosure and 808 = patch of paving in brown earth

other

81.188 mylonite arrowhead 81.466 mylonite

142

Silts and possibly make-up in an erosion hol771 low in the entrance-way of the first enclosure.

Chalco/EBA Fine Beaker

80.134 hornfels

183

Superficial soil fill of the central area of the enclosure.

other

81.505 mylonite short end scraper

BIII

142

The largest of the shallow features within the enclosure

829_831

Indet non-corky

BIII

142

The largest of the shallow features within the enclosure

832

Indet

107

Topsoil above uppermost dark brown fills in the enclosure

354

Chalco Early International Beaker ASH 39

In 107 but inside 120

120

Uppermost dark brown fills in the central area 563 of enclosure.

EBA Food Vessel

80.44 glass, modern; 80.51 chert

810_812

?E/MN non-corky

81.702 burnt bone

BIII 81.601 quartz x5 BIII 80.135 quartz arrow

Appendices \ 1179

Calanais Survey and Excavation, 1979-88

107

Uppermost dark brown fills in the enclosure

726

Prob E/MN Corky

120

Uppermost dark brown fills in the enclosure

360

?E/MN ‘Non-corky’

120

Uppermost dark brown fills in the enclosure

357

Chalco Early International Beaker ASH 39

120

Uppermost dark brown fills in the enclosure

358

Chalco/EBA Beaker

120

Uppermost dark brown fills in the enclosure

361

Chalco/EBA Beaker

120

Uppermost dark brown fills in the enclosure

359

Prob Chalco/ EBA prob Beaker

120

Uppermost dark brown fills in the enclosure

239-42

E/MN ‘Non-corky’

Summary

bank — Four sherds, all from Context 130 in

the northern bank sector: E/MN Corky ?Heb Inc, E/MN Heb Inc?, Indet non-corky and ?E/MN Non-corky. These should reflect the contents of the turfs and soil used to build the first bank. interior fills — These were clays and included three E/MN Hebridean incised sherds, two possible beaker sherds, and, in the locally highest clay patch 167, a sherd of Food vessel ASH 75. A lower patch 165 included a probably E/MN corky sherd. A near-central pit (149) contained a possible fine beaker sherd.

entrance silts — Three sherds, one fine

beaker, and Indet and Indet non-corky, from the largest area of silting 142. The silts also contained a mylonite and a quartz arrowhead and stray bits of mylonite and quartz.

In 107 but inside 120.

E/MN sherds (one of somewhat ambiguous affiliations) and a sherd of Food Vessel ASH 75. They also contained find 81.54 burnt bone in a turf line 803.

upper fills — The upper fills of the enclosure

reflect ploughing with the possibility of movement of soil and sherds from elsewhere in the area. They included one probably E?MN corky sherd and two non-corky ones (one of somewhat ambiguous affiliations), two bits of Early International Beaker ASH 39, three Beaker sherds (one of somewhat ambiguous affiliations) and, from an uppermost fill, a piece of Food Vessel ASH 75. They also contained two finds of glass, two pieces of chert and a fragment of a mylonite short end scraper (81.505). The lack of Hebridean Incised sherds is notable. This appendix is referred to in Chapter 7.12.

miscellaneous — The slab impressions on the ledge by the Ring were filled with material interpreted as washed down from the cairn. They included two sherds from the Food vessel ASH 75. reduction — Contexts outside the northern sector of the bank, interpreted as reflecting reduction of the bank by ploughing, included two

1180

Appendices \ 1180

Calanais Survey and Excavation, 1979-88

Appendix 4 Green clays 810 and 812 and East Row Stone 30 The grey-green sandy clays in BIV/BV: layers 810, 810.1, 812 and 812.1 were sandwiched between plough soil 141 and thin black worked soil 160 and its cognates 192 etc. The crucial interpretational problem is the nature and date of the green clay 812 which was later than the pit for East Row stone 30. The closely related layer 810 should also be taken into account. Until careful reanalysis of the labelling of potsherds in the Finds Book and on the finds bags after the latter had been recorded by Alison Sheridan (Chapter 18), it was not apparent that what is now labelled 810.1 went with 812 in BI, northeast of the northern part of BIV/BV and that 812.1 went with 810 in the southern part of BV. 810 (and now 810.1) are used for seemingly definitely disturbed layers and 812 (and now 812.1). are used for seemingly relatively undisturbed ones, as follows. The finds in them are listed in the main text (Chapter 7.12). —— 810 Disturbed green sandy clay in the southern part of BV: although on the bags the area was identified merely as BV other records make it clear that this was in the southern part of the trench. This sandy green clay 810, in BV towards the south end of BIVS/BV trench, was thin and it was certainly disturbed by later intrusions. For instance a small pit or root hole 834 with glass was found below 810. —— 810.1 Disturbed green sandy clay in BI: this thin green sandy clay lay north of East Row 30 and south of the modern ditch in an area where layers ran together. 810.1 was disturbed, and it was close to the pre-peat topsoil and thus cannot be regarded as well-sealed. —— 812 Green sandy clay in BV: although on the bags the area was identified merely as BV other records make it clear that this was in the northern part of the trench. —— 812.1 Green sandy clay in BVSX: bag labelling and the Finds Book show that despite being given the Area label 821 these finds were in the southern part of the trench.

One idea to explain the existence of 810 and 810 can, I think, be rejected. Given the evidence for dynamic post-depositional soil processes in the northern part of BV, and the strong iron pans in layers 892 ‘soil like 141’ under 812 and 810.1, it might be argued that the light colour of clay 812 reflected leaching-out of organic and iron-rich elements. But the green iron-rich compounds in 810/812 should have been oxidised during such processes. I therefore think it can be excluded as a possible explanation. That said, in any future excavation nearby the layers should be examined carefully for a better understanding of local soil developments. These green clay layers lacked sherds from ASH 42, the insular Beaker probably associable with a burial redeposited during the formation of plough soil 141; nor did they include any other specifically late types of beaker. The only two Beaker sherds which were not diagnostically early or anonymous (Cat 538 and 556 Illus 18.21) had stab decoration, not very common at Calanais. Thus, as with the green clay in BVS, the assemblage may have been an early Beaker one apart from the residual Neolithic sherds, the presumably intrusive Food Vessel sherd (Cat 567) in disturbed clay 810.1 in Area BI to the north of East Row stone 30 and a possibly intrusive possibly 6th century AD sherd (Cat 756_757) in the southern part of the trench. Hypotheses about origins and dating of 810 and 812 The clays can be assigned to any of several phases between construction of the green clay platform under the cairn and digging of the stone-holes for the stones of the East Row. a. Phase 7b to 7c with subsequent disturbance perhaps as late as Phase 10a or 10b: The clays may have come from disturbance of the fills of graves dug for people using early International and AOC Beakers. b. Phase 7c: These sandy green clays could be contemporary with the manipulation of green clay Appendices \ 1181

Calanais Survey and Excavation, 1979-88

related to closing down of the green clay platform. Closure of the latter probably took place a period when Beakers were still exceedingly rare in the Western Isles. In this hypothesis it could be argued that the southern half of the structures on the green clay platform had been used for deposition of Beakers but the northern half was not used for deposition. When the timber buildings were decommissioned the green clay and potsherds in the southern half of the building were scraped up and dumped outside the Ring. c. Phase 8 (use of chambered cairn): The layers could have been taken from the chamber of the cairn during ‘over-vigorous’ cleaning. The chamber floor area was roughly 3.5 to 4 sq m, and the passage floor area was around 2.5 sq m, so if a 0.1 m layer of subsoil were taken from it that could account for the amount of green clay in layers 810 and 812 (plus for that matter 842). The ‘contaminating’ presence of the potsherds in (disturbed) subsoil of the chamber might explain why it was removed. d. Phase 9a: Alternatively the interior of the enclosure a few metres to the north may have supplied

green clay when it was dug down through earlier layers, possibly after the first enclosure was built and while its floor was mainly green clay. Given the absence of beaker sherds from 130 where it formed part of the enclosure bank (here I am excluding the similar material in a depression in BIN relabelled 130.1 during post-excavation), the early embanked enclosure would, in this hypothesis, have been in existence for some time when this levelling down was carried out, time during which many disturbances introducing beaker material to green floor layers there could have occurred. e. Phase 10b: Another source could have been the stone holes of the East Row, but spread at a somewhat later date after activities not represented in the areas we dug had led to incorporation of potsherds in the clay. Yet another possible source is the (admittedly very poorly understood because seen only in section) pit or linear ditch underlying the pit for Stone 30. The problem with these ideas is that it is not clear why Beaker sherds should be incorporated unless digging the stone-holes disturbed earlier graves.

Appendices \ 1182

Calanais Survey and Excavation, 1979-88

Appendix 5 The sequence in Area D and E This summary runs from early to late. Appendix 5.1 The early ditch The earliest large feature discovered was the ditch 921 in Area DI which was about 0.8 to 1m wide and no more than 0.15 m deep. It appeared to be part of the shallow encircling ditch of an enclosure. If it was circular it is impossible to estimate its diameter at all closely, particularly because the east end of the ditch was overlain by the trough separating cultivation beds 1 and 2. The end of the enclosure ditch near the central monolith was not investigated fully; it appeared to be much disturbed. It is therefore not clear whether there was a terminal at that point although it did seem to end. To pre-empt the discussion in Chapter 24 where all of the Areas are considered together, there were no features elsewhere which might have been part of the same ditch. The ditch must date to before the cultivation beds, and the best estimate for the start of the latter is sometime between 2980 and 2510 cal BC. Appendix 5.2 Cultivation beds The area contained shallow cultivation beds with a wavelength of on average about 1.5 m. The ambiguity of some of the evidence demands that some detail should be taken with a pinch of salt. But Bed 4 in Area DII, DIII and DIV, and Beds 1 and 2 in DI did seem to be well-attested and there was a little evidence for Bed 3. Again, I shall discuss cultivation beds in Chapter 24, pulling together the evidence from all of the Areas.

Appendix 5.3 Pre-Ring old ground surfaces and related layers A persistent turf line, 334-lower to the north with 365-lower to the south, covered much of the area of DI. In places another old turf line (also labelled 334 and 365) formed over it, above spreads of clay or soil. Broadly speaking although the turf lines were in place before the cairn was built, it is impossible to be sure how they related to the erection of the central monolith or that of the Ring. In discussing the evidence I suggested that the earlier old ground surface was that into which the monolith and Ring were set and the younger one had formed on top of clay spreads derived from spoil from Ring pits. But this generalisation must be treated with caution. Turf tended to form on any exposed soil or clay at Calanais. A new turf line could merge with an old one. Breaks in a turf line could heal over and become unrecognisable. Appendix 5.4 The central monolith There was no direct evidence from Area D to suggest that the monolith was earlier than the Ring. It was definitely earlier than the chambered cairn and although there was some ambiguity in the evidence the monolith was probably also earlier than a light timber structure built on a platform of green clay to its east. Although the evidence suggests the timber structure was built on an area levelled up using clay from the Ring stone pits it does not strictly speaking prove that the Ring was later than the central monolith. Appendix 5.5 The Ring There was no unambiguous evidence for the date of setting-up of the Stone 47 of the Ring on Appendices \ 1183

Calanais Survey and Excavation, 1979-88

Area DV, because of later ground-working and soil processes. The evidence from round the base of the stone could be interpreted as suggesting a ramp inside the Ring. It was too tenuous to provide much interpretative weight. Further excavation would resolve the point. It would be fascinating if there proved to be some reason to pull the stone up from the enclosed area, rather than from outside.

lack of direct evidence that it was dug to support a standing stone, and given its similarity to the pit under the passage in Area BIWX which almost certainly did not support a standing stone, it should instead be compared to the pits found on sites discussed in Chapter 24. Much of Pit 917 remains unexcavated; further excavation would allow testing of the idea that it held a stone or a timber. More cogently, perhaps, its date could be ascertained. The possibility that it was related to the rows and part of the avenue is discussed in Chapter 24.

