Wargames Research Group - Armies of the Macedonian and Punic Wars 359-146 BC

Armies of the Macedonian and Punic Wars Organisatlon, tactlcs, dress and weapons. 188 lllustratlons

by Duncan Head Drawings by lan Heath Perslan, Greek, Bolotlan, Spartan, Athenlan, Phokian, Altollan, Achalan, Tarantine. Syracusan, Macedonian, Thessalian Successor, Antlgonld, Epelrot, Ptolemalc, Kyrenean, Seleucld, Pergamene, Bactrlan and lndlan Greek, Maccabean, Thraclan Blthynlan, lllyrlan, Scythlan, Bosporan, Sarmatlan, Saka, Parthlan, lndlan, Carthaglnlan, Numldlan, Spanish, Celtic, Galatian Roman, Latln, Samnlte, Campanlan, Lucanian, Bruttian, Apullan and Etruscan armles.

A WARGAMES RESEARCH GROUP PUBLICATION

FRONT COVER: BLEPHANTS AT RAPIDA, 217 BC At lcfi, a Ptolcmaic elcphant ofthe small Nonh African specie.s. Two fighting crew may havc bccn tbc most this smallcr brccd could carry. Thc saddlecloth comcs from a Ptolcmaic wall painting, from the military settlcment of Marissa in Palcstinc, but othcrwisc no clcar rcprcsentations of Ptolemaic clcphants cquippcd for battlc survive, so this reconstruction uscs the same sourccs as figure 185. On thc right is one ofthe largcr Indian elephants in Sclcucid scrvice (figure 184).

"Tbc soldiers in the towcrs made a fine fight ofit, jousting with tbeir pikcs a.n d stabbing at each other, but the beasts tbemsclvcs fougbt still bcttcr, battling with ali their strength and butting at cach otber witb thcir forebeads."

(Polybios).

(Painting by Ian Hcath)

Armies of the Macedonian and Punic Wars 359 BC to 146 BC Organisation, tactics, dress and weapons. 188 illustrations

by Duncan Head Drawings by Ian Heath

Persian, Greek, Boiotian, Spartan, Athenian, Phokian, Aitolian, Achaian, T arantine, Syracusan, Macedonian, Thessalian, Successor, Antigonid, Epeirot, Ptolemaic, Kyrenean, Seleucid, Pergamene, Bactrian and Indian Greek, Maccabean, Thracian, Bithynian, Illyrian, Scythiao, Bosporan, Sarmatian, Saka, Parthian, Indian, Carthagioian, Numidian, Spanish, Celtic, Galatian, Roman, Latin, Saronite, Campanian, Lucanian, Bruttian, Apulian, and Etruscan armies.

A Wargames Rescarch Group Publication

INTRODUCTION This book deals with armies ofthe Hellenistic period, the age ofthe dominance of Macedon and the Hellenistic kingdoms which rose in the wake of Alexander the Great's conquests, from the accession of Alexander's father Philip II in 359, unti! the Romans defeated the Acbaian League and annexed Greece in 146. lt also covers the rise ofRome from obscurity to unquestioned mastery ofthe Mediterranean, ending with her destruction ofCanhage in I46 after the epic con test known as the Punic wars. As the period open ed, the Greek hoplite was stili the dominant troop type in the Mediterranean world, but he was supplanted by the Macedonian pike phalanx, and that in turn was overthrown by the Roman legion. Geograpbically, tbis book reaches from Spain, westernmost theatre oftbe Punic wars, to India, eastemmost land to feel tbe weight of Macedonian anns.

This replaces tbc fìrst work published in tbis series on early armies, Phil Barker's book ofthe same title, published in 197 I. Though bearing the same title as Phil's, this is an entirely new book rather than a revision or new edition. lt covers a number ofnations not dealt with by the earlier book, takes account of a great dea! of work published since then, and above ali covers che whole subject in greater detail. This reflects a graduai but far-reachiog cbange io tbe role ofthe series; originally brief outline booklets, in response to demand for more exact and specific informatioo they bave become mucb more tborougb and detailed, though obviously eacb title covers coo great a canvas in both space and time to bope to be completely comprebensive. As ever in the series, I have relied above all on ancieot literary, artistic and archaeological evidence, ratber tban the uneven and sometimes untrustworthy views of modem writers. I ha ve tried to indicate, within the limitations ofspace and formar, the evidence or arguments I rely on for my views. A note on spelling; it is becoming fashionable to speli Greek names in a scyle closer to the originai, rather than in che Latinised forms to which we are more used. I have done this in most cases, but not with those names ve ry common in their Latinised forms; thus you will find Kassandros, Epeiros and Koinos instead of Cassander, Epirus and Coenus, but I have stuck to Philip, Alexander and Ptolemy ratber than restoring the more accurate Philippos, Alexandros, Ptolemaios. I bave made no attempt to be consistent with foreign oames known cbiefly tbrough Greek sources, since neitber Darius nor Dareios, for instaoce, is really preferable when the originai is closer to Daryavausb.

I am very grateful to PhiJ Barker and Bob O'Brien ofWRG for giving me the chance to write this book, especially as it was Phil's originai version whicb start ed me off on a serious ioterest in warfare ofthis era. Those who bave put me in tbe way ofuseful information are coo numerous to list, but I must express my debt to Thom Richardson, who ove r the years bas belped me with vast amounts ofinformation, especially on tbe Macedonian and lndian froots. I am also above ali, of course, grateful to my parents, for putting up with me for rnucb of tbe time tbis book was struggling imo existence.

Duncan Head July 1982

Copyright

© Duncan Head 1982

IUustruions © lan Heatb 1982

2

ORGANISATION AND COMPOSITION OF ARMIES THE P ERSIANS Persian armies were traditionally heterogeneous multi-national affairs, reflecting the size and varied population of the empire, and including mercenaries from beyond its borders, ofwhom the Greeks were a vita! part ofthe anny.

There seems to have been no real standing army, only quite small guard forces kept by the Great King and the provincia! satraps, plus some garrisons. In wartime these were supplemented by mercenaries and provincia! levies.

Persian numbers are a great problem; clearly their annies were very large, and ancient writers, impressed by this, give vast figures which are often mutually inconsistent and rarely credible, up to 200,000 cavalry and 1,000,000 infantry. Real figures are hard to arrivc at. Tam, working on reliable numbers such as the forces raised in the same area after Alexander's death, suggested Darius lll might have been able to call on 45-50,000 cavalry; Marsden estimated 34,000 at Gaugamela. Perhaps slightly over half ofthese were light cavalry of various sons. The number ofinfantry fielded probably was vast, especially including the half-armed servants who followed thc army in great numbers, but few were much use.

The royal guards included a unit of 1,000 picked cavalry cali ed the royal kinsmcn, and 1,000 infantry called "applebearers" from the omaments of their spearbutts. There may have been other guards, as Diodoros mentions the household troops ofthc palacc as wcll as thc two units above; in 480 the guard contained four 1,000 man units, two ofhorse and two offoot, and pcrhaps this was stili the strength. T hc famous ten thousand Immonals, the core of earlier armies, seem to have disappeared. The satraps, the provincia! governors who combined civil and military functions, kept their own guard units, usually of Persian cavalry- Cyrus the Younger's 600 cavalry in 401 BC are thc best known. Some satraps also had Grcek mercenary bodyguards. Locai garrisons carne under the satraps' control, while they were responsible for raising the levies oftheir province and often commanded them.

Among the Persians themselves, and probably the Medes who were their partoers in empire, there existed a military training corps, the kardakes, open in theory to ali young men. As kardakes were not paid, in practice only the sons ofthe nobiliry or ofother families rich enough to suppon them would belong. T hey did not usually form pan ofthe army, but trained with bows and javelins, accompanied the king when be hunted as this was considered good training for war, and were used as a police force. For the banle oflssos Darius lll called out the kardakes in an attempt to forma reliable infantry force. On graduating from this corps a Persian was liable for service in the army proper, usually as cavalry as these were the uppcr classes. Persis could raise perhaps 3-5,000 cavalry. The lower classes fought as good light infantry, "all ofthcm archers and slingers" as Diodoros says. Peukcstes raised 10,000 such from Persis alone in the Successor wars (he had raised 20,000 for Alexander, but thesc were not ali Persians, some being men ofthe nearby hill tribes).

Basic Persian unit organisation was decimai - Herodotos records officers commanding tens, hundreds, thousands and ten thousands. However Stra bo says the kardakes were organised into fifties, so a company of50 may havc been the basic tactical unit.

Many cavalry were raised from military colonists, who held land from the king, usually outside their ancestral province, in rerum for military service; they were responsible for providing their own cquipment. By the 4th cenrury many ofthese colonists were in dire financial straits, as while their military duties remained constant the burden oftaxes increased vastly, and the produce ofthe land had in many cases to be divided between an increasing number ofheirs. The colonists are best known from documents from Baby Ionia, where settlemcnts of jews, Arabs, Syrians and Indians are known. These colonist cavalry may all have been equipped and fought in roughly Persian style. How far through tbe Empire the system spread is not clear; our evidence is mainly from the relatively advanced western provinces. We know of a seulement of Hyrkanians in Lydia, who probably provided the Hyrkanian cavalry who fought at the Granikos. Xenophon seems to have a low opinion ofthese colonist cavalry, contrasting the 4tb century cavalry with the nobles and their retainers who be believes formed the originai Persian cavalry force. This older system seems stili to have obtained in Bactria and Sogdia, whe re the resistance to Alexander was led by powerful local nobles and their followe rs. This was a less centralised system, with the individuai warriors owing allegiance to their leaders rather than directly to the king. lt may have applied throughout Iran and in areas like Armenia and Kappadokia with Iranian-influenced aristocracies.

Some infantryalso held land from the K ing, at lcast in Babylonia whcre holders of"bow land" seem to have fought as infantry on the same terms as the colonist cavalry; these will be the Babylonian infantry who fought at Gaugamela. Many infantry, however, werc conscript levics or mcrely armed servancs and camp-followers oflittle military value - Lycian camp-followers with javelins are recorded at Cuna.xa in 40 1, for instance. At Gaugamcla Darius mounted some of his levy infantry in an attempt to increase his cavalry strength; perhaps the Albanian,

3

Cadusian and Tapurian cavalry may have been the comingents in question, as these are hill people who would be expected to fìght on foot. Some garrisons, no doubt of somewhat better infantry, were maintained, like the 1,000 Carians, stiffened with I 00 Greek mercenaries, wbo beld Kelainai in Phrygia against Alexander, or the H ebrew garrisons in Egypt, which were bereditary communities ofsoldiers paid by the king rather chan holding land from him. Barbarians from the fringes of tbe empire, or hill cribesmen nominally subject co the Great King but in practice indepcndent, could be hired to fighe for sacraps or che king himself. Xenophon foughc a sarrap of Armenia wicb Mardian and Khaldian merccnaries in his army; Bacis, the governor ofGaza, hired Arabs to defend his city against AJcxander; while tbc good Mardian archers and less reliable Uxii bili tribesmen at Gaugamela were probably similar merccnaries. The Saka and lndian cavalry at Gaugamela were also mercenaries or free allies. The Indians are said to ha ve brought 15 clephants with tbem, while Curtius couples tbem witb tbe "Rcd Sea men", the cri bai levies from the Pcrsian Gulf coast who formed part ofthe rear linc, which suggests the lndians too may ha ve provided some unreliable infantry.