Appendix 5.6 The platform and slots under the cairn

Appendix 5.8 The chambered cairn

There were remnants of a clay platform under the cairn. The slots on it and immediately under the cairn were slight and if they supported structures they must have been lightly built and short-lived. In the interpretation favoured here the slots were contemporary with one another even though one cut an old ground surface which seemed to have grown over the other. It is very easy for ground surfaces to be diachronic. Maybe the inner part of the structure was more durable than the outer, or it was more significant and was kept in good repair after the timbers in the outer slot had decayed or been removed. In Chapter 24 the slots are considered along with similar ones found on Areas H and B. From pollen analysis it seems probable that the turf outside the cairn continued to grow long after the cairn was built. Appendix 5.7 Pit 917 Pit 917 was the best candidate amongst the several pits recorded in Area DI for a stonehole unconnected to known standing stones, but given the

The cairn was shown to include an inner core or chamber wall and an outer cladding. The original chamber wall included a deeply curved rear wall rising up the mound round the base of the central monolith, rather than the square end shown by Sharbau shortly after it had been cleared of peat. There was a neat outer kerb, probably of laid slabs, which also rose up over the mound of clay at the base of the monolith. It seems to have been polygonal rather than rounded. All but its basal level was robbed in prehistory. Appendix 5.9 Burials and ploughing In Area D subsequent activity focused on the area south of the cairn. It was used for deposition of burial-related or domestic material in low piles and small shallow pits. It seems fairly likely that deposition alternated with ground-working over a considerable period, with deposition perhaps ending some time in the first few centuries of the second millennium BC. Ploughing, perhaps for cultivation, continued long after that. Its complex chronology is discussed in Chapter 24.

Appendices \ 1184

Calanais Survey and Excavation, 1979-88

Appendix 5.10 The rubble platform

Appendix 5.12 Unexcavated archaeology

The rubble platform in the southeast part of the area remains an enigma. Given the amount of Victorian activity in the area it is conceivable that it started life as a depot for cairn stones during trial trenching prior to cairn-rebuilding. It is also conceivable that it reflects more general stone clearance. But we could not entirely exclude the possibility that it was built before peat covered the place.

In Area D only DI Chamber and Area DIII were completely excavated. The cultivation bed levels in DII were very nearly bottomed. They were only partly removed in Area DI and important remains may survive including, perhaps, features related to the early ditch 921.

Appendix 5.11 The Victorian cairn and Lady Matheson’s Path The amount of disturbance during the post-peat period was a surprise. It is not at all clear whether the square back end of the cairn was a Victorian addition or a change made before peat started to grow continuously in the first millennium BC. Probably in a second phase of Victorian activity a sondage was cut into the cairn under where Lady Matheson’s Path was planned to be and then the cairn was trenched round almost to the central monolith to trace the line of the original kerb. A new kerb was built and earth and stones tossed back behind it. The material thrown back into the initial sondage was revetted by a neat slab-built wall and Lady Matheson’s Path was built over it. Area D was landscaped by addition of earth and turf (perhaps with some stone clearance) after which only minor maintenance took place until our excavations.

Area E summary description Most of Area E contained a plough soil but to the southwest there was a rock outcrop close to the surface. In soil above the bedrock was a small cluster of Hebridean incised wares sherds. Probably later pottery suggests subsequent cultivation and the sherds may have been collected during cultivation and put to one side. Taken in conjunction with the evidence from antiquarian accounts (Ashmore 1984; Ponting and Ponting 1979 appendix C), the archaeological evidence makes it improbable that Stone 35 was erect in any part of Area E before peat covered the place. There was no convincing evidence for any other standing stones having been present. The bleaching of Stone 35 was different on the top and bottom halves. That suggests that one half lay below the peat which covered the site from the first millennium BC to the early modern period and the other did not. It is conceivable that this latter piece was one of the ‘extra’ stones recorded on early plans (Chapter 3: Previous Studies). The history of use of the stone may be quite complex. It is possible, for instance, that it was originally used as a lintel over the chamber or the passage; perhaps it was broken in antiquity so that its parts could be used as kerb stones. But that is pure speculation.

Appendices \ 1185

Calanais Survey and Excavation, 1979-88

Appendix 6 Area F Summary description After the end of the glacial period the ground in which Area F was set out may all have been as undulating as the area west of the Avenue. Human activity was dated by charcoal spreads including fragments of hazel charcoal dating to some time between 3640 and 3380 cal BC and between 3330 and 2890 cal BC. The pottery assemblage consisted of small undiagnostic non-corky fragments, apart from one possibly early neolithic sherd, two rim sherds from Hebridean Incised Ware pots, one in a superficial deposit but the other in clay filling a pre-cultivation depression, and a sherd with harsh very gritty fabric found in the upper mineral soil on FVIII. In contrast to the areas near the Ring, there were no sherds from Grooved Ware, Beakers or Food Vessels. The area inside the Avenue must have been levelled purposefully, for otherwise turf lines would not have survived in the depressions. This activity led to the digging up of soil containing the earlier piece of 4th millennium BC charcoal. The later piece of charcoal provides a terminus post quem for the cultivation which allows it to be as early as the latter part of CaN-2c at Callanish Leobag, the earliest period of cereal-growing in the area. But it may have started much later and cultivation may have continued even until initiation of peat growth. Some of this pre-cultivation levelling-up, whatever the date of the rest, appears to have been roughly contemporary with erection of the southernmost stones of the western side of the Avenue, because a thin turf line covered both the fill material (green clay containing a piece of Hebridean incised ware) and the edge of the mound around Avenue stone 12. However that mound was brown clay rather than the green clay forming the mound round Avenue stone 11 and there is a possibility of more complicated stratigraphy near Stone 12 than appeared in Area F. There was a slight suggestion of a curving linear depression in the area enclosed by the Avenue stones, earlier than the local levelling-up. Too little

of it was seen to be sure that it was not in fact part of an amorphous depression. The other seemingly prehistoric features found in Area F may well relate to agriculture or grazing. Stones were found in piles which presumably reflect clearance either of cultivated plots or pasture. In general they lay on top of a cultivated soil. However, similar piles of stones lay in peat. When MacCulloch visited before 1819 stones 11 and 12 were not visible. Shortly before 1857 all three stones were shown largely peat-free on Callender’s ‘birds-eye view’. By the end of 1857 all of it been cleared except for a thin skim of amorphous peat. Fibrous peat subsequently formed over parts of the area. There seem to have been attempts to fill up waterlogged areas with flat stones (under some of which glass was found). Area H Summary description Area H was not cleaned down to natural subsoil and some 0.1m of deposits remain unexcavated in part of the area. The excavated sequence started with creation of cultivation beds with a wavelength of about 1.4 to 1.6m, although only one trough was well-preserved. Turf-line 766 formed. A thin layer of soil 756 / 777 overlay this, following the contours of the cultivation beds and best preserved trough. The preferred interpretation is that Ring stone 42 was erected through turf line 766 and that soil 756/777 was spoil from the pit for the Ring stone. Alternative interpretations are possible. It is conceivable that the Ring stone was erected after turf line 751 / 758 had formed. The stratigraphy suggests that this is less likely that the interpretation offered above. But if it was, then some of the green clay found above turf line 751 / 758 might have come from the nearby stone pits. Pollen evidence suggests that if that was so then the green clay would have had to be stored somewhere else before being spread under the subsequent area of the cairn. That does not seem very likely. Appendices \ 1186

Calanais Survey and Excavation, 1979-88

Soil 756 / 777 was ground-worked, which seems to have destroyed the top of slot 795/1001, if it was originally cut from the top of the soil. Another slot may have existed to its south although the evidence consisted only of a shallow feature in section. If it had been a slot it would have been removed by the chamber of the chambered cairn. Turf line 751/758 began to grow. After a time long enough for it to have become well-developed imported green clay was spread out over the area. At some time before that stones had been piled in a mound round the base of the standing stone and the green clay lapped onto this mound. On this green clay platform a light structure or successive light structures were built. If it was a single building it will have consisted of an inner and an outer timber wall or screen. But an alternative interpretation is that a light timber circle about the same diameter as the subsequent cairn was superseded by a smaller structure. The inner gulley defining the latter enclosed a space similar to that subsequently occupied by the chamber of the chambered cairn and passage. Once the structure had been abandoned the inner slot was filled with clay containing by then ancient charcoal and above that organic material which had a peaty appearance. The fills included two probable domestic Beaker potsherds. Above a patchy turf line covering Area H a sturdy chamber wall and an outer cairn were built. The outer cairn was probably faced with a wall of laid slabs. The two pieces of charcoal from the cairn which were radiocarbon-dated seem likely to have been residual. An early International beaker sherd was incorporated in the cairn; it, supported by the probable domestic Beaker sherds from the underlying slot, shows that the cairn was built after c. 2500 BC. The details of the find spot of the Beaker sherd in the cairn makeup show that it was far more likely to have been part of the original cairn than to have put there during the later disturbances; and that it is also quite unlikely to have fallen down through layers of stones from a secondary filling level. But the possibility that it was intrusive cannot be absolutely excluded. There are enough demonstrable errors in the record that it is in the abstract conceivable that another one has affected

this detail. But I have no reason to think that it did. A balanced interpretation, while giving primacy to the observed stratification, should acknowledge the currently perceived gap between the latest likely date for building a chambered cairn and the earliest likely date for International Beakers. The scientific evidence for the date at which building of chambered cairns ceased is very unsatisfactory (Ashmore 2004a). But there are no (other) chambered cairns in Britain with beakers in primary building contexts. In current models none are thought to have been built after c. 2600/2500 BC, and many scholars would prefer an earlier date. There are also uncertainties about what dates should be assigned to the earliest International Beakers. Some see Beakers appearing widely on the borders of the Atlantic around 2600 BC (Cottiaux et al 2007, 154) and reaching the Paris Basin, for instance, perhaps c. 2600-2500 BC (Cottiaux et al 2007, 157) or only after c. 2500 BC (ibid 155). Sheridan suggests that the first Beakers probably reached Britain after 2400 cal BC (Chapter 18, this volume). That is soundly based on the many available radiocarbon dates associated with Beakers in Britain; but because all the reliable dates are from burials the possibility that the earliest Beakers in the Western Isles were used in domestic contexts at around the date ascribed to early Beakers in other Atlantic-bordering areas cannot be excluded. It has been suggested that some small tombs were established in coastal areas in the late 3rd and the 2nd millennium BC ( Jones & Thomas 292). Their model has more of the status of a hypothesis than a theory, because of the acknowledged lack of dating evidence for the primary construction and use of the cairns, but it provides a framework for dating the Calanais tomb to the second half of the 3rd millennium BC. Its is discussed in further detail in Chapter 24: Discussion and Conclusions. My own view is that the difficulties currently perceived in an apparent gap between the last building of chambered cairns and the first appearance of International beakers in Atlantic areas in Britain are quite as likely to be resolved by new data as they are to be confirmed. The preferred interpretation purely from the Appendices \ 1187

Calanais Survey and Excavation, 1979-88

stratification and nature of context 736 at Calanais is that the International Beaker sherd reached the position in which it was found when the cairn was originally built, with the proviso that the stratigraphic evidence is not strong enough to entirely preclude an origin for the sherd in secondary activities. The original position of the north-east chamber orthostat and its form suggest that the wrong stone was in position in 1980. The recumbent slab in the chamber is a better candidate for the original orthostat. The outer part of the cairn was robbed of its stones, including its entire kerb apart from one stone. All but the lowest few layers of the chamber wall inner face may have been removed at this time along with any remaining chamber roofing. Eventually a slab large enough to have been a roofing slab was erected to form a new outer facing to the cairn; two smaller slab-like boulders continued the line of the kerb to the east. Stones and soil were thrown back behind the slabs building up the cairn to an even height. The material capping the cairn included charred cereal grains dating to between c. 1600 and 1400 cal BC. It also included many sherds of a fine Food Vessel (ASH 75) with ornate

impressed decoration. Because this deposit was of material scraped up from somewhere nearby, and because it was unsealed, the act of capping the cairn, the cereal grains and the potsherds need not have been of the same date as each other, and the barley and potsherds do not provide a terminus post quem for the capping. Some rough repairs to the inner face of the chamber may have taken place at this time, although the evidence suggested that the repairs visible when we excavated the chamber occurred in the Victorian period after peat had been cleared from the area in and around the Ring in AD 1857. The chamber was emptied in 1857. At the same time a pit may have been dug to investigate the footing of the external kerb slab. The only surviving account of the clearance refers to black deposits containing pieces of cremated human bone. Alternative interpretations for many other details could be promulgated, and in earlier versions of this narrative the text was overloaded with provisos. In this version those alternatives have been implied or alluded to in passing rather than described in detail because they do not affect interpretation of the recognisable structures built on Area H.