The most important mercenaries, however, were the Greeks, thc backbone of tbe army. There were mosùy hoplites, though a few peltasts may also bave been used, as a few were by Cyrus the Younger in 401. From thc late 5th ccnrury Greeks were regarded as indispensable, whether fìghcing other Greeks, Egyptians, or rebel satraps, and their generals often roseto high command. Ancient authors usually referto 20 or 30,000 Greeks in an army, though tbis may be an cxaggeration. In 350, to fight Egypt, Artaxerxes III received hoplites from his allies in Grcece, 1,000 Thebans and 3,000 Argives (whose commander, Nikostratos, Oamboyantly imitated Herakles by wearing a lionskin and using a club in batùe) and to supplement these he raised 6,000 men from his subjects, the G reek cities of the Asian coast. Thcse would probably bave been much inferior to t he mercenaries, as they would have had little expcricnce; Asiacic Greeks fighùng against Persia in che 390's certainly displayed a marked lack of courage and enthusiasm. The Carians were anotber source of subjecc hoplices.

Scythed chariots wcrc used, 200 ofthem ac Gaugamela. In Babylonia, the land held for milicary service included "chariot land ". As tbc scytbed vehicles were che only warchariots used by Persian ac chis era, it must ha ve been these which were supplicd by tbe holders ofBabylonian "chariot land".

Artillery was in use by the 330's, in many cowns ofthe western provinces - H alikarnassos, Tyre and Gaza used it against Alexander - and also in tbc Iranian hcartland, whcre the pass of tbc Pcrsian Gates was defcnded with catapults in 330. These were probably ali non·torsion bolt·shooting engines (see figure 186).

THEGREEKS The arrnies ofthe Greek city·states were based on a levy ofthose citizens prosperous enough to equip themselves as hoplitcs, close-fìghting infantry equipped at the start of our era with the craditional chrusting spear and large round shicld. In most arrnies they were organised in units called taxeis subdivided into /ocho1: but though the names were more or less standard the size ofunits they designated seems to have varied. ln the mercenary Ten Tbousand of Xenophon about 400 BC, lochoi seem to have been 100 strong, except that one ofthe originally independent forces that made up the army had lochoi of50; Spartan /ochoi seem to bave been larger. Similarly, taxis is used in the Ten Thousand fora body of200 men, while at Athens it mcam one ofthe large tribal regiments of about 1,000 men. It is thus difficult, wben wc bear for instance ofThcban troops manoeuvring in lochoi, to know exactly what size of unit is mcant. In tbc early 3rd ccnrury some states rcplaccd tbc traditional hoplite with thureophoroi with ovai shiclds (see figure 41}, and tbese in rum were supplanted by Macedonian·style pikemen. But the term hoplite could be, and was, used for both oftbese new troop typcs; it mcant little more specific than "an armed man", an infantry· man equipped for dose combat (the word is uscd, for instance, for Roman legionaries and for Cchs by some Greek writers). These infantry were supplemented by cavalry recruited from tbc richer citizens, and sometimes by tbe poorer classes serving as light troops. In the 4th century Thessalian and Boioùan cavalry wcrc stili rcgardcd as the best in Greece; tbe T hessalians kept their reputation, Polybios in tbe 2nd cenrury saying thcy wcrc irresistable when charging in formation, though slow and awkward iftbey had to fight individually. The Boiotians' reputarion declined, however, and by the end ofthe 3rd ccntury the Aitolian cavalry could be described as che best in Greece.

Except for the much more thoroughly trained Spartans, these citizen levies were part·time troops. As youtbs the citizens would undergo a period of basic training, emphasising weapon handling and limited unit drill, but thereafter tbeywere generally called out only in wartime. Tbey were paid while on servi ce. Officers we re frequently clcctcd, and were little more professional than tbeir men. To an extent thìs traditional picture was already changing in tbc 4tb cenrury, tbe picrure ofunrclicvcd amateurism being alte red by the existence of professional drillmasters and "tacticaJ advisers" wbo set out to improve tbc standard oftraining, experienced mercenary officers serving as gencrals for their home states, and elite forces offull·time troops which some cities had begun to maintain.

4

Citizen t roops were supplemented by mercenaries, usually Greeks. Mercenary hoplites were far from unknown in G reece, but were more typical of the armies of "barbarian" rulers who needed to hire good heavy infantry. ln G reece itself cities more oftcn hired specialist peltasts and archers to supplement the citizen hoplites (save that the peltast does not scem to bave caught on among che Grecks ofltaly and Sicily, where mercenary hoplites remained common). Mercenaries wcre generally ben er drilled than citizcn troops, with consequently better tactical flexibility, more experienced and less prone 10 panie. In addition, citizens were ofcen reluctant 10 serve, especially for long periods. However merceoaries could be difficult to contrai when faced with the prospect of loot, even pillagiog friendly territory. Noo-Greek mercenaries were rare io our era (except in Sicily); Thracians had been popular earlier, but Greek peltasts were now available to do their job, while for much of our period M acedon controlled the source of supply. Thracians and lllyrians are occasionally found, though, while Massilia (Marseilles) not surprisingly employed the neighbouring Celts, at least as cavalry.

Many large Greek armies were alliaoces of severa! states, sometimes unstable aod disturbed by argumeots over precedence. The army at Chairooeia in 338 was an alliance of Atheos, Thebes, and severa) smaller statcs; the "Lamian" war, the revolt against Macedon after Alexander's death, was a broad coalition led by Athens, Aitolia aod t he Thessalians, who deserted Macedon in mid-bat tle; while Thermopylai was defended in 279 by an Atheniao generai most of whose troops were Aitolian or Boiotian. In the 3rd cenrury the small city states tended io any case to be overshadowed by larger federai states such as the Achaian and Aitolian leagues.

Boiotia Tbc Boiotian citics formed a league which in the mid-4th century was domioated by its largest member, Thebes, whose power was binerly resented by ber neighbours. The constirution in force till 387 had divided Boiotia into elevcn districts, each electing one boiotarchos who sharcd in the rnilitary command, and providing 1,000 hoplites and I 00 cavalry; four ofthe eleven districts were taken up by Thebes and those cowns directly subject to ber. In our period this constitution no longer applied but command was stili in the hand of elected boiotard1oi and che available manpower was similar; this is confirmed by the 10,000 Boiotiao hoplites who served in the allied army at T hermopylai in 279. Field armies were usually smaller - 7,000 hoplites is a figure ofien fouod in thc 4th century.

T here were only 500 cavalry ai Thcrmopylai, bui tbc proportioo was often higher-800 to 5,000 foo1 ai the Nemeia in 394, 2,200 foot aod 1,300 cavalry in 313. This latter is higher thao the earlier establishment figure for the whole Lcague. In che 4th century Boiotian cavalry wcre highly regarded, and in the era of Thebao dominaoce aficr Leuktra in 37 1 their hoplites too were experienced and coofident; before Chaironeia 1hey wcre called the besi soldiers in Greece.

There was an eli1e T hebao corps of 300 hopli1es, che Sacred Band, fulltime soldiers garrisoned on che Kadmeia, Thebes' ci cadei, and said 10 be composed of pairs of lovers. Command of che Sacred Band was anothcr clective office. At Chairooeia, where the corps were wiped out, they fought as che Theban front rank, but at some other baules they seem to have acted as a separate body. Boiotian light infantry had earlier been good and numerous; they are not often meotiooed in the 4th century, but certaioly existed.

When Alexander attacked Thebes the Boiotian cities of Plataia, Thespiai and Orchomenos eagerly fought against ber. Deprived ofthese allies, Thebes armed foreign residents, refugees and freed slaves, forming these into a body which fought separately from the citizenry. The city was destroyed, aod though rebuih ii never regained ics former dominance. The absence ofTheban ambi1ions left che Boiotian league a much more stable organisation, but its military reputation dedined.

Tbc 3rd ceotury Boiotian army is well documented by a series ofinscriptions, and provides ooe ofthe best examples of the chaoges in G reek infantry armamcnt. Soon after che Galatian invasion of 279 Boiotia re-equippcd ber infantry with the thureos (sce figu re 4 1); thcsc thureophoroi are stili sometimes cali ed hoplites in inscriptioos. About 245 the 1/111reophoroi are in rum replaced by peltophoroi. These were probably not che traditional peltast, as the name might at flrst suggest, sioce a switch to the old pelte is hard to uoderstand when the tl111reos was as suitable, if not more so, for the peltasts' role, and it would be out ofkeeping with what seems to be the generai trend in Greece. But the Macedooian shield is somctimes callcd pelte, and the unusual form peltophoroi may be used because the more precise term peltast is inaccurate. lt seems that the Boiotians have adopted Macedonian equipment, the first independent Greek state to do so. Certainly they became an ally ofMacedon after 245, and the presence of an agema in their army reinforces Macedonian parallels. Inscriptions recording che annua! conscription ofyouths for military training suggest a strengt.h ofabout 17,000 men in the early 3rd cenrury, falling to 13,500 later in the century when peltophoroi were introduced, perhaps because the Macedoniao-sryle phalanx needed more drill en masse than the more open and individuai style ofthe thureophoroi, and maoy ofthe poorer citizens could not afford tbc time to crain, so do oot appear in the records.

5

The army was still commanded by the boiorarchoi in the 3rd cenrury but they were replaced by a single federai straregos, along the lines of other leagues, in the 2nd. lnscriptions reveal other officers and give a good idea of the composilion of the army; each city elected onc or more officers of peltophoroi (apparently only one in peacetime, more to lead the troops to war); officers for two elite infantry units, the epilektoi (picked men) and the agema; probably one was a standing force, thc other the pick ofthe part-time citizen soldiers, called out when the full levy was not needed; a hipparchos to command the city's cavalry and ilarchoi to lead each squadron (Thespiai, whicb under the early 4th century constitution had provided 200 cavalry, now had four ilarcho1); a 1ara111i11archos in cbarge of the new Tarantine cavalry (see figure 44); an archcr officcr; a slingcr officcr, a post whicb was only fillcd in warrime, perhaps because tbe slingers were mercenaries only hired then; and k1111agoi, rcsponsible for the guard dogs which patrolled the city walls.

Sparta Sparta was well past her peak in tbis period, having lost both manpower and territory aftcr Lcuktra in 371, and no longer able to cali on large forces of allied troops from thc Peloponncsian statcs. True Spartan citizens were a minority caste, the Spartiatai; the bulk ofthe population ofLakonia, Sparta's province, were helots, serfs who only served in tbe army as servants perhaps fighting as light infantry. There were lower grade ci1izens at Spana itself, 1he hypomeiones who had been demoted from full Spanan status, perhaps becausc of poverty, and the 11eodamodeis, helots given a low grade citizenship in return for mili1ary scrvice as hopli1es. They may have shared the rigorous boyhood training and discipline which made Spartan hopli1es so famous and effective. The citizens ofthe other towns ofLakonia were perioikoi, obliged to figh1 for Sparta, and probably supcrior to most othcr citizen hoplites, though lacking full Spartan training.