Appendices \ 1188

Calanais Survey and Excavation, 1979-88

Appendix 7 Modelling connections

Illus Appendix 7.1 Increased connectivity through addition of a few long distance links. In discussing regionality I said that one good reason for spreading comparisons over long distances is that local changes may sometimes have depended on changes elsewhere. The reason for failure of a previously successful social nexus may be loss of links between other areas rather than any events at the place itself (taking ‘success’ here in the limited sense of creation and survival of a rich material culture). If social networks are crucial to an understanding of how change took place it is important to grasp their common characteristics. Network behaviour is in some ways independent of the particular entities joined up in it, provided it is possible to establish and to lose both short and long distance two-way connections between them. Both in a formal sense and in practice the elements of a large, sparsely connected social network

can be ‘close to each other’ (Watts 1998, 5, 40). In the ‘egalitarian’ networks studied by Watts, a small number of long distance links can create a high degree of overall connectivity (Illus 24.1), which translates for instance into people frequently travelling long distances, meeting the inhabitants of distant areas, influencing them, being influenced by them and returning to influence their own neighbours. Other studies emphasise ‘aristocratic’ networks. In these some nodes start to attract more connections than others. Once they have more connections they seem (irrespective again of the particular entities making up the connection-set) to become unusually attractive to yet more (Buchanan 2002, 111-20). This too appears to be a universal behaviour where there are simple low cost two-way links. In some networks, however, over time the less successful nodes catch up with the more successful ones, returning the system to a near-egalitarian state. Thus the success or failure of the people at one site was inextricably linked to the success or failure of those in other places. In other words, it was probably the existence of low-cost two-way links between productive groups of people at geographically various locations which led to success. I suspect that this model applies to Calanais. Formal analysis demands identification of criteria for measuring success, the strength and number of links and other features. By way of example, a study of maritime interactions in the Aegean Bronze Age incorporated social as well as geographical factors (Knappett et al 2008). The analysis was based on 34 sites spread around the Aegean, for each of which a fixed intrinsic carrying capacity, reflecting its local resources, was assessed. The relative importance of each site was a variable reported with other output figures of the model. The social size of the site was represented by multiplying these two figures together. The links between sites also had physical and relational values, on the one hand the geographical Appendices \ 1189

Calanais Survey and Excavation, 1979-88

distance and on the other the fraction of effort which the people at one site put into maintaining a link with another site. Knappett and his colleagues calculated a cost-benefit social potential for each possible configuration of site and link variables, based on the costs of supporting a local population, the benefits of exploiting local resources and the costs and benefits of maintaining links. The benefits of one site’s maintaining a connection with another were based on the product of their populations and a number which reduced as the distance between sites increased. Thus Knappett and his colleagues encapsulated a proposition that the people living in large sites preferentially chose to interact with one another. The network was approximately optimised for many combinations of the input variables to minimise the overall ‘cost-benefit’ social potential, and the implications of the output figures were assessed. One problem identified by the authors with their own approach is that the choices made in setting up the model (which for instance in their case emphasised the role of socially large sites) affected the outcome. But their model was robust, producing similar outputs from many starting points of the input parameters. For the set of 34 Aegean Bronze Age sites it produced both historically plausible results and some less obvious ones which seem to fit the historical record. It showed the importance of some relatively small sites; for instance they suggested that the eruption of Thera, which destroyed an important link, permanently affected sites on Crete. They emphasised that their model was a tool for thinking rather than a mechanistic problem-solver and discussed possible improvements such as using transport times rather than geographical distances as input factors. In a subsequent study of the same region they concluded that the Thera event did not destroy the network; it was sufficiently resilient to

sustain overall levels of activity; they attributed the changes following the eruption to an increase in exchange costs (Knappett et al 2011, 1019-1020). Rather than transformations at Calanais itself it was maybe changes in ideas amongst the other peoples of the lands bordering the Irish Sea that changed perceptions of cost-benefits of connections with Calanais. But it is hard to judge whether the models described by Knappett and his colleagues (along with others like them) could provide a realistic strategic research agenda for the late 4th and 3rd millennia BC in the north and west of the Hiberno-British archipelago. My impression is that their approach would require considerable modification, for instance because there is far too little data to measure the carrying capacity of relevant sites in Britain and Ireland. It might be feasible to develop some measure of their likely social importance and the likely strengths of their links with others; but if modelling is to be worthwhile much preliminary work must be done on examining the date from excavated sites in the context of their local landscapes and interpreting existing material in museum collections, as recommended by Clarke (2004). Another problem in devising such a research framework will be agreeing what the evidence for connections between places means. For instance, one of the interesting facets of the model proposed by Sheridan (2009) and Pailler and Sheridan (2009) to explain the occurrence of possibly late 5th millennium Breton-style artefacts in Western Scotland is that the contacts between N W France and the countries bordering the Irish Channel have not been shown to have been two-way. The circumstances which they envisaged would thus not include networks with the ‘universal’ behaviour described above. In other words, there need not have been an enduring high level of social connectivity between people in Brittany and those on the coasts of the Irish channel.

Appendices \ 1190

Calanais Survey and Excavation, 1979-88

Appendix 8 Phasing Table Table 24.1Â Phasing of early features north of the Ring from about 6500 to about 3000 BC Phase Date BC

Date source Events

Pollen

Started at some date between c. 6600 and c. Woodland clearance 6200 cal BC. in CaN-2a and 2b at Calanais Leobag.

C14

Foragers in Avenue Date between 5305 and 5215 cal BC area?

C14

Foragers or farmers Date between 3720 and 3530 cal BC in Avenue area?

6500

6000 5500 1 5000

Dating and comments

4500 4000 2a 2b 3500 3a

C14 and pollen

Date between 3640 and 3380 cal BC. Soil Farmers in Avenue disturbances. Early cereal cultivation started area and near Calanear Calanais Leobag at some date between nais Leobag.? 3875 and 3605 cal BC

Pollen

Cultivation finished Date between 3490 and 3020 cal BC.

C14

Early farming in Avenue area?

Original deposition of charcoal and early to middle Neolithic pottery in settlements nearby.

Charcoal date between 3300 and 2890 cal BC. Soil disturbances. Charcoal dated to between 3330 and 2890 cal BC, in round terms between c. 3300 and 3000 BC. Imported with soil to Calanais many generations later in Phase 7.

Table 24.2aÂ Phasing of features in and near the Ring between about 3000 and about 2250 BC Phase 4

Date BC 3000

Date source Events/processes

Pollen

Early ditch

Comments

Before cultivation beds.

Appendices \ 1191

Calanais Survey and Excavation, 1979-88

Pollen

Cultivation beds and subse- At beginning of CaN-3a, start quent flatter cultivation between 2983 and 2510 cal BC.

Turf formation

2950 6a?

Stratigraphy Monolith erection? 2900

6a, 6b?

Stratigraphy 2850

6b?

Stratigraphy, pollen

C14

Interrupted slot? Other use of Ring?

C14

Interrupted slot? Other use of Ring? Continued turf formation?

2650

2600

A piece of birch charcoal from a charcoal spread in pre-Ring soil 871 produced a radiocarbon age (AA-24959 4140+/-45 BP) corresponding to a date between 2880 and 2580 cal BC. That suggests that the Ring was set up after 2880 cal BC.

The new turf is that above green Ring erection? Continued clay spoil from Ring pits but turf formation? Use of Ring where not covered by clay the old area? Interrupted slot? turf continued to grow

2700

6c?

Monolith erection? Ring erection?

Ring erection? Turf formation? Use of Ring area? Interrupted slot?

2750 6c?

Near beginning of CaN-3a. 3000/2850 BC After 5b turf growth, and interpreted (but not demonstrated) as before Ring.

Stratigraphy, analogies

2800 6b?

TPQ some date between 2890 and 2620 cal BC from charcoal.

Appendices \ 1192

Calanais Survey and Excavation, 1979-88

Clay platform and light timber structures built.

Possibly 2450/2350 BC?

Clay and soil imported with Phase 3 artefacts and charcoal.

Layers on Areas D and H, with Phase 3 charcoal and also potsherds. Preferred date shortly before clay platform built.

Indirect C14 and interpretation

Rows start now?

In general undated but East Row Stone 30 probably erected at some date between 2560 and 1690 cal BC; earlier than green clays with very early Beakers.

Artefacts

Slot fills include probable fine Closure of green clay plat- and â&#x20AC;&#x2DC;domesticâ&#x20AC;&#x2122; Beaker sherds. form structure? Dumping of Green clays 810 and 812 with very early Beakers and no obviclays 810 and 812 in BV? ous later ones.

7a?

Stratigraphy

7b?

Clay and soil imported Speculation with Phase 3 artefacts and charcoal.

2550 2500

Interpretation

7b? 2450 2400 7c

7d 2350 8

Artefacts

Enclosure Stage 1, stakeInterprehole defined after disuse of tation / timber structure and after stratigraphy cairn construction

9b 2300 10a

Chambered cairn building

InterpreEmbanked enclosure Stage tation / 2 earthen bank building,? stratigraphy

C14

Possibly as early as this, or even part of a long term process starting earlier.

After 2500 cal BC because the primary cairn and immediately pre-cairn contexts included an International Beaker sherd and at least one fine Beaker sherd. Starts around 2400/ 2300 BC?

Starts around 2400/ 2300 BC?

Pine charcoal dated between 2490 and 2290 cal BC could beGround working and burials long here, or Pomoideae charcoal dated between 2120 and 1770 cal BC Appendices \ 1193

Calanais Survey and Excavation, 1979-88

Table 24.2bÂ Phasing of features at Calanais from about 2250 BC Phase Date

Date source

Events/processes

Comments

10c

Starts around 2400/2300 BC?

Ground working and burials

See 10a

2250 BC

10d

Avenue starts now?

No specific dating evidence. Perhaps building on from a North Row? Stones set up over several generations? Birch charcoal dated to between 2030 and 1750 cal BC and Pomoideae charcoal dated between 2120 and 1770 cal BC, from an ard mark north of the enclosure, could belong here.

2000 BC 10e

C14

Ground working including ploughing

11a

Artefacts; Indirect C14 and interpretation.

Wall-based enclosure Stage 3 building

Interpretation

Ground working including ploughing

Analogy

Wall-based enclosure Stage 4 building

11b

12a

Between 2100 and 1750 BC? After deposition of barbed and tanged arrowheads; also a possible terminus post quem between 2040 and 1770 cal BC.

Later period of plough soil 141 belongs in pollen zone CaN-3c which started between 1880 and 1520; and was earlier than 139, so dates to between 1880 and 1690 cal BC. Undated; possibly in use at some date between 2150 and 1750 BC, judging by an analogy with Ardnave, Islay.

12b

Interpretation

Ground working including cultivation

Interpretation

Kerb erection

After 2040 and perhaps after 1525 cal BC; may post-date ruin of fourth stage enclosure?

Stratigraphy

Cultivation

C14 and pollen

Encroachment of peat

Undated except cannot be later than steady local peat growth.

1500 BC

1000 BC 15

Growth of peat

Includes layer 139, by hypothesis, dated to between 1940 and 1690 cal BC by a piece of heather charcoal.

Between 920 and 400 cal BC near the east end of the East Row Continued use of area of the stone setting, dissuading local peat growth?

Appendices \ 1194

Calanais Survey and Excavation, 1979-88

500 BC 15

If Phase 16 real.

Growth of peat

Continued use of area of the stone setting, dissuading local peat growth?

If Phase 16 real.

Growth of peat

Continued use of area of the stone setting, dissuading local peat growth?

If Phase 16 real.

Disturbance and erection of slab in passage? Growth of peat

Possibly in the second half of the 1st millennium AD judging partly by occurrence of late Iron Age pottery. But this activity may be post-1857.

BC/AD 15 AD 500 16

17a

AD 1000

Peat growth

AD 1500 Documentary

Cutting and post-medieval peat clearance

North end of setting cleared before mid 19th century. MacCulloch surves before 1819. Worsaae visits 1846.

Documentary

1857 clearance; 19th excavations and disturbances

Pit digging at earliest in AD 1846 but probably after 1857

Documentary

Modern conservation

From 1882 to 1980

18c

Documentary

Archaeological excavation

18d

Documentary

Post-excavation conservation

Resisitivity survey 1979, excavation 1980 and 1981; intrusive sampling etc 1982; excavation 1986.

17b AD 1850 18a AD 1880 18b AD 1980

Improvement of drainage and returfing and path-building.

Appendices \ 1195

Calanais Survey and Excavation, 1979-88

Appendix 9 Scottish Ritualised Landscapes It is a moot point whether the large chambered cairn clusters of Caithness, at Loch Calder, Loch of Yarrows, Sordale Hill and Dunbeath (Davidson & Henshall 1991, Fig 5) should be regarded as ritualised landscapes in the sense defined above. If current ideas about the date of these complex and multi-period monuments are correct they are centuries earlier than less ambiguously ritualised landscapes. The most concentrated cluster is at Loch of Yarrows with an average of slightly more than 1 cairn per square kilometre but the Sordale Hill and Loch Calder clusters are nearly as dense (ibid, 17; Fig 6). Davidson and Henshall suggested that in general the cairns might lie close to settlements. An alternative interpretation is that the clusters were in an area thought to be particularly suitable for funerary monuments by several distant communities. Particularly if this latter speculation is right, the areas will have had some particular meaning for those who buried their dead there. But apart from a favoured chamber orientation between ENE and south and however tempting it is to assume that the massive cairns made particular beliefs material, there is little about them to suggest what part the cosmos played in their belief systems (ibid, 79; Fig 23. Two of the most prominent ritualised landscapes of much the same period as Calanais are that centred on the stone rings at Machrie Moor and that around the great stone rings in the middle of Orkney, which includes other features both above and below the surrounding modern plough soils. The multi-period complex in Crinan/Kilmartin Valley, overlapping in date with the others, may be another instance even though it has no surviving tall-stone rings. The monuments around Balnuaran of Clava and the several clusters of recumbent stone circles in NE Scotland, up to a millennium later than Calanais, may be a different expression of integration of landscape and cosmos. On the other hand, the favoured interpretation of the distribution of the stalled cairns of Rousay is that they each represent a territory belonging to

a household or small community, and the same may be true of the Clava cairns and recumbent stone circles. There were also timber and earth ritualised landscapes, on the plough-lands, for instance at Forteviot Leadketty with a 265m diameter enclosure and entrance avenue, henges and burials (Noble 2009, 14-17; 2010; Brophy, K & Noble, G 2011). Other examples exist at Meldon Bridge with its large enclosure and entrance avenue (Speak & Burgess 1999) and Dunragit with its multiple concentric timber rings and massive burial mound in Wigtownshire (Thomas forthcoming). The Balfarg/Balbirnie complex in Fife included early neolithic pit scatters in the east and the west, a possible long barrow, a major henge, timber mortuary structures, a smaller henge and a small stone circle (Barclay & Russell-White 1993; Barclay 2005, 88). Thus the visible clusters are the remains of a once more widespread type of landscape containing ritual sites, some far more impressive in their time than Calanais. The ceremonial centre at Machrie Moor on Arran included at least seven stone circles, two of which were built of exceptionally tall stones. Two of the lower rings were precede by timber ones. To its north-west were extensive strip field systems measuring up to 200m by 50m, and thought to belong to the Neolithic, period. Both to the north-west and south-west open fields of the Late Neolithic to Early Bronze Age were found and Neolithic remains were found also in a third area. Boundary dykes of the late Bronze Age and Iron Age were abundant, perhaps serving to distinguish in-field from out-field (Barber 1997, 144-145). Barber argued that settlement during the Neolithic period was much more abundant than has previously been suggested (ibid, 149). As at Calanais the visible stones at Machrie Appendices \ 1196

Calanais Survey and Excavation, 1979-88

Moor overlay remains of earlier activities. Excavation of two of the rings of low stones (numbers 1 and 11) revealed a complex history (Haggarty 1991). Charcoal from pits dug during the Neolithic and containing Neolithic pottery was dated to between 3710 and 3380 BC (GU-2321, 4820+/-50 BP) while another date, between 4500 and 4140 BC (GU-2320, 5500+/-70 BP) is either anomalous, or reflects a mixture of Neolithic and Mesolithic charcoal, or, conceivably, very early Neolithic settlement (ibid, 57-8). A complex timber ring was set up on Site 1 including two non-concentric circles and a horse-shoe setting of 5 massive timbers, and Grooved Ware was found in post-holes of the inner ring and the horse-shoe setting. Charcoal from the inner ring was dated to between 3350 and 2920 BC and between 2950 and 1950 BC (GU-2316, 4470+/-50 BP; GU2325, 3980+/-180; ibid. 63). A simple circle of timbers was set up on Site 11 (Haggarty 1991, 61, Illus. 5) and charcoal from it was dated to between 2900BC and 2350 BC (GU-2324, 4080+/-90 BP; ibid, 63), leaving open both the possibility that the two circles were contemporary and that they were successive. After a phase of ploughing and erection of fences, with which sherds of Beaker and Impressed Ware pottery were found, stone circles were erected on both Sites 1 and 11. Circle 1 was almost exactly on the Site 1 timber ring. Circle 11 was exactly on the circuit of the earlier timber ring and most of the stones were set equidistant between the timbers of the earlier setting. The locations of the holes for these stone seem to have been marked out in advance by timber posts. A terminus post quem of between 2210 and 1920 BC was provided for the erection of Circle 11 by oak charcoal (GU-2323, 3690+/-50 BP; ibid, 76). Subsequently burials were inserted into the area. The low stone circle on Site 11 is significantly later than its timber predecessor and there are 19 chances out of 20 that the interval between them was between 250 and 900 years if the excavatorâ&#x20AC;&#x2122;s interpretation of the stratigraphy was correct; there is no reason to suggest that it was not. This is quite extraordinary; given that ploughing will have removed physical traces of the timber rings, either the precise site of the timber circle was remembered for at least ten generations or there

existed for many generations precise rules for the placement of structures relative to other features in a sacred landscape, perhaps in this case the tall stone rings. If the rings of tall stones at Machrie Moor are of much the same date as Calanais they should also be of a broadly similar date to the timber circles on Machrie Moor 1 and 11. This conjunction of stone and timber settings could be explained through the analogies with Madagascan sites propounded by Parker-Pearson and Ramilisonina (1998), who suggested that the timber structures were built for the living and stone ones for the dead. However it does not match well with their idea that these different functions were carried out in separate large areas. The area round Machrie Moore was the subject of a case study by Barnatt and Pierpoint (1983, 101) to distinguish between two hypotheses: that stone circles may be viewed as scientific instruments and that they were impressive focal points of a society that did not use such structures to support any interest they had in astronomy. Many archaeologists and archaeoastronomers no longer accept that dichotomy (Ruggles 1999). The main value today of their paper lies in its examination of the placement of the rings in what must have been, if Barberâ&#x20AC;&#x2122;s deductions are right, a well populated landscape. Starting from the proposition that the monuments should be in the best possible position for astronomical observation if the latter were important to the circle builders Barnatt and Pierpoint concluded that there was no evidence in their siting for an interest in precise alignments on the main solar or lunar events, but that the rings may have been placed to exploit a horizon notch indicating midsummer sunrise (Barnatt and Pierpoint 1983, 112). There are problems with their study in the number of prior assumptions made about which lunar and solar events should be selected for the assessment of astronomical potential (Ruggles 1983, 116; 1999, 133). Also, their analysis was based on declinations centred around 2000 BC, and did not take into account the long history of use of the ridge on which the rings sit. Barnatt and Pierpointâ&#x20AC;&#x2122;s conclusions have to be treated with caution (Ashmore 1999c). That said, the main facAppendices \ 1197

Calanais Survey and Excavation, 1979-88

tor in the placing of the group of circles may have been their prominence in the local landscape. As things stand the concept of a ritualised landscape has to be employed in a much more limited way at Machrie than at Calanais. From currently available evidence it seems that people looked towards a single ritual focus rather than imbuing many locations with connections to a wider cosmos. However, detailed examination of the moorland along the lines pursued by Curtis and Curtis at Calanais, along with those suggested above for enhancing understanding at Calanais, may one day change that impression. Around and after 3000 BC a rich society expressed itself in settlements and monuments around and between the southern part of the Loch of Harray and the Loch of Stenness in Orkney. For ease of reference I shall refer to this area as the Heart of Neolithic Orkney although strictly speaking that term refers to the World Heritage site which includes Skara Brae and excludes many monuments near to the World Heritage area (Historic Scotland 1998, 39). It is sufficiently well known and publicised that I shall keep my account very brief; indeed the sometimes indiscriminate use of the archaeology of Orkney in discussions of the prehistory of other Scottish and indeed also English regions has raised protests which merit some sympathy (see for instance Barclay 2000). From an Orcadian perspective the singling out of this area has been seen as invidious. Other parts of Orkney contain 4th to late 3rd millennium landscapes which are exceptionally well endowed and the Heart of Neolithic Orkney may be only part of an archipelago-wide ritualised sacred geography. But the extraordinary non-domestic buildings in the settlements at Barnhouse and Ness of Brodgar, and the massive ceremonial and funerary structures in this area makes the case for its being part of a ritualised landscape particularly simple. The Heart of Neolithic Orkney also contains several 4th and 3rd millennium burial monuments including standing stones, massive boundaries

and two major settlement foci (Historic Scotland 1998; Richards 2005; Card 2010). Putatively 2nd millennium BC monuments are also fairly abundant, including large and small burial mounds and small stone settings (Historic Scotland 1998). The settlement at Barnhouse, on the shore of the Loch of Harray close to the Stones of Stenness, included Structure 8, a hall with a central hearth, a dresser against the rear wall and a complex entrance arrangement incorporating another hearth. It sat on a platform constructed of yellow clay and bounded by a substantial wall. The general arrangement of hall and platform is reminiscent of that at the chambered cairn of Maes Howe, just across the Loch of Harray (Richards (ed) 2005, 205 Figures 6.2-6.3). Further, taken in isolation, the archaeological remains of the central hearth and elaborate entrance arrangements of the hall have more than a passing similarity to the central hearth at Stones of Stenness and the features to its north. Richards has noted that their obvious differences concealed similarities in their modes of construction and, possibly, their purposes (Ritchie 1977, Fig 4; Richards 1993, 304-307; 2005, 221-2). Another settlement of broadly the same date as Barnhouse has been discovered at Brodgar Farm, between Stones of Stenness and the Ring of Brodgar (Ballin-Smith and Petersen 2003, 102-3; Card 2007; 2008; 2009; 2010; see also www.orkneyjar. com/archaeology/nessofbrodgar/). Landscape resistivity and magnetometry survey show that it was both complex and extensive and the excavations by Orkney Archaeological Trust and Orkney College demonstrate considerable stratigraphic complexity. It included four substantial buildings, the latest (Structure 10) more than 16m long by 10m wide, surrounding a central area. The stones composing some of them were painted; many others bore incised geometric designs and areas of pecking, while red and yellow sandstones were incorporated in their builds. They appear to have had slate-like slab roofs. They stood in a massive enclosure measuring c. 125m by 75m, the north wall of which was 4m wide and survives in places to over 1.7m tall. The pottery associated with the complex was Grooved Ware. A massive deposit of cattle bones is interpreted as evidence for feasting Appendices \ 1198

Calanais Survey and Excavation, 1979-88

possibly associated with the “decommissioning” of Structure 10. The complex is currently interpreted as “primarily non-domestic and ... an integral part of this rich ritual landscape” (Card 2010, 1-3). No radiocarbon dates have yet ( January 2012) been published, but unpublished dates are said to indicate a start-date for the complex before c. 3200/3100 cal BC and the cattle bones, thought to date within a very few centuries of the end of use of the complex, are reported to date around 2300 cal BC. The pottery from Ness of Brodgar was apparently all Grooved Ware. No sherds of the putatively earlier Orcadian pottery style, Unstan Ware, have been reported nor any of the (mainly later) Beakers styles. Another massive land division, of unknown date, has long been known slightly further north (Historic Scotland 1998). Large burial mounds formed another feature of the local Orkney landscape before the time Calanais was abandoned to peat growth. Big Howe near the Stones of Stenness and the mound just north of Bridge of Brodgar may include funerary structures under much later putatively secular ones, and there are four large burial mounds round the Ring of Brodgar. Further to the north the Ring of Bookan appears to extend the area regarded as suitable for major ceremonial and burial monuments. Smaller tumuli also survive locally in areas without a long history of ploughing. There are no relevant absolute dates for the Ring of Brogar. The dates available for silting up of the ditch (SRR-502, 2210+/-60; SRR-503, 2320+/-50; Renfrew 1979, 39-43) belong in the Iron Age. Recent work by Richards et al (2009) has not as yet improved the dating evidence. In my view the likeliest date for construction of the Ring is between 2500 and 2000 BC, since it has two diametrically opposed entrances; that estimate is based on the frequent occurrence of Beakers at two-entrance henges. If that is right the Ring of Brodgar was built 5 centuries later than the likeliest building date for the Stones of Stenness. The Stones of Stenness, the hall at Barnhouse Structure 10 at Ness of Brodgar and the chambered cairn at Maes Howe illustrate different but related aspects of the beliefs of a coherent social

group. Different parts of the landscape may have held different meanings for the people of this area. Modern ideas about the Stonehenge / Durrington complex in England provide food for thought. Parker-Pearson and Ramilisonina used analogies with recent cultures in Madagascar to suggest that the area around Stonehenge and Durrington Walls was divided up between a domain of the living and a domain of the dead (Parker-Pearson and Ramilisonina 1998, Figure 6). It was suggested that the period of building of the sarsen stone settings at Stonehenge (centred around 2400 BC) and that of the timber settings at Woodhenge and at Durrington Walls (2500 to 2000 BC and 2500 to 2100 BC respectively) may have been broadly the same (ibid, 314) although that becomes a moot point with the proposed re-dating of the sarsen structures at Stonehenge to between 2580 and 2470 cal BC (Parker-Pearson et al 2009, 26). In the area of the living the massive earthwork at Durrington Walls with its internal timber settings and the timber setting inside the henge at Woodhenge were places for ceremonies. In the domain of the dead, the stone monument at Stonehenge was a place of the ancestors. In between was an area of a few square kilometres through which people could pass from one domain to the other, perhaps using the Avenue as a ceremonial route for introducing initiates to the ancestors. That area was part of a liminal zone surrounding the domain of the ancestors (Parker-Pearson and Ramilisonina 1998, Figure 7) and was used for early Bronze Age burials. The Ring of Brodgar may fall into the same broad period as the monuments discussed by Parker-Pearson and Ramilisonina. Leaving to one side the difference between stone and timber, which may have had little relevance to Orkney, maybe there was a change of focus; Stenness and the nearby settlement at Barnhouse were left to the ancestors while Brogar became a centre for the living. Be that as it may, the ritualised landscapes at Calanais and the Heart of Neolithic Orkney differed considerably from each other despite the existence of tall-stone rings in both areas. Central Orkney contained two large settlements, and several large funerary mounds; it did not have the Appendices \ 1199