The number of full citizens had long been in steady decline, sapping Spartan military power, and the use of 11eodamodeis was one response 10 tbis decline. Aristotle quotcs less 1han 1,000 Spartiata i (probably at the time of Leuktra) and Plutarch 700 about 250 BC. These ci1izens were organised aftcr Lcuktra in 12 lochoi. Previously the /ocl1os had been about 150 strong, a subdivision of a mora of 600. Now i1 seems three lochoi were rcgarded as equivalent 10 an old mora; perhaps the strength had risen 10 about 200, but perhaps declining manpower meant that three /ochoi totalling 450 had to be used where previously a mora of600 would ha ve been appropriate. lflhe internal organisalion ofthe /ochos was unchanged, each one was divided into two pemekostyes, each oftwo e11omotiai, each of 36 men, a total for the locbos of 144 men, plus officers. Whcther 150 or 200 was the strength ofthe new lochos, 12 of them clearly need more men tban 1,000 Spartiatai could provide, so neodamodeis and perhaps hypameiones must havc been included (thougb the former bad earlier formed separate units). Therc were enough neodamodeis to send 1,000 to Egypt with Agesilaos in 362, so there were probably enough ofthem to make up the 12 lochoi, making it unnecessary to assume that perioikoi were incorporated in the lochoi; thcse must bave been organised separately. lt is wonh noting tbat any estimate ofthe strength oftbe loc/1os can only be approximate, because the strength fielded would vary depending on the extent ofthe callup. The Spartan army was lcvicd on thc basis of age-groups, in terms ofso many years above military age (which was 20). Thus on a 35-year callup men from 20 to 55 would serve, on a 40-year cali up men from 20 to 60, with a corresponding increase in the sizc of each lochos. The younger age-groups could be ordercd to cbarge out from the ranks against a mobile enemy, and so werc probably lighter armed; such must ha ve been the "lightest equipped" men wbom King Areus took in pursuit of Pyrrhos from Sparta to Argos in 272, along witb Cretans and apparemly the royal guard, ifthat was the eli1e force Pyrrhos wiped out in one clash on the march. This guard may have been 300 strong.

Spartan cavalry was traditionally not oumerous (600 men in 394) and poor quality, because whilc rich Spartiatai were obliged 10 maintain the horses and arms, the riders were substitutes, often unfit, who only met their horses and equipmcnt on thc eve of a campaign. Xenophon claims the standard had been improved by the use of mercenaries, perhaps to replace the unwilling substitutes.

At this stage Spartan armies rarely contained allies, as her old unpopularity persisted, but Agis II had Peloponnesian allies for his war witb Macedon in 330, and Messene probably helped Sparta against Pyrrhos. Nor could Sparta afford many mercenaries except a few Cretan archers; but Agis Il was subsidised by Persia, and ofhis 20,000 foot and 2,000 borse, I 0,000 were mercenari es, some veterans oflssos.

The decline in Spartan manpower led to radical reforms. In 242 King Agis IV proposed a redistribution ofland to create plots for 4,500 Spartan citizen soldiers and 15,000 perioikoi, the Spartans to be organised into I 5 units of200 or 400, and the numbcr 10 be rcache~ by enfranchisingperioikoi and foreigners. Agis claimed this was a return to the old ways; to others it looked more like a socia I revolution and Agis was assassioated before be could carry it out. But Kleomenes III in 227 carried through a similar reform, restoring the old stem training and discipline, which had apparently lapsed, and giving citizenship to 4,000 perioikoi who were re-armed witb the Macedonian sarissa (Sparta 6

seems to bave gone straigbt from traditional hoplite equipment to tbe Macedonian style without ever adopting tbe thureos). He later added anotber 2,000 freed belots also with sarissai. Plutarch confirms the tota!, saying 6,000 Spartans fougbt at Sellasia in 222 out of a tota! of 20,000. Also present were about 5,000 mercenary infantry, probably 1,000 cavalry, and 8,000 perioikoi (apparently unreformed so probably with traditional hoplite arms) and allĂŹes. Kleomenes could only afford the mercenaries because ofsubsidies from Egypt, and even then bis resources were strained. His mercenaries included Tarantine cavalry and Cretan arcbers, wbile many if nor ali bis other mercenary infantry were ligbtly armed.

Despite the destruction ofthe citizen phalanx at Sellasia, Spartan citizens continued to fight in Macedonian style under the later "tyrants"; the phalanx is now treated as one body, implying ali the citizenry are now armed in tbe same way, including the perioikoi. The rulers bowever placed mucb more reliance on their mercenaries, probably for both military and politica} reasons, Machanidas in 207 leading them in preference to tbe citizen phalanx. Tbey stili included Cretans, Tarantines and many ligbtly armed infantrymen. Nabis, figbting. Rome in 195, bada citizen levy of 10,000 (probably including perioiko1), 3,000 mercenaries and 2,000 Cretans, later joined by anotber 1,000 mercenaries and 2,000 Argive citizens (plus even more mercenaries in garrisons). Again he put all his confidence in his mercenaries (wbo included Aitolians) some of whom formed his bodyguard. After Nabis' death Sparta was forced into tbe Achaian league.

Athens The citizens of Athens were organised into ten (later twelve) tribes, wbicb were the basis of tbe military organisation. Command was given to ten generals, stratego1~¡ at the start of our period they may stÏll bave been elected one from eacb tribe, but wben Aristotle wrote bis Co11stitutio11 o/ the A1he11ians later in tbe 4th cenrury election was open to ali regardless of tribe, to encourage the cboice of able and experienced men. This may ha ve been one ofthe reforms inspired by the Atbenian army's poor showing in the Chaironeia campaign. One strategos was responsible for expeditions abroad, one for home defence, two for the port of Piraios, one for the fleet; che other five had no specific duties. Even this was an advance on the earlier system whereby ali ten shared command. By the 2nd century they all had specific jobs; one "generai ofthe hoplites" for foreign campaigns, one for home defence, as before, but now three for the fleet, three for tbe Piraios, one for Eleusis and one for provisioning. In Aristotle's day there was al so a hipparchos for the cavalry garrisoning Lemnos, but later Lemnos and Salamis each had a generai and a hipparchos, wbile Imbros and Skyros bad a generai each. Also elected were a ta.xiarchos to command each tribe's infantry, a phylarchos for each tribe's cavalry, and two hipparchoi who had overall command oftbe cavalry.

By the 2nd century, a taraminarchos was added to lead the "Tarantine" cavalry; these took pan in state games organised on a tribal basis, so were citizens, not mercenaries as the first "Tarantines" had been.

Each tribe provided a taxis ofhoplites, subdivided into lochoi of unknown size, and a cavalry unit called phyle afcer the tribe itself. Athenian ligbt infantry were notoriously poor in numbers, training and equipment, as tbe lower classes generally served in the fleet; only a handful of javelinmen were at Mantineia in 362. The citizenry bad become very unmilitary in the 4tb cenrury and relied wberever possible on mercenaries - "lt is quite certain tbat during the greater pan ofthe 4th century Athens had used mercenaries more freely than any other Greek city" to quote G. T. Griffitb - and the xenologos, commissioner for tbe mercenaries, was a regula r post. Some foreign expeditions consisted entirely of mercenaries, sucb as tbe 150 borse and 4,000 peltasts sent to aid Olynthos against Macedon in 349. Even Demosthenes, tbe great proponent ofcitizen armies, urging Athens to send 2,000 foot and 200 borse abroad to figbt Macedon, only asked tbat a quaner be citizens. Xenopbon, writing The cavalry commander about 360 says that Athenian reluctance to serve applied even more to the cavalry, because of the expense of mounted service. The cavalry sbould bave mustered 1,000 men, a phyle of I 00 from eacb tribe, but was badly under strengtb; be recommends recruiting 200 mercenaries to bring tbe force up to establishment and raise its standards. He says tbe ten phylai should drill in two groups of five, implying tbe force usually fougbt in two groups, probably eacb hipparchos leading balftbe cavalry on one flank ofthe foot. Within tbe phyla1~ every ten-man group was led by a dekadarchos; Xenophon says tbese formed the front rank, implying a deptb often ranks, wbich is deeper tban Greek cavalry usually fought. Eacb phylarchos had a second-in-command so tbe phyle could operate in two bodies.

Tbe inexperience of che citizens cost Athens dear at Chaironeia, and as a result a form of conscription was intr oduced, and ali free-born citizens were given military training. lt bad always been the tbeoretical obligation of ali men who could afford boplite arms to train and fighe as sucb; now tbe obligarion was extended to poorer men as well, and tbe training appears to have become more thorougb. Lists of conscripts during tbis period suggest about 12,000 men aged from 20 to 50 might bave been available, at least in theory, but when seven of the ten taxeis marcbed out to Thessaly in the Lamian War of32 I, they mustered only 5,000 men (accompanied by 500 borse and 2,000 mercenaries).

7

Aristotle says the conscripts (epheboi) would be called up at 18, and issued with a spear and shield by the state. For one year they would be in garrison in the Piraios, training with hoplite arms, bows, javelins and catapults; in the second year they would be sent to garrison the border forts and trained in route marches. After the Greek defeat in the Lamian war, however, citizenship was drastically reduced, and conscription confined to the remaining citizen class. About 300 it was abolished completely and the ephebic corps fell from 800 (i.e. an intake of 400 per year on the reduced franchise) to 30, losing ali military significance. The whole population was stili prepared to rally round ifthe city itselfwas attacked (as in the siege by Demetrios in 295 or that by Sulla in 86) b ut generally Athens was contene with a small force of mercenaries and citizen volunteers, who seem also to have been professional soldiers (either an elite part ofthese, or the whole citizen volunteer force, was known as epilekroi), kept in various garrisons, chiefly in the border forts. In 200 when Philip V of Macedon attacked the city its army was assisted by a Pergamene garrison, probably also mercenaries.

We have no record of Atbens' hoplites adopting either the thureos or the Macedonian arms. This may simply result from gaps in the sources, but it may be that Athens, no Ionger aspiring to great power status in the 3rd cenrury, saw no need to update her armed forces.

Phokis Phokis was usually an insignificant power, but in the "Sacred War" ofthe 350's, under what was nominally a series of elected generals, but in practice not far from a dynasty ofhereditary tyrants, was able to field large armies by using the sacred treasures of Apollo's sanctuary at Delphi to hire mercenaries and subsidise allies. The conringem of500 borse and 3,000 foot sent to Thermopylai in 279 was probably not far from the full ciuizen muster. DurĂŹng the Sacred War there was an e lite force of 1,000 picked citizen peltasts; something of a rarity, but many of the Phokian citizen troops were probably similarly lightly armed, as they are recorded climbing precipitous slopes to hurl javelins at the Galatians in 279. Phokian citizen hoplites certainly existed, though.

Mercenaries in the Sacred War were paid higher rates than usual to tempt them into ignoring the sacriligious source ofthe funds; the result, according to our not unbiased sources, was an even rougher lot than mercenaries generally were, "the worst knaves, and those who despised the gods". Mercenaries brought the total force to over 20,000, and the tactics used aganst Philip of Macedon in T hessaly suggest there was stili a high proportion oflightly equipped infanrry, perhaps mercenary peltasts. Phokian armies may have been very short of cavalry; Onomarchos had only 500 borse to 20,000 foot in bis last battle against Philip, although he was assisting the Thessalian tyrant Lykophron ofPherai, who would be expected to have a few horse ofhis own. Onomarchos had earlier had a large force ofstonethrowing engines, probably early non-torsion models (see figure 186), which he used to g¡ood effect against Philip.

The anillerymen, too, will have been mercenary professionals. Subsidised allies included Athens and Sparta; after Onomarchos' death his successor Phayllos was helped by 5,000 Athenian citizen foot and 400 horse, joined by 1,000 Spartans and 2,000 Achaians, plus the now exiled Lykophron of Pherai and 2,000 ofhis own mercenaries.

The Aitolian League Aitolia was a poor country, the people living in villages not cities, and fighting as light javelinmen. Bue in che early 3rd century they re-organised themselves imo cities (which must bave made it easier to assemble the citizen troops for drill and training) and thereafter the League expanded rapidly umi! it included, as members or allies, most ofthe small states of centrai Greece. The army was commanded by a strategos elected every autumn.