Calanais Survey and Excavation, 1979-88

profusion of small rings and stone settings found round Calanais. That the Kilmartin Valley supported what I have here called a ritualised landscape has long been recognised. Not only was it ‘… certainly a centre for ceremony and burial for an extended period …’ (Ritchie 1997, 83) but ‘The concentration of sites and the intervisibility of many within the confined valley floor allow us to conjure notions of a numinous landscape with an importance that might extend beyond the immediate locality’ (Ritchie 1997, 79). The Kilmartin valley contained both timber and stone structures. The latter have been long known, with a linear cemetery of massive cairns, along with a henge, standing stones including the decorated stones at Nether Largie and Ballymeanoch, and the stone rings at Temple Wood. It also contains numerous decorated rock outcrops (Ritchie 1997, 77- 84) and has produced fine artefacts, including highly decorated pottery of many periods and jet jewellery (Ritchie 1997, 80-84). The discovery of major timber monuments followed excavations in advance of gravel quarrying near Upper Largie in the late 1990’s. They revealed a long rectangular timber enclosure or pit-defined cursus, a pit-defined avenue, a pit and a pit or post-ring complex, Beaker graves, cists and burials, an elliptical timber ring and a short row of posts (Terry 1997, 19-21 fig 8; Ellis C 2000, 16; 2002, 145; Ellis and Crone pers. comm.; Cook et al 2010). Kilmartin is one of the few areas in Scotland where there is enough scientific dating evidence to (start to) form a long chronology. The 25 radiocarbon ages from Upper Largie include three suggesting forager activity in the second half of the 5th millennium BC. These activities at Upper Largie (and doubtless elsewhere) led to the creation of a substantial amount of charcoal. Nine ages were obtained in all from the postholes of the long rectangular enclosure; the youngest suggesting a date in the second quarter of the 4th millennium BC. It, the Clyde Cairn at Nether

Largie South and the massive cairns of the late 3rd and 2nd millennia, at least one overlying what may have been free-standing earlier stone and earth structures, overlapped the creation of timber and stone settings of various kinds including the timber ring at Upper Largie. A human burial belonged in the centuries around 2000 BC. The timber circle at Upper Largie seems to have belonged around 1500 cal BC judging by the youngest age from it although other ages suggest a somewhat earlier date in the second quarter of the 2nd millennium BC (Ellis 2002, 145; Ellis and Crone pers. comm.; Cook et al 2010). The stone ring at Temple Wood was preceded by a small oval ring of stones to the north, which itself seems to have been later than a small timber ring. The latter may be as early as 3500 cal BC but strictly speaking the radiocarbon age from it may relate to activities before the ring was set up (Ashmore 1997, 238). The southern ring has been seen as flattened, and at 13.5m by 12.5m it was similar in size and shape to that at Calanais. It was embanked and its interior was filled with boulders under and among which were small kerb cairns and burial cairns. Most of the 11 radiocarbon ages from this part of the site were obtained many years ago from samples of mixed charcoal, and most of them had large standard deviations. Although they give a very broad indication of the chronology of the long use of the site for burials they do not provide many useful dates for the site phasing or the inhumation burials. The most reliable of them suggest activities continued until the last half of the 2nd millennium BC but they were not directly associated with the cremations deposited in the ring (Scott 1989; Ashmore 1997 253, 266-270, 283). However, recent dating of cremated human bone has shown that Burials D, Primary E and Secondary E have indistinguishable dates suggesting burial between about 1420 and 1270 cal BC (Sheridan 2008, 202). Cist burials reported from other excavations near Upper Largie have been dated to around 2000 BC and to the centuries preceding 1000 BC (Mercer and Rideout 1987; Ashmore 1997, 263, 268; Cook et al 2010). The Beakers from Kilmartin show many simAppendices \ 1200

Calanais Survey and Excavation, 1979-88

ilarities to Dutch ones. Presumably that reflects contacts with the eastern areas of Britain. Irish Bowl Food Vessels from three cairns suggest long-distance western connections and Whitby jet suggests contacts with NE England in the last few centuries of the 3rd millennium BC; (Sheridan 2008; Sheridan pers.comm). Smaller cairns and short rows of standing stones may be mainly of the later 2nd millennium BC. A cremation in the socket of the holed-stone at Ballymeanoch has been dated to between 1370 and 1040 cal BC (GrA-28613 2970+/-40 BP Sheridan 2005, 183). Many of the isolated short stone rows and stone pairs of the Kilmartin Valley are, at minus 30°, closely aligned on moonrise or set at the major standstill limit. A smaller number of alignments lie between minus 21° and minus 26°. Each of these is close to a row or pair with an alignment of minus 30°. Ruggles has suggested several explanations for this pattern including the possibility that rows and pairs of the latter group were aligned on midsummer full moon at a random point in the 18.61 year lunar node cycle but when the moon was seen to set further and further south another ‘more correct’ alignment was built (Ruggles 1999, 109). That said, claims of alignments relating to the earlier stone settings at Temple Wood and Nether Largie need reassessment in the field (Ruggles 1999, 59, 231 note 79). The apparent contrast between the location of the timber monuments in the upper part of the valley and those of the cairns and stone settings downstream raises the question whether the concept of zones for the living and zones for the dead can be applied to the Kilmartin Valley. But it is perhaps only the small amount of exploration of timber monuments in the valley which allows distinction to be drawn. It seems likely that there were strong connections between the people who lived in Kilmartin Valley and those of other areas along the western sea-ways in the 4th millennium BC and later. There were also connections with the Netherlands, Ireland and Yorkshire (Sheridan 2008; Cook et al 2010). The sacred landscape was on present evidence created and perpetuated by societies which were materially much richer and more widely connected than those around Calanais.

Clava cairns lie mostly along the coast of the inner Moray Firth and the rivers running south with a concentration in Strathnairn (Bradley 2000, 2, 184). The monuments at Balnuaran of Clava include two passage cairns, a ring cairn and a small rough kerb cairn, while in an area of about a square kilometre around them there are at least six other monuments (Bradley 2000, 5, 175-8). Radiocarbon dates from hazel charcoal on the old land surface under the cairn at Balnuaran of Clava imply a date for construction which lies between 1920 to 1680BC and between 2010BC and 1740BC (Bradley 2000, 115-6; AA24234, AA-25433), or in round terms about 900 to 600 years later than Calanais. The two passage cairns are orientated towards the point where the midwinter sun sank below a nearby hill (Thom 1966, 18; Bradley 1998, 136, 142; 2000, 122-5). The passage of the more northerly cairn still points at midwinter sunset, although the alignment would have been better 4000 years ago, but while the passage of the more southerly cairn points in the same direction midwinter sunset is blocked by trees (Bradley 2000, 122). The more southerly cairn lies on the line indicated by the passage in the north cairn. In Bradley’s words, ‘seen from the other monument, the summit of that cairn merges with the horizon. As a result the sun appears to set on top of the cairn itself ’ (Bradley 1998, 142). There may also have been orientations on midsummer sunrise and while lunar horizon phenomena could have been observed they were not positively indicated. Bradley perceived Balnuaran of Clava as unusual in Scotland, because he thought that indications of interest in solar movements were uncommon (Bradley 2000, 126). By and large Clava cairns each had a local significance; they dominated their immediate localities but were neither placed to be visible from afar nor to have unusually good views outward. They seem not to have been designed to be intervisible (Bradley 2000, 178-80). The layout of the monuments does not seem to have favoured prominent Appendices \ 1201

Calanais Survey and Excavation, 1979-88

horizon features (Bradley 2000, 182). But a sample of 25 cairns suggested that their orientations clustered round midwinter sunset and the major and minor lunar standstills (Bradley 2000, 181-2). Thus in the lands south of the Moray Firth (with outliers on the Black Isle to the North) there seem to have been many small ritualised landscapes. In places they were so close to one another that they probably formed a single large one. The pattern is far richer and more extensive than that visible around Calanais. The recumbent stone circles of North-east Scotland have a distribution largely confined to Aberdeenshire. On the limited reliable evidence currently available, at least some of them were built over half a millennium earlier than the radiocarbon-dated cairn at Balnuaran of Clava. Beaker sherds have been found in some abundance at them- and they were used as cremation cemeteries around 1000 BC (Welfare 2011, 162-167). There are no convincing prototypes for them outside the area but they share many traits with other Scottish ‘Bronze Age’ mon-

uments (Welfare 2011, 252). It is important here to take on board Tilley’s point about labels (Tilley 1998). Indeed in rejecting 85 out of 156 northeast Scottish structures claimed at one time or another to be recumbent stone circles Welfare deprecated the tendency to lump sites together into self-reinforcing typological schemes obscuring the true variety of cairns and megaliths in the area. Types of monument in the northeast ‘that can be tentatively identified and which would repay further study might include: small open circles; closed cairns surrounded by both graded and ungraded circles; cairns surrounded by platforms; other ring cairns with and without circles; and circles which have other types of settings in their circumferences.’ (Welfare 2011, 258). In 1984 Ruggles concluded that a highly significant general orientation was evident. That orientation was from the interior of the circle outward over the recumbent stone towards horizon points centred on the SSW, generally avoiding local horizons (Ruggles 1984, S77). He and Burl showed that overall there was no conclusive evidence for precise observations of the sun or the moon, and although the orientations did suggest an interest in the rising or setting of the major standstill moon this did not explain all the sites (Ruggles

Illus Appendix 9.1 A variety of ritualised landscapes Appendices \ 1202

Calanais Survey and Excavation, 1979-88

and Burl 1985, 57-8). Welfare (2011, 213) has argued that it is more likely that the general orientations of recumbent stone circles and Clava Cairns reflect an interest in the setting sun. But I must admit that I do not see why the ideas must be mutually exclusive, nor why they cannot have varied over time and space. No strictly contemporaneous settlements have been demonstrated. Bradley has shown that the local distribution of lithic scatters around Tomnaverie was not centred on the circle (Bradley 2005, 46-7), which may suggest that it was on the fringes of areas used for domestic and agricultural purposes. But Welfare catalogued a large number of circles with evidence for possibly contemporary nearby landscape elements and suggested that they had settlements close to them. He contrasted their distribution with that of large burial cairns, some ‘found deep in the glens that penetrate the mountain hinterlands’ (Welfare 2011, 64-67). In terms of possibly ritualised landscapes the most significant aspect of the Clava and recumbent stone circles distributions is the impression they give of social landscapes mixing ritual, burial and (hypothetical) domestic sites. But the existence of the circles does imply that observance of some contemporary rituals was focussed in particular places and to that extent the landscape was partially ritualised. It must be stressed that Illus 24.70 is based on very incomplete information because of the small number of excavated sites and the limited amount of archaeoastronomical survey. I suspect that the differences which it encapsulates will diminish with future archaeological discoveries. For example the position of ‘Early Kilmartin valley’ on the ‘astronomical’ line may be over-cautious. It is based on reservations expressed by Ruggles (1999) about claimed early astronomical alignments. Certainly the orientations of some of the stone rows there suggest that by the second half of the 2nd millennium BC local people had a well developed interest in astronomical events. Also, I have not included the evidence from Lowland Scotland, although the monuments so far excavated, for instance at the Balfarg complex in Fife, at Forteviot in Perthshire, Broomend of Crichie in Aberdeenshire, Dunragit, Holywood and Holm in

the south-west represent different patterns again. Nevertheless, even the few examples of ritualised landscapes discussed above suggest that there was no overarching template to their creation. Barnatt (1989, 5) has suggested that stone circles and related monument forms may ‘to some extent form parts of a continuum’, particularly the smaller ones. He thought that small stone circles, ring cairns, kerb cairns and other round structures may, functionally, have expressed very similar ideas in different ways in different places and times (Barnett 1989, 9-13). Welfare expressed a similar idea after his exhaustive study of recumbent stone circles in northeast Scotland, going so far as to write ‘ ... similarities as might appear to exist are better seen as shared traits in contemporary manifestations of country-wide beliefs’ (Welfare 2011, 259). There is indeed some indication that there were broadly favoured belief systems in Scotland even if local expressions varied. The examples discussed above suggest that there was a stronger interest in building structures referencing the movement of the moon than to the sun, with Maes Howe and Balnuaran of Clava the most obvious exceptions. (although as noted above Welfare (2011, 226-8), perceiving mindsets distracted by the idea of precise alignments on mainly lunar events, and proceeding from the new orthodoxy that the orientations of recumbent stone circles and Clava cairns were not astronomically exact, preferred to explain their orientation as referencing the sun). Midwinter orientations seem to have predominated over midsummer ones, although how far that is due the relative lack of investigation of timber monuments is an intriguing question given the idea that stone monuments were for the dead and timber ones for the living (Parker-Pearson & Ramilisonina 1998). But more generally there seems to have been no over-riding uniformity of practise. At Machrie Moor, there seems to have been a central sacred place where many monuments were built close to one another, but the surrounding landscape was not ritualised. The main setting at Calanais dominates its landscape and it is inter-visible with many of the smaller settings around it. The sacred landscape in and around the Heart of Neolithic Orkney included a variety of substantial monuments well Appendices \ 1203