Aitolia was unusual for the peri od in making next tono use of mercetiaries (the only case recorded is 1,000 Cretans hired from Knossos in 218) both because they were themselves "a more warlike people than Greeks in generai tend to be" (Livy) and because they were voor; Aitolia exported men as mercenaries rather than hiring more. Philip V of Macedon observed that Aitolian mercenaries were often found in opposing armies, and that tbe League wou ld unofficially allow its citizens evento take service against its own allies. Aitolians were also noted raiders and pirates, greedy for loot, but sometimes rash and ill-organised in their foraging; they seem in generai to bave been poorly disciplined. Citizen manpower may bave totalled 20,000 men (Griffitb's estimate); the largest force fielded seems to have been 12,000 infantry and 400 cavalry in Thessaly in 321. There were 7,000 hoplites, 790 light troops and a few cavalry at Thermopylai in 279; the home levy active at the same time was predomioanrly javelinroeo. The army included an elite infantry unit of epilekto1; led by an epilektarcl1os. The proportioo of cavalry seems to ha ve beeo low, even by Greek standards; though there were 400 horse to 3,000 foot (plus 500 ofthe Cretans) in ooe force in 218, tbere were ooly 400 to 6,000 foot allied to Rome in 197, 200 horse to 3,000 foot in Thessaly helping Antiochos m. The 1,000 cavalry sent to help Nabis ofSparta in 192 may have been dose to the League's whole mounted strength. By 200 Aitolian cavalry were reputed to be the best in Greece, especially dispersed and io single combat. Ali we know ofunit orgaoisatioo is a refereoce to oulamoi, small squadrons ofuncertain strength.

8

Despite the high proponion of hoplices at Thcrmopylai, Polybios in the 2nd century says several times that Aitolians wcrc ncither equipped nor trained for dose fighting on lcvel ground, but were more at home in rough cerrain. Either hoplites had disappeared by the time hc wrotc, or there were so few that thcy wcrc not significant, or else Aitolian "hoplites" were lighter than othcr nations'. lt is tempting to see Aitolian hoplites becoming 1hureophoroi, as others are known to have done; it would fit the terrain and tbeir craditions, and explain Polybios' comments. Some oftheir allies certainly had heavier troops; wc hear of sarissai in a joint Aitolian-Ambrakiot force in 189, but Ambrakia had been Pyrrhos ofEpeiros' capitai, and be will have introduced the sarissa, so this may cell us nothing about Aitolian equipment.

Aitolian troops frequently fought in alHance with Elis against the Achaian league; for inscance in the winter of218/7 the Aitolian generai Pyrrhias raided Achaia with 1,300 Aitolians, 1,000 Eleian cit izen foot and 200 oftheir horse, and 500 mcrcenaries in Eleian pay. Elis' troops included Tarantine cavalry. In the early 4th century she had possessed two picked infantry forces, " the three hundred" and "the four hundred", but there is no evidence that these stili existed. Whether Eleian citizen infantry had retained the old hoplite equipment or bcen re-armed is not known.

The Achaian League The hitherto unimportant league of che small towns of Achaia began to expand in 251 when Aratos of Sikyon deposed his city's tyrant and annexed it to Achaia. Thereafter, under Aratos and then Philopoimen of Megalopolis, the league expanded to become one of the major powers of Greece, embracing most and briefly ali of the Peloponncsc. Achaia proper could probably not raise much more than che I 0,000 men Aratos led in 245; but as tbc league expandcd it could field 20,000 against Spana in 228, and talk of raising 30 to 40,000 in 170, though that was nevcr donc. For the last battle with Romc, the League raised only 14,000 foot and 600 horse, even after supplementing the citizens with freed slaves, but this was a last-ditch levy possibly supplemented by survivors of an earlicr defcat, while many ofthe rich may have beld alooffrom the radical leaders ofthe time. Tbc Acbaian army was commanded by stracegoi clectcd annually in carly summer, at tbc start ofthe campaigning season. T ill 225 two S1ra1egoi wcrc choscn, thereafter one with a deputy called hipparchos (master ofhorsc). Each town had two apoteleio1~ commissioners responsible one for the town's cavalry, one for che infantry. Subordinate field commanders called hypos1ra1egoi also appear.

Early military organisation seems to have been patchy, a generai levy oftbe citizens with mercenaries hired when needed. In 219 however the League found itselfunable co hire any mercenaries, bcing unable to pay thosc who had servcd in che recent war with Sparta. Aitolian raids harrassed tbc western citics, and the citizen levy was presumably too cumbcrsome to dea! with them, as the cities ofDyme, Pharai and Tritaia raised a "private" mercenary force of 300 foot and 50 borse, despairing ofthc League's organisation. As a result ofthis display offederal impotence, in 217 the League resolved to maintain a standing army of8,000 mercenary foot and 500 horse, plus a picked citizen forcr- of3,000 foot and 300 cavalry. The picked citizen force may bave existed before this reform, at least on paper, as che cl ice force sem to Sellasia in 222 was ofthe same strength. Tbe full citizen levy could stili be ca.lled out when nccdcd. At this stage citizen infantry were lightly armed as 1hureophoroi (cvcn apart from citizen skirmishers, who included archers and slingers). PhalangĂŹtcs are mcntioned at Kaphyai in 220, but may not bave been Achaian, as che Maccdonian generai in the Peloponnese was there with his troops. In 222 1,000 Megalopolitan exiles were armed as bronze-shielded phalangites by Macedon; they fought in this gear at Sellasia, marched to but missed Kaphyai two years later, and 500 ofthem served in the standing army of217. In 208 Philopoimcn reformed the rest ofthe citizcn infantry with Macedonian equipment. Achaian "peltasts" are stili occasionally mentioned after this, so some citizens may have kept light equipment.

Philopoimen also reformed the hitherto inefficient citizen cavalry, recruited like other Greek borse from che nobles and che rich. They were organised in /ochoi, whicb bere probably means "files" as it does in the tactical manuals, no doubt of 8 men, grouped into dilochiai, double-files of 16; ou/amoi perhaps of 32; i/ai perhaps of 64, though Asklcpiodotos' manual says Grcek i/ai wcre of 128; hipparchiai ( 128?); and syntagmai (256? or perhaps one city's contingcnt, and thus ofvariablc sizc?).

Merccnaries were various; in 208 thcy included lightly-armed infantry, "Tarantine" cavalry, Illyrians, and probably the 1horakilai (figure 42). There were also thorakilai among the 3,000 foot and 500 borse at Kaphyai in 220, who were probably ali mcrccnarics as the citizens bad bcen dcmobilised. Livy mentions Tarantincs again in 192, along with Cretans, and in Philopoimen's last campaign in 183 Crctans and Thracians appear, thc lattcr probably tbe same as the Thracian cavalry mentioned in 197.

9

T a r as Taras (Roman Tareorum, modem Taranto) was Sparta's only overseas colony, and in 1he 4th century the most powerful of the Italiote Greek cities - many of her neighbours had been wcakened or destroyed by internecine warfare or by the expansion ofthe native Oscan peoples. Taras had a fairly effeciivc democratic constiruùon which provided internal stability, and dominated 1he Italiote cities' usually ra1her ineffective defensive league. Sbe aimed 10 expand into Apulfa, brieOy succeeding aftcr 334.

Apart from a powerful Oeet, Strabo says Taras could mobi lisc 30,000 infantry, 3,000 cavalry and 1,000 hipparchoi.

This was probably only a paper strength, but more than 20,000 men could certainly be fìelded. Hipparchoi usually means cavalry officers, but must bere denote some elitc cavalry unit. lf most of Taras' cavalry were the light skirmishers known as T aramines (see figure 44) the hipparcl1oi may have been heavy cavalry. The citizen infantry in Pyrrhos' time are called a white-shielded phalanx. This does not prove wha1 their equi pment was, as traditional hoplites formed a phalanx and could have whi1e shjelds. But 1he similari1y ofthe name /eucaspis phala11x to the /e11caspides of 1he Macedonian phalanx suggests that either Pyrrhos or Alexander of Epeiros had re-armed the Tarantine, like the Epeiro1e, infantry in the Macedonian style. Apart from citizens T aras could afford 10 hire substantial mercenary forces, and may have kept a standing mercenary force in peacetime; Agathokles, future tyrant of Syracuse, served as a mercenary ofT aras when an exjle about 320, in time of peace.

The citizens' military qualities dedined during our period, and Pyrrhos found them lazy, luxurious and unwilling to fìght. This decline, as well as tbe ceaseless pressure of the Oscans, help explain Taras' reliance after 350 on foreign condottieri. In 343 King Arcrudamos ofSparta landed with a Spartan and mercenilry force; he was hllled in baule, and his men surrounded and wiped out with javelins, by the Lucani. In 334 Alexander, king ofEpeiros and protegé of Philip of Macedon, began a series ofsuccessful camprugns, overrunning most of Apulia and defeating the Lucani, Brunii and Samnites. But Taras withdrew suppon, fearing his ambitions for persona! power, and he was surrounded by thc Lucani. Trying 10 break out, he killed the Lucanian ge nerai in single combat but was killed himselfby one ofthe 200 Lucanian deseners in hjs army. Next was Kleonymos, an adventurer ofthe Spartan royal house, who brought 5,000 mercenaries from Greece, recruited 5,000 more in ltaly, was given 20,000 Taran tine citizen foot and 2,000 horse, and was joined by other Italiote citi es and the Messapii. Again he was rejected by Taras as he was too ambitious.

Finally in 280, when Rome was che enemy, T aras called io King Pyrrhos ofEpeiros to head a grand coalition. He imposed conscription and a harsher discipline thao 1he T araotines were used to, retraining the army. After his withdrawal T aras was soon caprured by Rome, and stayed in Romao bands except for revolting to Hannibal from 212 to 209. But now ber troops were ofvery low calibre, military organisation and training probably having almost entirely lapsed; " they had no staying-power, being inferior to the Romans both in weapons and tactical skill, not to memion detcrmination and sheer physical strength". Some ofthe infantry now fought wich javelins.

Syr acu se Chief of 1hc Grcek cities of Sicily, Syracuse alterna1ed shon-lived democratic and oligarchie regimes with autocracies. Tyrants were able to gain and keep power because ofthe need for strong authority to organise resistance to Carthage, and often dominated the other Sicilian cities. Most successful ofthe tyrants in our era was Agathokles, who ruled from 317 to 289 (from 304, as hlng). H ieron (ryrant in 269, hlng in 265) took Syracuse into alliance with Rome in 1he fìrs1 Punic war; in che second, the city joined Canhage in 215 only to be captured and sacked by the Romans in 212, dcspitc thc famous cfforts of Archimedes in her defence.

Syracusan regimes of wha1ever complexion depended heavily on mercenaries for an effective fìeld army against Carthage, as the ci1izen troops were inexperienced and unreliable, even when fìghting in a cause they supported; Dion, fìghting the 1yrant Dionysios Il in 357, was nearly beaten wheo hls citizen troops panicked in the face of a surprise auack, and was only savcd by his more reliable mercenaries. lt was similar elsewhere in Sicily; Diodoros contrasts the ci1izens of Akragas "in morale far inferior since the citizen army had been formed with indulgcnce and in a sheltered way oflife" with 1he mercenaries ofSyracuse "trained in military service and in constant campaigns."