Calanais Survey and Excavation, 1979-88

spaced out from each other; a large settlement existed nearby to the east of the Stones of Stenness and a collection of large houses or halls, possibly with a mainly non-domestic function, has been explored to its north. The recumbent stone circles of northeast Scotland seem to be parochial sacred places each similar to the others. Further west, at around the same time, one or two of the sacred places incorporating Clava cairns were larger than others but seemingly did not otherwise differ. In this discussion I have favoured the ideas put forward by Bradley (2000), Jones and Thomas (2010) and Welfare (2011) that in the changing societies of the late 3rd and early 2nd millennia BC Scotland ‘ancestral’ monuments were built, either as a reaction by natives to incomers or as an attempt to claim local ancestor kinship by incomers. The creation of new ‘ancestral’ monuments occurred both in areas where real but ruinous ones were fairly common (Calanais and Embo) and in areas where they were rare (Clava and the lands of the recumbent stone circles). Yet there are hints that the apparent lack of earlier large funerary and ceremonial monuments may be more apparent than real. The Clava style platform cairn and massive graded kerb at Stoneyfield, Raigmore may be some 500 to a thousand years earlier than Clava cairns. But there are severe problems with the radiocarbon dates from the site. A structure like the two timber enclosures at Balfarg Riding School, built of perishable materials (Simpson 1996, 62-3 Illus 10), underlies the cairn. Grooved Ware was associated with it. The ‘permanent’ stone structure may be considerably later despite Simpson’s (not unreasonable) argument that its placement around the timber structure leaving the latter near-central suggests ‘near-continuity’ (Simpson 1996, 65-6). So I do not find the evidence demonstrates a visible pre-Clava ‘ancestral monument’ in the sense Bradley appeared to mean. Yet in some other areas such as the Kilmartin Valley it may be more realistic to see continuity, neatly encapsulated by the linear cemetery there. The chamber of the early cairn at Nether Largie South contained a secondary built cist associable with Beakers; the earliest burials in the massive Kilmartin Glebe cairn further north in the linear cemetery were accompanied by Food Vessels. The

one monument was reused while it was in a good state of repair, ancient though it was at the time, and the other was built anew. As remarked above (24.1.2) the most fruitful way forward in disentangling these similarities and differences is through local and regional studies of evidence from surveys, excavations and museum collections and subsequent comparisons between the results from other similarly studied areas, with any temptation to restrict interpretation to local comparisons tempered by a recognition that some long-distance links did exist and that occasionally there were wider-scale changes (Beek 2011, 45).

Appendices \ 1204

Calanais Survey and Excavation, 1979-88

Appendix 10 Stratigraphic diagrams of contexts with finds

Illus Appendix 10.1Â Area A Appendices \ 1205

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.2Â Area BI and BIVN Appendices \ 1206

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.3Â Area BINX, BII and BII/III Appendices \ 1207

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.4Â Area BIII Appendices \ 1208

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.5Â Area BIVWX Appendices \ 1209

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.6Â Area BIWX Appendices \ 1210

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.7Â Area BV and BIVS Appendices \ 1211

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.8Â Area C Appendices \ 1212

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.9Â Area DI Chamber Appendices \ 1213

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.10Â Area DI Appendices \ 1214

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.11Â Area DII and DIII Appendices \ 1215

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.12Â Area DIV Appendices \ 1216

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.13Â Area DV Appendices \ 1217

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.14Â Area E Appendices \ 1218

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.15Â Area F Appendices \ 1219

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.16Â Area HI Appendices \ 1220

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.17Â Area HII Appendices \ 1221

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.18Â Area HIII Appendices \ 1222

Calanais Survey and Excavation, 1979-88

Illus Appendix 10.19Â Area S Appendices \ 1223

Calanais Survey and Excavation, 1979-88

Appendix 11 Analysis of the pottery and other finds from cairn contexts in Areas B, D and H This is referred to in Chapter 24 Discussion and Conclusions 24.12.3. PRE-CAIRN All of the pottery (10 catalogue entries) was E/MN corky or probably E/MN corky apart from 48_53 which included a Heb Inc sherd and 1032_1033 which might instead have been from a corky fine Beaker. Table Appendix 12.1 Artefacts from immediately pre-cairn contexts Area Ctxt Description and interpretation

Compact green clay layer with a very minor sand component and some rotted grits, with a surface pattern of polygonal cracks 752. HII 760 Cut by linear feature beneath chamber 773. Notes Often referred to in the narrative as a ‘green clay platform’.

Pot Cat No

Dark brown gritty clay noted in west corner of trench. Under 370 bedding layer for 1104_1105 DIV 375 cairn boulders. Over 332 = possible bank material plus 394 = turf line. DIV 375 As above

Medium grey-brown crumbly clay loam with some olive-green flaking, charcoal-rich 48_53 HII 747 and greasy when rubbed. Under 733 = lower fill of cairn. Over 760 = green clay platform. HII 747 As above

179_180

HII 747 As above

1034

HII 747 As above

1032_1033

HII 747 As above

1035

Orange/yellow speckled clay under 373 = DIV 391 redeposited green clay and over 394 = turf line. Interpreted as same as 339 in DII.

Details

Other

81.225 quartz flake 81.228 quartz 81.227 quartz flake

E/MN prob corky

81.283 quartz flakes

E/MN corky Heb Inc (5 sherds) E/MN corky

E/MN corky

81.209 quartz chip

E/MN corky or fine Beaker,

E/MN corky

81.320 quartz chip x2 Appendices \ 1224

Calanais Survey and Excavation, 1979-88

PRE-CHAMBER WALL SLOT AND PRE- PASSAGE WALL FEATURES The bottom fill 730 of a slot under the chamber wall included two probable Chalco/EBA Beaker sherds and an E/MN corky sherd. The features under the passage wall included one E/MN corky sherd, 3 E/MN non-corky sherds , a sherd which was either E/MN non-corky or domestic Beaker and a fine probable Beaker sherd. Table Appendix 12.2Â Artefacts from pre-chamber wall and pre-passage wall contexts Area

Ctxt

Description and interpretation

Pot Cat No Details

HII

730

Grey-brown with slight green tinge, gritty sandy clay with small sub-angular 671 stones; charcoal-rich, uncemented and non-greasy. The bottom fill of slot 773

HII

730

The bottom fill of slot 773

672

Prob fine Beaker

HII

730

The bottom fill of slot 773

162_165

E/MN corky

BIVWX 878 BIVWX 881 BIVWX 881 BIVWX 881

BIVWX 885

Clay grey greasy clay in feature below passage north 67 wall

Dark green fill clay in feature below passage north 245 wall

Dark green fill clay in feature below passage north 246 wall

Dark green fill clay in feature below passage north 247 wall

Dark green fill below passage north wall, like fill of slot 883, the continuation of the slot on Area H

Prob domestic Beaker

Other 81.332 bone (burnt) 81.350 cremated bone? 81.352 = bone (burnt) 81.682 = quartz flake 81.667 quartz flake 0.02 gm of cremated bone from sieving by AOC in 2005/6

E/MN Corky E/MN Non-corky E/MN Non-corky E/MN Non-corky

735

?E/MN or Chalco/EBA Non-corky or dom Beaker

736

Prob Chalco/ EBA Prob fine Beaker Appendices \ 1225

Calanais Survey and Excavation, 1979-88

ARTEFACTS FROM THE PIT UNDER THE PASSAGE AREA Table Appendix 11.3Â Artefacts from the pit under the passage area

Area

Ctxt

BIVWX 889 BIVWX 889

Description and interpretation

Clay green pit fill of round pit in the passage Clay green pit fill of round pit in the passage

Pot Cat No

Details

Other

283

E/MN Heb Inc

841

?E/MN or ??Chalco/EBA Non-corky or dom Beaker

ASAL CAIRN LAYERS These layers included an E/MN Heb Inc catalogue entry (Cat 281) and an Indet one. Table Appendix 11.4Â Artefacts from basal cairn layers Area

DIV

Ctxt

370

398

Description and interpretation

Pot Cat No Details Other

Bedding layer of humic clay for cairn boulders - layer of dark brown/black clay with charcoal located among cairn boulders from middle of cairn to perimeter. Similar to 360 = greasy orange/brown clay with quartz chips, but on other side of kerb. Under 974 362 = soil on DIV and fill of upper chamber wall. Over 375 which may be redeposited natural from nearby monoliths. Note It shows ard marking (Film 1981.7.27) and may be equivalent to basal 315 or 320 on DI Patchy layer of green clay onto which rested base layer of cairn. Possibly redeposited 373 = capping over paving 381 around Monolith 29. Under 390 = bedding layer for 2nd layer of cairn boulders, and associated cairn boulders. Cut by 915 = linear feature. Over grey gritty clay layer, 905 = upper layer of several turf line under the cairn and chamber.

Indet

81.477 flint 81.509 hornfels 81.512 quartz scraper 81.554 quartz

81.552 quartz 81.591 quartz x 6

Appendices \ 1226

Calanais Survey and Excavation, 1979-88

DIV

Probable the same loose charcoal impregnated clay as found within lower courses of cairn material. Under 368 = 281 upper layer of stones located in vicinity of Monolith 29. Over 381 = “paving” of large slabs around Mono. 29.

910

E/MN Heb Inc

373.1 was yellow-brown at its surface and greener underneath. Contemporary with 355 = redeposited green clay. Under 362 = upper fill of chamber wall, and 380 = kerb of cairn. Over 381 = paving around Monolith 29, and 918 = turf line east of Monolith 29.

373.1

81.334 quartz (6.94/8.02, 1.15 BD) 81.335 quartz (7.15/8.12, 1.14 BD)

PRIMARY CAIRN Primary cairn layers included mostly E/MN sherds (10 catalogue entries), but they also included an E/MN Heb Inc sherd (174_175), E/MN or Chalco/EBA corky or Beaker sherds (168_171 and 1036–1038), Prob Chalco Prob Beaker sherds (541_542), and a Chalco/EBA sherd or sherds (351_352). Table Appendix 11.5 Artefacts from primary cairn layers Area

Ctxt

HII

733

HII

733

HII

Description and interpretation

Pot Cat No Details

Dark brown slightly sandy clay with more humus than 732, but more gritty and less greasy. Under 729 = fill 178 of cairn. Over 737 = outer cairn base and 747 = clay loam above 760 green clay platform.

176

733

177

81.674 charcoal E/MN corky sample - seen 23/10/95 E/MN corky E/MN corky

HII

733

E/MN or Chalco/EBA 1036–1038 corky or Beaker

HII

733

541_542

HII

759

Greyish very gritty clay just outside the base of the northern face of the chamber wall. Under 732 = lower fill of cairn interpreted as basal fill of cairn.

Other

Prob Chalco Prob Beaker

81.222 quartz flake

Appendices \ 1227

Calanais Survey and Excavation, 1979-88

HII

732

HII

732

HII

732

HII

Gritty slightly sandy clay basal fill of cairn. Under 728 = fill of cairn. Over 741 = base of inner cairn. Cut by 736 166_167 = yellowish-reddish clay by reddened stones. ?Same as 733 = fill of cairn.

732

E/MN corky

173

E/MN corky

172

732

E/MN corky

174_175

732

168_171

HII

736

Slightly reddish gritty clay, much less rooty than the surrounding 732, with a few fragments of burnt bone (Day Book 6) in an area of discoloured stone 351_352 in the cairn. Under 728 = cairn fill. In 732 = lower chamber fill. Interpreted as modification of 732 by a fire set when the cairn was being built.