While the democrats relied mostly on Greeks for their mercenaries, the tyrants imitated Carthage by hiring a wide variety of foreigners, less likely to sympathise with a possibly hostile citizenry. In 341 the democratic leader Timoleon had 4,000 mercenari es ( 1,000 deserted before the baule ofKrimjsos; some werc vcterans of the impious Phohlan looting of Delphi), 3,000 citizens, and 5,000 others (allies or perhaps additional mercenary units). Sicily stili seems 10 have relied on 1hc tr:idiiional hopli1e, but many ofT imoleoo's troops were lightly armed infantry, so some ofthe men he had recruited in Grecce may have becn peltasts.

IO

The tyrants could use foreign mercenaries for the peltasts' role, the barbarians being led by their own officers and armed in their native styles. The army with which Agathokles invaded Africa in 31 Ois rypical ofthe tyrants' mixed forces; a guard of 1,000 picked mercenary hoplites, 3,000 more Greek mercenaries, 3,500 Syracusans, 2,500 unspecified infantry (Sicilian allics?), 3,000 Samnite, Gallic and Etruscan mercenaries, and 500 archers and slingers. The figures conceal some cavalry, as these crossed to Africa with saddle-pads and bridles, hoping to get horses locally. By 308 at least 800 had done so. Agathokles also had siege engineers with him, using catapults 10 besiege Utica in 307. Ophellas, Macedonian governor ofKyrene, carne to his aid with 10,000 infantry, 600 cavalry and 100 chariors (mostly Greek mercenaries); Agathokles killed him and took over his army. Most may have deserted, because in 307, despite reinforcements from Sicily, Agathokles had only 6,000 surviving Greeks, 6,000 Gauls, Samnites and Etruscans, and 1,500 cavalry. Bui he had been joined by che locai Libyans to che cune of 10,000 infantry and over 6,000 chariocs, and carlier by some Numidians. None ofthese native allies were very reliable, however; in one clash Numidians on both sides withdrew, leaving Greeks and Canhaginians to fight it out, while the Libyans "only sat and looked on, being always rcady to change witb changing conditions".

Other mercenaries not mcntioned in this war were Sikels, the natives ofinland Sicily; Agathoklcs seized power in 317 with 3,000 men most ofwhom were Sikel mercenaries who had previously fought alongside him in the service of the Syracusan democracy. Ligurians are mentioned once, when Agarhokles massacred 2,000 Ligurian and Ecruscan mercenaries whom he could noi afford to pay. This was a common problem; a less drastic solu1ion 1han massacre was to give discharged mercenaries land to settle on, which Dionysos I had done with numerous Oscans.

His son Dionysios II had numerous Oscan mercenaries from Campania in the 350's, one of his gencrals being Nympsios from N eapolis in Campania, who has an Oscan name even though his city was stili chiefly Grcek. The famous Mamertini who seized M essana in 288 were Campanian mercenaries of Agathokles who had been sent home after his death. Iberians, used by Dionysios I, are not meoĂšoned in our era till 214, when we hear of African and Spanish troops who may in fact be from the Canhaginian army. At this stage thc Syracusan army could field 19,000 men, citizcns and mcrcenaries, including Cretan archers.

MACEDON: PHILIP AND ALEXANDER Before Philip T he early Macedonian kingdom fell naturally into two parts; "lower" or coastal Macedonia, ruled directly by the kings of thc Argead dynasty, and "upper" Macedonia, the tribal highland areas such as Orestis, Lynkos and Elimaia, rulcd by semi-independent dynasries which occasionally acknowledged the supremacy ofthe Argeads but were often al open war with them. Macedonian arrnies were fairly primitive. They were built round a core of excellent cavalry provided by the king's retainers (hetairoi, Companions) and nobles, bue there were only a few hundred of these. The infantry were an ill-armed disorganised levy of lowland peasan1s or highland herdsmen, probably mostly armed as peltas1s with a few archers. Sometimes Greek hoplites were available, from allied powers or the Greek scttlements of1he coast. Alexander I (498-454) is credited with an attempt to form an effective infantry bodyguard, 1he pezetairoi (Foot Companions, the name later given by Philip to his new phalanx) while Archelaos (413-399) may also have 1ried an infan1ry reform, but neither seems to have lasted. Occasionai highland princes may have tried to form good infantry forces, as Arrhibaios of Lynkos had some hoplites in 424. The lack of effectivc centrai authori1y or a strong army meant that Macedon was continually subject to civil war, Thracian and Illyrian invasion, and Greek intervention. Philip II (359-336) changed ali thar.

The reforms of Philip Il Philip successfully imposed royal authority on upper Macedonia, and locai separa1ism was hencefo rth a negligible factor. His creation of a national army was probably imponant in providing a focus of na1ional sen1iment and unity, as much as giving him the power to enforce ir. Many members from the highland dynasties held high command under Philip and Alexander, sometimes leading men from their own province, bui apparently conteni 10 be pan of the royal army ra1her than quasi¡independent lords. Their sons were al coun in the corps of royal pages, paides basilikoi, serving both as hostages and as the next generation's commanders in training.

Early in 358 Philip could only field I 0,000 infantry and 600 cavalry; in 352 this had risen 10 20,000 foot and 3,000 horse, including Thessalians and probably mercenaries. At Chaironcia he had 30,000 foot and 2,000 borse (not needing many cavalry to guarantee superiority in Greece). By the end ofthe reign, to judge from Alexander's forces in 334, therc were 24,000 Macedonian foot and perhaps 3,900 cavalry under arms, not 10 mention other nationalities.

This new army, financcd by the rich mines of Mr. Pangaion, was trained and disciplined to high standards. The

Il

men were tougbened by peacetime training including route marcbes io full kit, wbicb seems to bave been an innova¡ tioo, by banning women in camp, baggage wagons, and such weakening customs as hot baths. In contrast to the mobs of non-combatants who followed Greek and Persian armies, Philip limited servants to one groom per cavalryman, and one servant per ten infantry, carryiog essentials like handmills to grind graio. Even these seem to have fought; we see grooms at Gaugamela, behind tbe infantry, dealing with Persian cbariots. Incentive to serve under these demanding conditions was provided not just by pay and loot but by a developed pay structure which not only gave extra pay for promotioo to file-closer or file-leader but also for distioguished service. A deep respect and affectioo for the king was soon added.

Most ofthe cavalry were provided by enlarging the old corps of royal companions, which was thrown open not just to the old Macedonian nobility but to "new men" of Greek as well as Macedoniao origin, dangerous immoral ruffians according to ooe not uobiased Greek writer. The new men were given large estates io oewly cooquered lands, ensuring their loyalty to the king. Only 800 strong at one point in the reign, there were 3,300 Compaoion cavalry at the end.

The bulk of the infantry formed a new sryle phalanx, by 334 apparently organised in 12 or 15 caxeis of about 1,500 men, each caxis raised from one province. They were grouped in files called dekades, tens, so may have fought ten deep under Philip. Diodoros ascribes the organisatioo ofthe phalanx and the development ofits dose formati on and distinctive long pike, the sarissa, ali to the early months of Philip's reign, but there is reason to suspect that the sarissa may not have beeo iotroduced at once (see note to figure 32). The iofantry were called pezetairo1~ Foot Companions. The judicious extension ofthe Companion title seems to ha ve been ooe oftbe steps used to emphasise attachment to the crown. (Some infantry were called aschecairoi - town companions? - an obscure title of uncertain significance.)

There were also 3,000 hypaspists, "shield-bearers", a guard iofantry unir in three units of 1,000, one ofthem being the agema or elite battalion. The oame hypaspists suggests they may have originated in shield-bearers or retainers ofthe cavalry. They were the ooly native Macedonian force oot organised and raised by province, no doubt to foster their loyalty to the king and counterbalance any ideas ofprovincial separatism io the rest ofthe army. As guard infantry they filled the role of Alexander l's old pezecairo1~ who (ifthe corps stili existed) might conceivably have provided the nucleus ofthe new pezerairoiphalanx, the hypaspists then being raised to replace them as guards; but this is hypothetical, and the old pezecairoi may alternatively have become hypaspists.

While these forces were vita! to Philip's victories io pitched battle, for srnaller operations, raids aod interventions in Greek states' politics, he relied on Greek mercenaries. "You hear tbat Philip goes where he pleases" said his enerny Demostbenes "not by marching his phalanx ofinfantry, but by brioging io his train light infantry, cavalry, archers, mercenaries and other such troops". Mucb more ilexible than citizen uoops, mercenaries could be put into the field at any time ofthe year, which was essential to Philip's way ofmaking war; "summer and winter are alike to him ... he has no particular season for rest". He may have had 6,000 or so mercenaries at Chairooeia. Finds from the siege ofOlynthos confirm he had archers aod slingers, probably merceoary.

Also essential was his effective siege train, the best ofits time. His engioeers may have made major advances in catapult development (see note to figure 187). Philip's series of quick, successful sieges in Chalkidike and the nearby coasts cootrasts sharply with the tedious, long drawn out nature of most Greek siege warfare.

Thessaly Thessaly was very like Macedon - only pan urbanised, dominated by the nobles and their excellent cavalry, with traditions of unity but usually divided. The post of ragos, elected ruler and milita ry commander of ali Thessaly, was recogoised, aod uoder a strong cagos Thessaly dominated tbe perioiko1~ the hill peoples living round the Thessalian plain; but the post was often vacant. Jason of Pherai, cagos io the 370's, estimated he could raise 6,000 cavalry, over 10,000 hoplites, aod ionumerable perioikic peltasts. The pamphleteer Isokrates estimates Thessaly's cavalry strength at over 3,000, perhaps more realistic. Jason's successors could not keep power, and in the 360's Tbebes intervened aod reorgaoised Thessaly into a league of cities. Philip II, called in to help Thessaly against Phokis, was elected tagos, and restored the old orgaoisation into four provinces, cecrarchia1~ to give his aristocractic supporters more influence; but the cities seem still to bave been important io raising troops. Thessalian cavalry fought for Philip against Phokis and in 34 1 in Thrace. Alexander, his successor as cagos, took to Asia 1,800 Thessalian cavalry in i/ai ofunequal size, that from the city of Pharsalos being the largest and best. Most ofThessaly joioed the Greek revolt of32 l but thereafter remained subject to the Macedooian kings, and provided cavalry, till 196 when Rome freed pan and gave the rest to Aitolia. Thessalian infantry, rarely mentioned, were perhaps not distinguished from Macedooiaos by the 3rd century.

12

Alexander Alexander's mĂŹlitary organisation was essentially that he bad inherited from bis father. Wc bave mucb more information on his allies and mercenaries than on Philip's. The infantry hc took to Asia numbered 12,000 Macedonians, 7,000 allied Greeks, 5,000 Greek mercenaries, 7,000 T hracians and Illyrians, and 1,000 archers and Agriancs. Cavalry were 1,800 Macedonian Companions, 1,800 Tbessalians, 600 Greeks, and 900 light cavalry prodromoi, Paionians, and Thracians. To this tota! of32,000 foot and 5,100 horse should probably be added tbc survivors of the advanced force of 10,000 foot and 1,000 borse, Macedonians and mcrcenaries, despatchcd by Pbilip. The higb proportion of cavalry is in sbarp contrast to normai G rcek practice, and Pbilip must bave been building up thc Maccdonian cavalry strength with the invasion of Persia in mind. Reinforcements did not alter the basic strucrure and balance ofthe army. Antipater was left in Macedon with 1,500 borse and 12,000 foot.