HII

736

279

E/MN corky 81.103 quartz 81.108 carbonE/MN corky ised hazelnut one sherd 81.110 quartz Heb inc flake 81.114 quartz scraper E/MN or Chalco/EBA corky or Beaker Chalco/EBA International Maritime Beaker

E/MN non-corky

81.104.81 piece of cremated bone.

PRIMARY CHAMBER WALL The primary chamber wall contexts had 11 E/MN corky catalogue entries and one E/MN Heb Inc. Table 11.6Â Artefacts from primary chamber wall layers Area

Ctxt

HII

770

Description and interpretation

Pot Cat No Details

Very dark grey clay, more compact than the overlying 769. Over 771 = top turf line, 773 1040_1043 E/MN Heb inc = linear feature under chamber wall. Layer D of the stonework had 770 between its cracks.

Other 81.329 quartz flake (2.60/1.80) 81.330

Appendices \ 1228

Calanais Survey and Excavation, 1979-88

HII

770

E/MN corky

HII

770

218_219

E/MN corky

HII

770

7_9

E/MN corky

PRIMARY PASSAGE WALL The primary passage wall contexts had 3 E/MN corky catalogue entries and one E/MN Heb Inc. Table 11.7 Artefacts from primary passage wall layers Area

Ctxt

Description and interpretation

Pot Cat No

BIVWX 824

Passage North Wall fill

316

BIVWX 887

Passage North wall green clay

842

BIVWX 887

843

BIVWX 887

844

Details

E/MN Heb Inc

Other

E/MN corky E/MN corky E/MN corky

SECONDARY CAIRN This had 10 catalogue entries for certain, probable or possible E/MN corky sherds, 1 catalogue entry for a corky E/MN or Chalco/EBA Beaker sherds and another for a none-corky one, 1 for a Chalco/ EBA fine Beaker and 11 for the Food Vessel ASH 75. There were also several pieces of quartz and one piece of flint. No glass or modern ceramics were found, in this way contrasting quite sharply with chamber contexts. Table 11.8 Artefacts from secondary cairn layers Area

Ctxt

HII

708

HII

708

HII

708

HII

708

Description and interpretation Pot Cat No Details Slightly greasy clay blanketing the cairn under turf and Min592 istry grit and over 712 = upper fill of cairn.

EBA FV

ibid

Indet ?E/MN corky

ibid

1022

?E/MN corky

ibid

1030

Prob E/MN corky

723

HII

708

ibid

1023

HII

708

ibid

1031

Other 81.597 quartz 81.479 flint

E/MN corky or Chalco/EBA Beaker

Chalco/EBA Fine Beaker

Appendices \ 1229

Calanais Survey and Excavation, 1979-88

HII

708

ibid

593

EBA FV

HII

708

ibid

560

EBA FV

Loose gritty dark brown clay fill between upper stones of cairn. Under 708 = gritty clay Under Ministry grit. Over 728 = fill of 594 structural stones of cairn plus 729 = more compact version of 728. Interpreted as secondary filling related to erection of kerb slab 711.

EBA FV

HII

708

HII

712

HII

712

HII

712

HII

ibid

712

1024_1029 EBA FV 591

156

E/MN corky

158

E/MN corky

157

712

724

HII

728

HII

728

Loose unconsolidated gritty clay with many fine fibrous roots amongst the structural stones of 725 cairn. Under 712 = upper fill of cairn. Over 732 = lower fill of cairn.

HII

EBA FV

E/MN corky

E/MN or Chalco/ EBA Non corky fine Heb inc or Beaker Prob E/NM non-corky

161

728

81.589 quartz x4

81.77 quartz 81.79 quartz 81.564 quartz lump

E/MN corky

159_60

E/MN corky

SECONDARY CHAMBER WALL Secondary Chamber wall contexts had 35 catalogue entries for certain or probable E/MN corky sherds and 2 for probably E/MN non-corky sherds. There were 2 entries for E/MN non-corky or Chalco/EBA domestic Beaker sherds. Other finds were quartz and glass. The two glass finds came from under the inner wall stones of layer C of context 769 and suggest that the upper wall had been re-set or rebuilt after peat was cleared from the area. Table 11.9Â Artefacts from secondary chamber wall layers Area

Ctxt

HII

768

Description and interpretation

Very loose light medium brown sandy rooty soil with many fragments of rotten stone between stones of layer A of chamber wall over 769

Pot Cat No Details

Other

18_26

81.670 quartz x 2

E/MN corky

Appendices \ 1230

Calanais Survey and Excavation, 1979-88

HII HII

768 768

HII

769

HII

769

HII

769

HII

27_28 29 Dark brown very loose clay with a little sand friable rooty fill of chamber wall varying slightly in amount of humus and in colour to brown with a green 181 tinge. In layer C 769 was ‘very rooty and loose towards the chamber - more like 768’ according to Plan 26

769

182_200 201_206

769

1039

Prob E/MN non-corky

E/MN noncorky or Chalco/ EBA domestic Beaker

E/MN corky

Not in catalogue

81.327 glass 81.328 glass 81.669 quartz flake

E/MN corky E/MN corky

E/MN noncorky or Chalco/ EBA domestic Beaker

PREHISTORIC DILAPIDATION? Context 352 was a set of small features filled with black sandy clay under the basal levels of the main plough soil outside the cairn kerb. It contained catalogue entries for 4 E/MN corky sherds and 4 indeterminate sherds along with two fragments of quartz. Context 374 was a slightly lower level of black compact clay forming a margin round the cairn. It contained 5 catalogue entries for certain or probable E/MN corky sherds, 1 for an E/MN Heb Inc or non-corky sherd and in addition there was an empty bag labelled ‘Hebridean’. There was1 entry for ?Chalco/EBA ?dom Beaker and 3 for Indeterminate sherds. The interpretation of the features is difficult. They may reflect both down wash (374) and removal of stones during ploughing (352) at an early stage of ground-working, before Beakers became abundant at Calanais, along with initial cairn dilapidation. Table Appendix 11.10 Artefacts associable with dilapidation of the cairn Area Ctxt Description and interpretation

352

Pot Cat No

Black, slightly sandy clay small features outside the cairn. Under layer 315 plough soil; it directly underlay 369, the basal level of 315. Over turf line 365 114_116 which overlay clay 389 (area of mottled black/brown clay. There are complications here which did not appear in plan and it is not clear what 352 represents.

Details

Other

E/MN corky

81.511 quartz 81.550 quartz chip 81.294 no pot in bag

Appendices \ 1231

Calanais Survey and Excavation, 1979-88

352

928_931

352

E/MN corky Indet

Not in catalogue

374

Margin of black compact clay with charcoal around perimeter of cairn Under 320 and 369, i.e. lower level of 315 producing Hebridean ware. Over 622 Turf line 334 and probable linear bank feature 388. Interpretation Thought to be mineral percolation from cairn.

374

624

?E/MN corky

374

120

E/MN corky

374

625

374

?E/MN corky

121

374

623

374

340

374

953

374

945

Indet

81.275 flint flake (5.40/6.90, 1.15 BD) 81.473 flint 81.270 peat was entered in finds book but deleted and the finds number reassigned

E/MN Corky Prob E/MN corky

E/MN Heb Inc or non-corky Empty bag saying ‘Hebridean’

374

?Chalco/EBA ?dom Beaker

946_952

Indet

LATE KERB SLAB SLOTS ON AREA BIWX Table Appendix 11.11 Artefacts in late kerb slab slot fills, probably associable with dilapi-dation of the cairn Area

Ctxt

BIWX

852

Description and interpretation

Pot Cat No Details

Black clay soil in late kerb slab socket 785 at base of cairn

Other

E/MN Heb Inc

Appendices \ 1232

Calanais Survey and Excavation, 1979-88

BIWX

852

Black clay soil in late kerb slab socket 786

BIWX

855

Early kerb slab socket

BIWX

855

Early kerb slab socket

BIWX

855

Early kerb slab socket

787

BIWX

855

Early kerb slab socket

788

Early kerb slab socket

789

Early kerb slab socket

790

Early kerb slab socket

791

Early kerb slab socket

792

Early kerb slab socket

793

E/MN Corky

E/MN Corky (but see under fabric) E/MN Corky

E/MN Corky

‘FACADE’ SLAB PIT AT PASSAGE ENTRANCE The ‘facade slab’ pit at the passage entrance included two E/MN Corky sherds, an E/MN Hebridean Incised sherd, 4 Beaker sherds, 3 E/MN Non-corky or Chalco/EBA Beaker sherds, 3 Indet or possibly domestic Beaker sherds, 2 Post-EBA possibly indeterminate IA or IA ‘Plain Style’ pottery and two completely indeterminate sherds. It is hard to explain the presence of the post-Bronze-Age sherds particularly if they belong in the first millennium AD. Table Appendix 11.12 Artefacts from the ‘facade slab’ pit at the mouth of the passage Area

Ctxt

BIVWX 859 BIVWX 859 BIVWX 859 BIVWX 859 BIVWX 859 BIVWX 859

Description and interpretation

Black greasy clay in passage facade socket

Pot Cat No

Details

E/MN Corky

422

Chalco/EBA Beaker

308 423 424

BIVWX 859

450

BIVWX 859

602

BIVWX 859

603

Other

E/MN Heb Inc

Chalco/EBA Beaker Chalco/EBA Beaker Chalco/EBA Beaker Early

Indet; ??Chalco/EBA?? dom Beaker Indet; ??Chalco/EBA?? dom Beaker

Appendices \ 1233

Calanais Survey and Excavation, 1979-88

BIVWX 859

604

BIVWX 859

614

BIVWX 859

615

BIVWX 859

863

BIVWX 859

865

BIVWX 859

866

BIVWX 859

867

BIVWX 859

864

Indet; ??Chalco/EBA?? dom Beaker Post-EBA: ?IA or later Indet; ??’Plain Style’ pottery Post-EBA: ?IA or later Indet; ??’Plain Style’ pottery Indet Indet

E/MN Non-corky or Chalco/EBA Beaker E/MN Non-corky or Chalco/EBA Beaker E/MN Non-corky or Chalco/EBA Beaker

AMBIGUOUS CONTEXTS, POSSIBLY DISTURBED BY VICTORIAN INTRUSIONS Table Appendix 11.13 Artefacts from ambiguous contexts possibly disturbed by Victorian intrusions Area

Ctxt

360

Description and interpretation

Pot Cat No Details Other

A layer of greasy orange/brown clay with quartz chips in the interstices between stones of the 620 cairn near the outer face of the chamber wall, Under 318 layer IV

932

621

Prob E/MN noncorky

81.478 flint? 81.336 quartz chips 81.585 quartz x 5 81.584 chert x 2 81.339 iron pan/pot?

Chalco/ EBA Fine Beaker ?E/MN noncorky/ Heb Inc

Appendices \ 1234

Calanais Survey and Excavation, 1979-88

DIV

362

Dark brown gritty charcoal-rich clay fill of cairn. Under 300 = turf. Over 370 = layer of dark brown/black clay bedding, and 373 = the clay capping over stone paving around stone 29. Interpretation Same as 360? Note DIV did not include much of the cairn fill. 360 was disturbed during the Victorian period. This layer may also have reflected Victorian disturbance.

81.556 quartz 81.557 quartz flake 81.558 quartz flake 81.559 quartz flake 81.562 quartz flake 81.563 quartz 81.581 quartz x 9 81.583 quartz x 2 81.476 flint 81.555 flint plus chert 81.560 burnt flint 81.561 hornfels 81.547 small pebble 81.645 glass

VICTORIAN SOUTHERN CHAMBER WALL FILL Table Appendix 11.14 Artefacts from Victorian southern chamber wall fill Area

Ctxt

361

Description and interpretation

Fill of lower chamber wall, dark brown friable clay thought to have been washed in rather than deliberate fill. Same level as 360 = greasy orange/brown clay. Interpretation The peat was probably washed in during the clearance of 1857. If this is correct it implies that the upper part of the southern face of the chamber wall in Area D was very dilapidated, as interpretation of early accounts of the clearance independently suggest; Sharbeau’s sketch may have been drawn after repairs had been made.

Pot Cat No Details Other

81.253 peat 81.598 quartz

VICTORIAN CAIRN FILL? Table Appendix 12.15 Artefacts from Victorian cairn fill Area

Ctxt

318

Description and interpretation

Cairn and soils in it including layers 318.1 to 318.4. This is a portmanteau label.

Pot Cat No Details

Other 81.342 “knucklebone” 81.346 quartz

Appendices \ 1235

Calanais Survey and Excavation, 1979-88

A loose grey, gritty soil behind false kerbstone 318 1 interpreted as Victorian reconstitution.

326

432_437

326

438

326

439_449

326

711

A loose black soil, filling stones of cairn - animal 318.3 burrows interpreted as prob- 619 ably disturbed by Victorian activities

318.3

420

318.3

421

318.3

334_338

Grey gritty patches which lay in a drift over the kerb.