The 12,000 M acedonian foot were the 3,000 hypaspists in chiliarchiai of 1,000 and 9,000 pezerairoi in six provincia! raxeis of 1,500. (Antipater's 12,000 infantry may bave been eight sucb raxeis, but more likely not ali were Macedonians; wbereas Alexander had 2,000 archers and Agrianes in 335 in Illyria he only took 1,000 to Asia; the otber balf probably stayed with Antipater.) Their originai organisation was based on tbc file of 1Omen. By the end of Alexander's reign the file, though stili called dekas, ten, was of 16 men. As the phalanx fought 8 deep at Issos, it has been suggested the file was 8 strong at the start of Alexander's reign; but the progression from IOto 8 to 16 seems unduly complex, and the figbting formation at Issos could have been in half files. I suggest Alexander simply increased tbe depth from I O to 16. This depth became standard and tbe tbeorist Asklepiodotos gives a phalanx organisation based on a file, /ochos, of 16, which does not reflect any known organisation exactly, but seems broadly to fit both Alexander's and later armies. T wo files of 16 make a dilochia of32, two ofthese a tetrarchia of64, two of tbese a raxis of 128, two ofthese a symagma of256 (the lowest tactical unit, with its own staff and standard), rwo sy11tagmara a pemekosiarchia, meaning "unit of 500'', actuaUy 512 rank and file, rwo of these a chi/iarchia, "unit of 1,000", in fact 1,024. After this leve! the units may cease to bear any relation to real organisation, but we have the merarchia of 2,048, phalangarchia of 4,096, keras of 8, 192 and pl1ala11x of 16,384. Comparison with known units suggests tbe basic structure is sound, but notali oftbe levels in Asklepiodotos' hierarcby bear much resemblance to real units. Pemekosiarchiai were till 33 1 Alexander's basic sub-division oftbe ra.xis, obviously three to each taxis. la 331 at Susa he re-organiscd the army so the chiliarchia became the basic uni t. Obviously tbis is not compatible witb a 1,500-man raxis, but by now tbc pezerairoi had been greatly reinforced; from 9,000 they may bave numbered 11,000 at Gaugamela, and 6,000 more joined at Susa. At tbis time, or soon after, a seventh raxis was formed. I suggest Alexander now bad enougb Macedonians to organise seven 2,000-man raxeis, each divided into rwo cl1iliarchiai, and each ofthese into two pe11tekosiarchiai. Wc do not know bow the pemekosiarchia was organised but a 256-man unit like the theorists' sy11tagma is likely, perhaps caUed lochos. Hypaspist chiliarchiai were no doubt similar.

Tbe allied Greek infantry would have been mostly hoplites. No great reliance was placed on tbem - they were bostages for their states' behaviour as much as a contribution to the army, and probably few were especially enthusiastic. The Argive contingent was left to garrison Sardis, tbc rest sent bome wben the army reacbed Media, thougb some re-enlisted as mercenaries, as did some ofthc Thessalian cavalry, wbo were discharged at tbe same time. Mercenaries were used in vast numbers. G. T . Griffitb addcd up thc figure.s for reinforccments, and concluded tbat to thc origina! 5,000 we re added at least 60,000 (including cavalry) by the end ofthc reign. By tbis time Alexander was not only ofTering great prospects ofloot but was the 011/y employer of mercenaries in the eastern Mediterranean. Only a fraction oftbis vast number served in tbc field army at any one time - Griffitb suggests perhaps I 0,000 - as most were used as garrisons or military scttlers. 26,000 Greeks settlcd in tbe eastem satrapies rebelled after Alexander's death. At Gaugamela tbe veteran mercenaries who had been with Alexander from the start formed a separate unit, and they may be distinguished on other occasions by references to xenoi, foreigners, as opposed to misrhophoro1~ mercenaries. We are never told whetber Alexander's mercenarics wcre boplites or peltasts. Tbose taken ovcr from Persia, and tbc allies who re-enlisted, were presumably boplites. Griffith suggests most oftbe mercenaries were peltasts, and givcn tbeir tactical role - at Issos apparently adding flexibility to tbe pbalanx, at Gaugamela some supporting cavalry, others in tbe sccond li ne - this sccms plausible. Cenainly tbey are not used as the mainstay of the li ne.

Ofthe Balkan infantry the Illyrians and some ofthe Thracians were skirmisbing javelinmcn; at Gaugamela there were two units ofjavelins, under Balakros and Sitalkes, onc of"infantry'', probably peltasts, with tbc baggage. Thc Agrianes and archers, altbough Alexander put great reliance on tbem, wc.re only about 500 cach at the start oftbc war. Another 500 Agriancs joined in Asia Minor. Tbey seem to have been a mix ofjavelinmen and slingers. By Issos there were two bodies of archers; it seems unlikely that so smalla farce as 500 would be split, so tbere wcre probably two units of 500 or so each. At Gaugamela they are distinguished as Cretan and Macedonian. No arcber reinfarcements are mentioned, but some may have been witb Pbilip's advance farce .

13

The 1,800 Companion cavalry formed eight i/ai, the basic tactical units. The difference of300 between this and the force left with Antipater is probably accounted for by a 300-strong royal ile. The otber 7 squadrons will then have been about 215 men cach, probably with an establishment of2 l Oas this figure, like 300, allows the formation of an exact wedge, the Macedonian cavalry's typical formaùon. Anùpater's 1,500 will then be 7 similari/ai. In 331 Alexander divided each ile into rwo loc/101~ probably because reinforcements had made the i/ai roo large to be convenient tactical units. By early 329 they werc rc-organised in ltipparcl1iai, probably 8 in number. Each lzipparclzia may have been 400-500 strong nnd composed oftwo i/ai, and they may have includcd some Persian and similar troopers from the stare. After thc rcturn from India tbere were only four lzipparclziai, probably bccause of losses in the Gedrosian desert marcb, and a fifth was raised, almost entirely Iranian, under lranian officers. These Iranian cavalary in thc Companions were armed with the Macedonian cavalry spear.

The prodromoi, scouts, are also called sarissoplzoroi, lancers, and are probably Macedonians. There wcrc 4 i/ai, probably about 150 each. They are last memioned at the Jaxanes battle in 329, and may thereafter have been incorporated imo the Companions - though since the latter's lzipparcl1iai already existed, the inclusion of the prodromoi cannot be connected to thc Companions' reorganisation. By now Alcxandcr was rccruiting Iranian light cavalry who could carry out the scouting, and hc may well ha ve felt thcsc Maccdonian lancers would be pul 10 better use in the Companions' shock role.

The Paionians sccm 10 havc only provided one ile. So probably did thc Thracians al thc start, but 500 more joined in Egyp1 and 600 al Susa. These wcre lef1 in garrisons in Iran, and none were in the army which invadcd India, though 5,000 joined thc army there la1er. Cunius says 3,000 Illyrian cavalry joincd Alcxander in Bactria, bui 1his seems doubtful.

Alexander's siege train was vcry good, since he inhcritcd che best in Grecce. Unlike Philip he now hnd stoncthrowing engines a1 his di sposai. Unusually, hc 1wicc uscd ca1apulcs in the ficld - to covcr rive r crossings in Illyria and Scytbia.

Eastern troops Aftcr beating Darius Alcxandcr made extensivc use of Iranian cavalry. Thc tirsi refercnce is noi long af1er Gaugamela, when the force which outflanked a mountain pass is said by Curtius and Polyainos to include 1,000 Scythian mounted archers. If correct, tbese must be mercenaries of Persia who had changed sides after the defea1 of their paymaster. By the time he reacbed Hyrkania Alexander had raised a unit of locai mounted javelinmcn, lzippakontistai. More locai cavalry were recruited for 1he invasion oflndia - Bac1rians, Sogdians, 1,000 Saka and Dahai mounted archcrs, Arachosians and Paropamisadai. A1 the Hydaspes batti e the lise ofunits makes no mention of lzippakontistai, bu1 1hey do appear later in 1he campaign. These javelin cavalry howcvcr seem ro have been che Arachosians and Paropamisadai, the originai hippakontistai having been left in 1he gnrrison of Bactria. No units of Persians, Medes or 01her west Iranians are mentioned, though individuals may have served in che Companions already, as tbey certainly did aftcr t hc lndian campaign.

In India Alexandcr used numerous locai uoops, from willing or coerced allies. The first sucb unit was 300 cavalry from the town ofNysa. 5,000 Indians, mostly Taxilans, were in tbe army at the Hydaspes. Later Poros, now an ally, provided 5,000 of his troops, plus elepbants. Tbe Oxydrakai soon after provided 500 chariots. Alexaoder also assembled a large elephant corps ofhis own, from captured beasts and donations from friendly kings. He collected about 200, and sent 1hem 10 Babylon to become pan ofhis standing army. He had no cali 10 use them in pitched baule, but clearly had military du ties in mind for 1hem.

After the rerum from India, 1he role ofthe Iranian troops increased, thanks in part to 1he growing dissatisfaction of the Macedonians. In 324 Alexander assemb led 30,000 young Asiatics trained and armed in Macedonian style, cnlled thc epigonoi, "inheritors". When the Macedonians mutinied he threatened 10 forma complete Iraniao army, infantry and cavalry alike, to replace them - presumably using tbe epigonoi and tbose eastern cavalry already in service. W bat became ofthe epigonoi when this pian lapsed is not clear, but Polyainos mentions I0,000 Persians and three more units of500 when describing Alexander's coun ceremonial; these may be epigonoi. Soon after, Peukestes the satrap of Persia brough1 to court 20,000 Persian, T apurian and Kossaian 1roops, and it was these Alexander used in an interesting experimcnt wi1h the phalanx, mixing them in with Macedonian 1roops. In each 16-man file the first three men were Macedonian pikemen, then there were 12 Persians with bows or javelins, and che lasr man was a Macedonian file-closer with pike. C learly this was :m :mempl to combine shock and missile power, and one wonders how it would have fared in battlc. U nfortunately Alexander's death meant it was never used in acùon.

14

THE SUCCESSORS On Alexander's death in 323 his army initially stayed united. Antipater continued 10 govern Macedon and command the Macedonian army, Perdikkas became regent and commanded tbe field army in Asia, while lesser gene rais and satraps made do with locai garrisons and mercenaries. Antipater used the army of Macedon ( 13,000 foot and 600 horse) 10 subdue a Greek rising, helped by 1roops from Asia (Leonna1os wi1h 20,000 foo1 and 1,500 horse, some Macedonians; Kra1eros wi1h 6,000 ve1cran Macedonians, 4,000 rccruits, 1,000 Pcrsian archers and slingers, 1,500 horsc). Perdikkas was killed by his own officers in 32 1, Antipater briefly succeeded him but died in 319, and thercafier the unity of empire and army dissolved rapidly. The arrnies of the warring Successor gene rais (many ofwhom took the title king) tben fall into three groups. First, Macedon cbanged hands severa! times, and wboever controlled it could cali out tbc Macedonian home levy. Tbese troops were as good as ever in baule, buL in what was efTectively a dvii war were likely to change sides and transfer their allegiance 10 a rivai candidate; mercenaries at least usually only cbanged sides after the defeat of an employer. Tbere was also a sbortage of Macedonians, tbe kingdom baving been depopulated by tbe exodus of soldiers 10 Asia, many of wbom died or settled there. These Macedonian based armies would be supplemented by Greek allies and by mercenaries. Second, generals with a strong power base in Asia could supplement tbeir nucleus ofMacedonian veterans witb locai troops.

Tbese provided ligbt infantry and cavalry, so t he Asian armies often bad bigb cavalry proportions - Antigonos had 28% horse at Paraitakene - while some were trained as pikemen, 10 increase tbe phalanx strengtb. So were some mercenaries, but mercenary peltasts seem 10 have been included in "pbalanx" totals; ce rtainly Antigonos bad some in 3 17 but ali his mercenaries are included in rbe pbalanx total in rhe sources. These armies also had most of Alexander's surviving elepbants, tbough Antipater took some 10 Europe. Third group is thc minor figures wbo could not command mucb Macedonian support and bad 10 rely mostly on mercenaries; thus Arrhidaios of Hellcspontine Phrygia in 319 had over 10,000 mercenaries 10 only 1,000 Macedonians, plus 500 Pcrsian archers and slingcrs and 800 cavalry. In many cases Lhe Iesser figures were unwilling allies ofthe greater, and their persona! ambitions help explain the high frequency of Lreachery and desertion in tbese wars - Perdikkas :md Eumenes were botb betrayed by rheir own oflicers, and lowcr ranking officers quite often desen. It was also common practice for defeatcd troops 10 enlist witb tbeir conqueror, and this was probably not confined 10 those strictly called mercenaries.

Outlines ofsome oftbe better documented armies follow. Seleucus and Ptolemy bave been dealt wirh along witb their descendants later.

Polyperchon Nomina1ed by Antipatcr 10 succeed him as regent in 3 19, Polyperchon was never generally accepted. AnĂšpater's army clected instead their commander's son Kassandros. Against him Polyperchon levied 20,000 Macedonian foot, 4,000 Greek allies (be had made great efTons 10 secure the loyal1y of the Greek s1a1es, buL some sĂšll joined Kassandros), 1,000 borse and 65 elephants (su rvivors ofthe 70 brough1 to Europe by Antipater). He joined up after reverses with Alexander's mother Olympias, who had raised an army in Epeiros, including Thracians and cavalry from Ambrakia. But she, with some oftbe elephants, was penned up in Pydna, starved out (the elephan1s were fed on sawdust, and died; tbc mercenaries resoned 10 ca nnibalism) and killed in 316. Polyperchon had a last fling in 31 O, allying witb Aitolia and otber anti-Macedonian Greek states to back a pretender to the Macedonian tbrone; they raised over 20,000 foot and 1,000 horsc. But he was persuaded over to Kassandros' side, and is last beard of as a subordinate in his service, commanding 4,000 Macedonian foot and 500 T hessalian horse.

Kassandros Kassandros began bis career in 318 with only 4,000 men, probably mcrcenaries, lent bim by AnĂšgonos as a thorn in Polyperchon's side. He took aver Athens, where hc set up a puppet regime, then was accepted as king ofMacedon and subdued much ofGreece; when he re lieved his ally Megalopolis from Polyperchon's unsuccessful siege he took over some ofthe latrer's elephants. In 302 he could raise 29,000 foot and 2,000 borse; as much ofGreece bad gone over 10 Dcmetrios 1hesc must ha ve been mos1ly Macedooians and mercenaries. Kassandros died in 298 and his sons could noi hold tbe kingdom.

Alketas Alketas was one of Perdikkas' loyal subordinates who af1er his commander's deaih carried on the war agains1 Antigonos. In 320 be bad 16,000 foot and 900 borse. 6,000 oftbese were the warlike and virtually independent Pisidians, and only these and his persona! guards s1ayed loyal after his dcfeat. The Pisidians were mostly light javelin-armed infantry. Both Alexander and Antiocbos lii, campaigning in Pisidia, met small forces of Pisidfan cavalry, so a few of Alketas' horse may have been Pisidians too, again probably javelin skirmishers. The rest ofthe force was no doubt mostly mercenaries.

15

Peukestes The satrap of Persis, Peukestes was popular with his subjects because of his fondness for Persian costume and language. He bad many Persians among his troops - I0,000 Persian archers and slingers, 3,000 men of mixed origin, pantodapoi, armed in Macedonian style (no doubt mostly Persians with Macedonian officers, and perhaps fĂše-leaders), 400 Persian cavalry and 600 Greek and Thraciaa borse, settlers left by Alexander from whom Peukestes no doubt raised his small agema (less than 300 men). He fougbt for Eumenes, deserted him at Gabiene, and was deposed by Antigonos.

Eumenes Eumenes ofKardia, a Greek from the Thracian coast wbo bad been Alexander's milicary secretary, became satrap ofKappadokia after tbe regent Perdikkas conquered it in 322. Wben Perdikkas and Antipater fell out, Eumenes raised an army to figbt Antipater's generals Krateros and Neoptolemos. Althougb be bad little military experience previously, Eumenes sbowed considerable military talent, and inspired great loyalty among many of his men, tbougb tbe Macedonians resented a Greek interfering in their private squabbles. He levied 20,000 foot, ofmixed ori gin and quality but including a nucleus of veteran but unruly Macedonians, and 5,000-6,300 good cavalry, many of tbem Kappadokians but including Greeks and Tbracians. H e beat Krateros but soon after was beaten by Antigoaos wben his cavalry commander deserted, and besieged in a Kappadokian fonress wich 500 borse and 200 fooc, till he surrendered on cerms.

Tbereupon he raised anocher army, 2,000 foot and perhaps l,000 borse from bis ex-soldiers and loyal Kappadokians, and to carry on tbe war witb Aotigoaos was given command by tbe regent Polyperchon oftbe 3,000 argyraspides, "silver sbields", under tbeir officers Antigenes and Teutames. This corps was Alexander's old hypaspists, but as they were stili at che same strengtb as under Alexander, while ali t!beir men were veterans ofhis wars - not one of them younger tban 60 (in 316) - tbey must bave been brougbc up to screngch wicb veterans of ocher units. One problem wicb tbis identification is that argyraspides and bypaspists appear as separate corps ac Paraitakene. But since argyraspides is occasionally used for tbe bypaspists even in Alexander's day the identification seems secure, wbile "bypaspiscs" is widely used for guards units of different kinds in different armies, so probably has anotber meaning ac Paraicakene. Tbey may be Eumeaes' 2,000 loyal veteran foot wbo joined him after tbe siege, or else tbe small forces of Macedonian fooc who served among the guards of che various eascern sacraps, assembled in one unit.

Eumenes recruited abouc 5,000 mercenaries in che coastal areas ofSyria, Anatolia and Cyprus, and marched east.

He tben persuaded mosc ofche eascern sacraps co join bim, including Peukesces of Pers.is and Eudamos oflndia, who brougbt 120 elephants be bad obtained by treacberously killing che lndian king Poros. Eumenes now bad 35,000 foot and 6,200 borse. The foot were 3,000 argyraspides and 3,000 bypaspists (botb numbered among tbe phalanx; bowever Alexander's bypaspists bad been armed, the need now was for as many Macedonian pikemen as possible), over 6,000 merceaaries, some witb Macedonian equipment, and 5,000 of mixed origin, pantodapoi, also armed in Macedonian style (3,000 of tbem were Peukestes' men). Tbe otber 18,000 were skirmisbers, 10,000 of tbem Peukestes' Persians, che rest no doubt a mixrure of mercenaries and native troops from tbe eastern satrapies.

Cavalry included Eumenes' bodyguard or agema of 300, probably bis loyal Kappadokians, anotber I 00 of his retainers (paides, "boys"; it bas been translaced slaves, but the word is used for tbe Macedonian royal pages, so chese may bave been similar young noble trainees and bostages), 900 Companions (chose of Alexander's Companions wbo bad not attacbed tbemselves to tbe bodyguard ofone or ocher generai seem to bave split between Antigonos and Eumenes), 1,000 Aracbosian and Paropamisadai ligbt borse, 500 Tbracian ligbt horse recruited from settlers in the east, and 3,500 who are difficult co classify as they are known mainly by tbeir leaders' names, but are probably mostly Macedonians or colonists settled in Asia, thougb some may bave been natives. U nit sizes are wildly irregular as generally eacb satrap kept bis own troops witb him, but at Paraitakene we do bear offlank and advance guards in small i/ai of 50, wbile Peukesces' and Antigenes' guards combined tO form one ile of 300 (a fine example of the imprecision ofunit terminology). Witb this army Eumenes fougbt Paraitakene and Gabiene, but was !et down by the unreliability of his reluctant subordinates and tbe treacbery of che Macedonian argyraspides. Wbea tbese deserted to Antigonos, be posted them to distant Arachosia, witb secret orders to the satrap never to !et tbem see bome again. He was not foolisb eaougb to use them in battle!

Antigonos and Demetrios Antigonos Monophrhalmos ("One-eye"), originally satrap of greater Phrygia, gradually rose in power and incorporated tbe troops ofhis defeated enemies iato his army. H e carne closest ofanyone to restoring Alexander's empire. Whea be first clasbed witb Eumenes in 320 be bad only 10,000 foot, halfofchem veteran Macedonians, 2,000 borse and 30 elepbaots; baving incorporated many of Eumenes' troops he fielded 40,000 fooc and 7,000 borse

16

against Alketas; took over most ofhis troops, and had 60,000 infantry and I 0,000 cavalry (and stili 30 elepbants) to figbt Arrhidaios. When he met Eumenes again in 317 Antigonos had joincd up with Pitbon, satrap of Media, and fielded over 28,000 foot, about 11,000 horse, and 65 elephants. Thc infantry were 8,000 Macedonians, 9,000 mercenaries (some of thcm peltasts), 8,000 pamodapoi with Maccdonian equipment, 3,000 Lykians and Pampbylians; ali these were reckoned as tbe pbalanx, and there was an unknown number oflight troops (Persians and Cretans are mentioned in the army on other occasions). Cavalry were Antigonos' own agema of300, perhaps 300 paides, 1,000 Companions, 1,000 Greek allies, 2,200 assorted settlers in the cast; ali these will bave been heavy cavalry. Light horsc were 1,000 Mcdes and Parthyaians, 1,000 Thracians, 2,300 ofthe new Tarantines. Others may have been heavy or light; 500 assorted mercenaries and 1,500 with Pithon.

The army Antigonos' Oamboyant, impetuous son Demetrios Po/iorketes, "the besieger" took to Gaza in 312 was of broadly similar composition, with a phalanx of 11,000, 1,500 skirmishers, 4,600 borse and 43 elephants. The phalanx was of 2,000 Macedonians, 1,000 Lykians and Pamphylians, and 8,000 mercenaries. Horse were Demetrios' guard of200, 800 Companions, 300 more heavy cavalry in 6 i/ai ofSO, 100 Tarantines in three i/ai of about 32, and 3,000 unspecified but mixed types.

Fighting Kassandros in Thcssaly in 302 Demetrios had numerous locai allies; he fielded 1,500 borse (including some Tarantines), 8,000 Macedonian foot, 15,000 merccnarics, 25,000 Grcek allies, and 8,000 light troops and piratcs. Fewcr cavalry than usual but not many could be transported to Grcece, or raised locally. The other kings were already combining against Antigonos, and he recallcd Dcmetrios' army which formed part ofthe 70,000 foot, 10,000 borse and 75 elephants wbicb fought at Ipsos - a mcasure ofthe growth of Amigonos' power. There were many Athenians in that army, probably mercenarics or volunteer allies who had come with Demetrios. Aftcr Antigonos' defeat, Demetrios was reduced to a freebooter, his powcr based on ships aod mercenarics. But in 294, intervening 10 "help" a son of Kassandros against Pyrrhos of Epciros, he took the throne of Maccdon, and wĂŹth Maccdonian rcsources overran Tbessaly and much ofGreece. However he was unpopular in Macedon, thc army mutinied, and Demetrios fell. He invaded Asia with 11,000 mercenaries, but after a desperate series ofcampaigns he was captured by Seleucus in 285. His son Antigonos Gonatas ("Knock-knees"?) succeeded to his few remaining possessions, but Gonatas' armies were to be different.