425

318.3

709

318.3

459

DI 356 chamber

Layer of friable dark brown clay. Possibly equivalent to 390 = surface on which the second bottom layer of cairn stones was laid. Under 350 = chamber wall. Over 355 = redeposited green clay.

Chalco/EBA Beaker Chalco/EBA Beaker

80.49 quartz arrow head 80.66 quartz 80.148 quartz 80.149 quartz 80.211 slate 80.210 glass 80.187 quartz *80.150 quartz

Chalco/EBA Beaker Chalco/EBA Beaker Prob E/MN non-corky Prob E/MN

80.66 quartz 80.68 flint 80.170 worked stone

Chalco/EBA Beaker Chalco/EBA Beaker E/MN? Heb inc?

E/MN noncorky or Chalco/EBA domestic Beaker Chalco/EBA Beaker

Not in catalogue

81.274 quartz x 2 (1.13 BD)

Appendices \ 1236

Calanais Survey and Excavation, 1979-88

VICTORIAN CAIRN INTRUSION FILL? The fills of this intrusion contained an Indet sherd and quartz. Table Appendix 11.16 Artefacts from Victorian cairn intrusion fill Area DI

Ctxt

Description and interpretation

Pot Cat No

Grey greasy surface within cairn 318 and 331 = clods of grey 333.1 927 humus. Over 360 = greasy orange brown clay.

Note 24.12.4: Analysis of the chamber fill

Details

Other

Indet

81.226 charcoal 80.147 quartz 80.146 quartz

and modern contexts

Prehistoric pottery was almost completely absent from chamber fills (the only possible case 81.244, a pot fragment, is not in the catalogue) and there were abundant Victorian to modern finds. Table Appendix 11.17 Artefacts from Chamber fills Area

Ctxt

Description and interpretation

351

Stone tumble within chamber.

DI cham- 357 ber

DI cham- 364 ber

DI cham- 367 ber

Rectangular slot cut into green clay 355 with fill of loose organic-rich clay and stones. Under 364 = clay fill of tumble in chamber. Interpretation Fairly modern. Could this be a cast of the ‘lintel’?

Remnant layer of friable black humic clay fill of stone tumble in chamber. Same as clay within 351. Under 351 = stone tumble in chamber. Over 355 = redeposited green clay. Within 925 = chamber. Stones wedged under SW chamber portal this century or last. Under SW portal stone. Over 372 = undisturbed green natural. Within 925 = chamber.

Pot Cat No Details

Not in catalogue

Other

81.634 glazed pot 81.655 china 81.644 glass 81.655 glass 81.703 recent leaf or other organic matter 81.212 peat

81.700 insect wing

glass

Appendices \ 1237

Calanais Survey and Excavation, 1979-88

HIII

709

724

740

757

Loose with many grits Ministry grit. Under 702 = turf. Over 720 = gritty wash down from cairn. Note Fill very similar to 731. Interpreted as modern disturbance in chamber.

81.646 china 81.652 china 81.643 glass 81.651 glass 81.653 leather 81.654 iron

Redeposited natural in a cut interpreted as Victorian exploratory trenching

81.94 glass/ obsidian

Light olive green sandy clay loam in chamber with black mottling. Under 720 = grit wash down from cairn. Over 742 = recent “turf line”. Interpreted as modern redeposited natural from excavation. in chamber

glass and china 81.153 glazed pot (1.27, 2.68/2.20) 81.210 china (1.27, 2.54/2.21) 81.223 china (1.36, 2.88/2.37)

Sticky sandy grey/green clay with yellow brown mottles and iron staining. Under 753 = slumped portal stone, 770 = displaced fill of bottom chamber wall. Interpreted as redeposited natural.

81.221 quartz x 1 (1.40, 2.90/2.24)

MODERN CONTEXTS Table Appendix 11.18 Artefacts from modern contexts Area

Ctxt

HII

706

HII HIII

Description and interpretation

Pot Cat No

Ministry grit on cairn, blanketing it [Day Book 5b]. Under 701 = turf. Over 704 = cairn, 1021 and 712 = upper fill of cairn, 708 = gritty clay [Day Book 5b]

701

Turf over cairn 704.

702

Turf in chamber - lusher than turf in Areas I and II.

Details

Other

Indet

81.660 glass 81.2 flint 81.657 glass 81.659 glass 81.574 quartz 81.575 quartz x 3 81.4 1929 florin (2.45/2.18) 81.650 china 81.649 glass

Appendices \ 1238

Calanais Survey and Excavation, 1979-88

Appendix 12 Looking to the future One pressing need is publication of the wide range of environmental and archaeological fieldwork carried out in the area in the last few decades. Complementing that, research on sea levels and underwater exploration in Loch Roag has a large potential for revolutionising understanding of the story of foragers and early farmers in the area; and study of the land surface beneath the deep peat covering much of the area will probably reveal well preserved archaeological features. The slight remains of cultivation beds at Calanais and the discoveries of better preserved but perhaps substantially later cultivation beds at Calanais Fields (Flitcroft et al 2000, Johnson et al in prep) suggest that much evidence for 4th to 1st millennium field systems survives under surrounding peat. Judging by the results from Calanais settlements may be well preserved, apart from loss of bone assemblages. There is also a fairly high likelihood that well preserved evidence for early human activities will survive in the innermost parts of Loch Roag. The seabed between Tob nan Leobag and the southern shore is less that 2m deep and the wide expanse of deeper water to its east should at some date have contained a freshwater loch (Illus 24.4; J Crawford pers. comm.). More broadly evidence for how the loch-side landscape has changed, and how people exploited it should emerge if the submerged parts of the Great Western Island (Illus 24.3) are examined in a similar way to that in which Doggerland is being explored. The east-west trending inlets and passages may be particular fruitful areas for detailed underwater studies despite the current erosion of 2nd millennium sites (Shepherd 1976). For example any search for hunter-gatherer burials in the Western Isles should take account of recent ideas about the original placement of such burials. The Møllegabet canoe burial was found in shallow water at what would have been the mouth of a river near a Danish Baltic island (Grøn & Skaarup 1993, 38). Whatever the precise reason for its placement

where it was found, it provides a clue to where similar burials might survive – possibly, for instance along the submerged stream channels of inner East Loch Roag (Illus 24.4). Of the many field systems and fragments of structures revealed by peat cutting on Lewis, perhaps Steinacleit, in the care of Historic Scotland, is the likeliest well-known structure to be a major 3rd millennium domestic building, at least in one phase. It reminds me of some of the pre-peat halls of Shetland and although few of the latter are well-dated a smaller but otherwise similar structure at Ness of Gruting belongs around the last quarter of the 3rd millennium (Barcham 1980, 504). Steinacleit is a possible starting point for investigating late 3rd millennium farming communities in Lewis. As described in detail above, the discoveries made by Ron and Margaret Curtis encourage a working theory that many of the small structures found near Calanais formed part of a richly symbolic landscape (Illus 24.30). To explore their model studies of the landscape should move into a new phase, comparing the loci reported by Curtis and Curtis with others nearby and developing and testing models of how people used and moved around the land. It hardly needs to be said that the smaller rings around Calanais along with the many sites and potential sites discovered by Ponting and Ponting, and Curtis and Curtis, deserve geophysical survey and some excavation during those broader landscape studies. Colin Richards’ work on the sources of the stones used to create the setting (Richards 2006, 182) could usefully be expanded through detailed petrological studies of the standing stones and the many outcrops of gneiss in the surrounding landscape. It could be extended to the smaller stone settings. The possibility that stones were sourced amongst the communities which had an interest in Calanais suggests that such studies should not be restricted to the area immediately round Calanais and the other stone settings. Appendices \ 1239

Calanais Survey and Excavation, 1979-88

Closer to the stone setting at Calanais, excavation in the area of Test Pit G1, near the pre-modern kiln to the west of the southern part of the setting might reveal early settlement. The pottery from the test pit could all be of 4th or 3rd millennium BC date. The possible wall-base in Test Pit G2 could reflect medieval or post-medieval activities but given its proximity to Test Pit G1 it could instead relate to much earlier settlement. Turning to Calanais itself, the true layout and function of the ditch cannot be resolved without further excavation. That should start with emptying of the backfill in Area DI and investigation of the unexcavated early soils, followed by re-examination and scientific analysis of the soils and turf lines on the Area D east baulk section (Chapter 9). Subsequently excavation between Area D and Area B, and completion of the excavation of Subarea BV, might trace more of the ditch circuit although that would perforce demand equally careful re-examination of the complex overlying layers, complicated by the presence of Ring stones 44 and 45. Because it would remove any subsequent chance of exploring the strata in the southeast part of the Ring and the area immediately to its east, investigation should involve a systematic comprehensive sampling and on-site scientific support; it should be attempted only if substantial investment becomes available. The late slots in sub-area DI continued under the east baulk. Reinvestigation might clarify the problems of their relationship to one another and to the local turf lines . The idea that the Ring was not all built at once did not receive support from excavation. Nevertheless, the devil is in the detail. Subsoil weathering between many of the stones may have removed evidence but some recognisable strata did survive under the low bank of stones between sub-areas DII and DIII. They may survive elsewhere, particularly round the bases of Ring stones 51 to 53 which are close together. Excavation there might show whether Ring stone 52 was, as suspected, secondary. The possibility that there was a substantial early pit in the area around East Row stone 30 can only be tested by new excavation. That should start with the emptying of BV and might as well be planned

with the idea that that area would be excavated to natural - that would allow investigation of the early soils and cultivation beds. If funds allowed excavation around BV could test the idea that there were early Beaker graves nearby and the idea (from Somerville) that there might have been an enclosure in the area. There is a slight possibility (if pinned down I would guess a 1 in 5 chance) that the Beaker sherd used to provide a terminus post quem for chambered cairn construction was secondary. A substantial proportion of the cairn remains unexcavated. A programme of radiocarbon dating of individual pollen grains from basal cairn levels might provide a conclusive answer. But given the plentiful evidence for importation of soils with by the ancient material in them very many grains would have to be dated and until dating costs are greatly reduced that will not be affordable. It would also be interesting to look for drops of tephra (volcanic glass) in the layers under the cairn given the possibility that a period of poor treegrowth in Ireland around 2354-2345 BC might have been due to the Hekla 4 volcanic eruption in Iceland Baillie 1999, 205). Parts of the enclosure bank survive, particularly in the main north-south baulk left between sub-areas BI and BIII (Chapter 7). Again dating of many individual pollen grains might illuminate its chronology. The palisade trench under the southern part of the enclosure bank was interpreted as going with the second stage enclosure. But it remains possible that it was earlier. The surviving strata immediately east of Ring stone 43 will contain evidence allowing complete resolution of this issue. More data is need from the area of the Avenue. Future fieldworkers should first improve geophysical information (Chapter 5: Resistivity Survey, Illus 5.1, 5.10) using ground penetrating radar, resistivity tomography and other techniques (Chapter 5: Resistivity Survey, Illus 5.9; Part 5.4). The results reported here suggest investigation of the possibility that the Avenue was preceded by a row or rows of pits, and the possibility that some stones were removed in prehistory. They should explore the idea that there was a route-way from the north to the Ring. In response to speculations Appendices \ 1240

Calanais Survey and Excavation, 1979-88

about the early form of the Avenue they could look for systematic differences between the archaeology of the southernmost stones and the rest. In all this they should bear in mind the fairly high likelihood that the area occupied by the Avenue was used for agriculture in the 4th millennium BC. During future excavation one productive endeavour would be exposure of a large area round Area F and the use of modern analytical techniques to explore the mounds round the bases of Stones 11 and 12, the mineral soil and the layers underlying it. Judging by the results of our work there is considerable potential for dating of Avenue stones 11 and 12, and for fruitful exploration and dating of the prepeat cultivation north of the Ring. The area east of Calanais seems to have an important meaning for the builders of the Ring and of the chambered cairn. The peat on the ridge

there might be a productive target for geophysical investigation and test-pitting, if permission can be gained. Also investigation of the small erect stone at Cnoc Sgeir na h-Uidhe, about 1.5 km east of Calanais (Illus 24.30) might reveal whether it could have been used as a marker for equinoctial sunrise. Lastly, Calanais has provided inspiration for creative work. Some fine examples are provided in the official guidebook â&#x20AC;&#x2DC;Calanais: The Standing Stonesâ&#x20AC;&#x2122;, thanks to Heather Delday, my then editor (Ashmore 1995; reissued by Historic Scotland with cosmetic changes and minor updates in 2002). Other first-rate examples are contained in Calanais, the book accompanying the 1995 exhibition of the same name (An Lanntair 1995). I hope that this report will help to stimulate further creativity.

Appendices \ 1241