Lys imachos Lysimachos was Alexander's governor of Thrace. In the first stages of the Successors' wars he kept out of the generals' intrigues, steadily exteoding his power over Thrace and its coastal Greek cities. He beat Scuthes ofthe Odrysai into submissioo but failed to conquer the Getai. Lysimachos was the most bruta! and savage of the Succcssors, massacring mercenari es Iest they dcscrt, taking citics and kingdorns by treachery, regarded by his rivals "as a villain and a barbarian", yet clearly an ab le and calculating politician and generai. In 322 he only had 4,000 foot and 2,000 horse, but grcw in power till he could contribute about 44,000 foot and 3,000 borse 10 the alliance with Seleucus in 30 I. As a result Lysirnacbos took over western Asia Minor. In 288 he took half Macedon when Demetrios wasdeposed, partitioning it with Pyrrhos, from whom he soon took the other half, and Thcssaly. He also siezed Paionia, exiling the rigbtful king. He was defeated and killed by Seleucus in 281, deserted by most ofhis Asian subjects. Ptolemy Keraunos, "the thunderbolt", took over the European part ofthe kingdom :md the arrny after murdering Seleucus, but be was soon killed and the army annihilated by the Galatians.

Lysimachos' early forces were presumably mostly Macedonians and mercenaries, and later he had access 10 the whole levy of Macedon and Thessaly. He would also have been able 10 raise troops in tbe Greek towns on the coasts ofThrace and Asia, but may not bave done so as he scems to have been very unpopular there. There are occasionai hints that, not surprisingly, be used Thracian troops. In 313 he faced a coalition oflocal Greek cities, Thracians and Scythians, but persuaded the T hracians 10 change sides and join him instead. The stele of a Bithynian officer rccords him killing a Thracian and a Mysian ne:ir Korupedion, probably Lysimachos' men in the battle of281.

Finally, the late writer Lucian mentions a Thracian infontryman witb sarissa and shield killing an unknown governor of Media, perhaps in a 3rd century context. Lysimachos is more likely than any other generai to have Thracians in his phalanx, and the governor of Media may be a subordinate ofSeleucus. Illyrian mercenaries were also used, or abused; Polyainos records Lysimachos massacring 600 Autariatai whose baggage had been plundered, lcst they dcsert, while 2,000 Autariatai, perhaps fearing a similar fate, deserted to Antigonos in 302. 800 Lykians and Pamphylians were with them, thc first sign Lysimachos was using Asiatic troops. As he had only just set foot in Asia and had no territory there to recruit in, thcse dcserters may ha ve been a garrison of Antigonos, captured, forced to enlist with their captar, and taking the first opportunity 10 remedy the situation. Thc Mysian mcntioned abovc suggests that later Lysimachos levied men in his new Asian domains. By this time he also probably had a few elephants, from tbe farce which had been Kassandros'. These passed to Keraunos, who was riding an elephant in the banle with the Galatians, and died when it threw him; his liead decorated a Gallic spear.

17

ANTIGONID MACEDON When Demetrios was taken prisoner his son Amigonos Gonatas inherited litrle more than a few mercenary garrisons in Greece. But in 280 the Galatian invasion reduced Macedon to chaos, and wheo in 277 Gonatas crushed one branch ofthe invaders at Lysimacheia he was able to gaio the Macedonian throne. At first he was reluctant to cali out the Macedonian national army: Macedon was depopulated and devastated by the invasion, and as yet the pcoplc had little loyalty to the new king. So he supplemented his Greek mercenaries by hiring large numbers of Gauls, once a whole tribe together, who were readily available, fierce and cheap. By 273 he was however using Macedonians - to his cost, as they deserted to Pyrrhos while the Galatians fought and died to the last man. Pyrrhos briefly retook Macedon, but Gonatas' mercenaries played a part in the fighting which soon led to Pyrrhos' death.

Gonatas recovered Macedon and the Antigonid dynasry ruled till Rome deposed Perseus in 168. By the end ofhis long reign (276-239) Gonatas' army would be much like che better documented armies ofhis successors. One difference was that Gonatas had a few elephants, probably less than 20, which had come down from Alexander (via Antipater, Kassandros, Lysimachos and Ptolemy Keraunos). None ofhis descendants had access to elephants.

In Antigonid armies cavalry played a minor role, amounting only to 5-10% ofthe tota) streogth. Aotigooos Doson only took 300 Macedonian horse 10 Sellasia, Philip V had 2,000 Macedooian and Thessalian at Kyooskephalai.

Perseus, without Thessaly, raised 3,000 Macedonian cavalry, the resuh of 30 years of generai peace since Kynoskephalai and a steady increase in the population. Perhaps too maoy of the nobility wem lO Asia with Alexander and never returned, and it was not till Perseus' day that the cavalry recovered in numbers - though their quality was always good. Thcy included a royal, or "sacred" ile. Philip V had 400 "household" cavalry in 219 who may be this force, and the 300 at Sellasia are probably the same unit at a lower strength. The native cavalry were usually supplemented by allies and mercenaries. 600 Greek allies and 300 mercenaries were at Sellasia, Philip V used Gala tian horse, while Perseus had 1,000 crack Thracians at Pydna and 200 Illyriao cavalry io garrison at Kassandreia.

A corps of native Macedonian "peltasts", not to be confused with the mercenary peltast, was the equivalent in status and perhaps equipment and role of Alexander's hypaspists. (Hypaspists appear in the Antigonid army only as the king's inner bodyguards and aides.) Within this corps was an elite body called the agema or, according to Livy, "The Conquerors", "chosen for their strength, vigour and youth" - though ali the peltasts were regarded as crack troops. Doson had 3,000 peltasts at Sellasia. Only 2,000 are mentioned at Kynoskephalai, but Philip V had 5,000 earlier, in 219. Since Perseus too had 5,000, of which 2,000 formed the agema, I suggest Philip's orgaoisatioo was the same, and the 3,000 "ordinary" peltasts have got themselves left out of the Kynoskephalai figures, or else included in the phalanx tota!. Plucnrch speaks of3,000 elite troops who fough to the death at Pydna, undoubtedly some ofthe peltasts and perhaps, ifthe figure is slightly out, the agema. T he peltasts were probably organised into chiliarchiai subdividcd like those ofthe phalanx.

The phalanx was proportionately more important than under Alexander or the Successors; it provided one-third to two-thirds of the strength of the army (34% at Sellasia, lower than usual because of the number of Greek allies present, 62% at Kynoskephalai, 49% at Pydna). Probably every free Macedooian was obliged to serve, property qualifĂŹcations no longer being important because thc troops were now armed by the king. Wc do hear ofbarbarians settled in Macedon being given land to hold in return for military service, but this may not have extended to native Macedonians. T he phalanx was divided into two probably equa! corps, the chalkaspides, "bronze shields", and leukaspides "white shields". Cl1alkaspides are sometimes found on distane expeditions without leukaspides, so may have bcen readier for prolongcd servicc, perhaps because thcy wcrc rccruitcd from thc youngcr men. The basic tactical unit of the phalanx was the speira apparently of 256 rank and file, divided iato 4 tetrarchia i of 64, each composed of 4 fil es of 16. The systcm is clcarly similar to thc thcorctical layout of Asklcpiodotos, but with alternate steps - the dilod1ia and taxis of the theorist - missed out. Four speirai formed a unit of about 1,000 called chiliarcl1ia or perhaps meros, under a hegemon. These were grouped into strattgiai; continuing the system would suggest strategia i of 4,000 men, but as Peter Connolly has pointea out this docs not lit the army at Sellasia where chalkaspides and leukaspides would be 5,000 men each. Hc thereforc suggests straregiai of5,000 (which would have the advancage ofmaking the peltasts equivalent to a strategia) but this in tum docs not lit Kynoskephalai, where the cwo divisions ofthe phalanx were 8,000 each! lt seems likely that at the very least the strategia varied in size with time, ifit was not in fact a totally ad hoc command, dependent on the numbers available; cenainly its commander, che strategos, bears a title of very flexible meaning.

Allied and mercenary troops were used to increase this force, balance thcir capabilities, and sometimes to take the risks while sparing the valuable Macedonian citizens. Greek allies and merccnaries werc used in the greatest numbers at Sellasia, where there were 3,000 infantry and 300 cavalry mercenaries, 600 assorted allied cavalry, 3,000 picked Achaian foot, 1,000 M egalopolitans, 2,000 Boiotians, 1,000 Epeirors and 1,000 Akarnanians. T he

18

Megalopolitans had becn armed as Macedonian phalangites by Doson; Boiotian and Epeirot armies already fought that way. The Akarnanians would be light infantry ofsome sort, Lhe Achaians tl111reophoroi. Philip Valso fough1 in alliance wi1h Achaia and had 300 Achaian slingers wi1h him in 219. The mercenaries a1 Sellasia included Cre1an archers, who werc found in jus1 abou1 every Antigonid army; Perseus had 3,000 a1 Pydna, and only 1,000 otber Greek mcrccnaries. The normai use of1he mercenaries on campaign sugges1s 1he resi of1hem were lightly armed too, probably th11reopl10roi. They were also much used for garrison duty, to savc calling up Macedonians in peacetime.

Galatians were never hircd again on the scale Gonatas had used them, bui Doson had 1,000 Gala1ian infantry at Sellasia, Perseus 2,000 in 169. Hc tried to hire 10,000 cavalry and the samc numbcr ofsupponing infantry from tbc Celtic Bastarnai, bui he baulked a1 1he cost and fought with no Celtic cavalry a1 ali.

Agrianes and Paionians also provided troops; there were 1,000 Agrianes at Sellasia, PhĂŹlip had 500 in 219, and Perseus had a mixed force of3,000 Agrianes, Paionians and Thracians settled in Macedon. Thc javelinmen and slingers grouped in units of400 at the Kallinikos skirmish in 171 may include these. lllyrians were also used; 1,600 at Sellasia were allies ofMacedon rather 1han mercenaries, under Demetrios of Pharos; Philip sent 400 to Crete in 219 and fielded 2,000 Tralleis at Kynoskcphalai. For some reason modem writers consistently cali these men Thracians, but Livy quite specifically says they were an lllyrian people. (They are probably quite dis1inc1 from a similarly named people found in Asian armies, Trallianoi rather than Tralleis; both get anglicised as "Trallians".)

Perseus had 1,000 infan1ry and 200 cavalry of the Penestai tribe in garrison at Kassandreia in 169.

Thracians, usually infantry, were often used in some numbers; Philip V had 2,000 at Kynoskephalai, while Perseus in 169 had 3,000 "free" Thracian mercenaries, and an allied contingent commanded by King Kotys ofthe Odrysai, 1,000 elite cavalry and 1,000infan1ry10 support them.

Finally, Philip V's garrison ofSikyon in 198 contained a numbcr ofl1alians; some were ve1erans ofHannibal's army who did not dare return to Roman rule, while others were desertcrs from tbc Roman flect.

EPEIROS Early Epcirot uoops appearas 1ribal irregulars similar 10 tbe lllyrians, bue by 280 when Pyrrhos went to ltaly he had a modem army organised in Macedonian style. He may have created 1he army himself, but another candidate is his predccessor Alcxander of Epciros; hc was a friend and ally of Philip II of Macedon, and his abili1y 10 mount an expedition 10 ltaly in 334 sugges1s be had a better organised force tban tbe old tribal levies. However wc have little de1ail on his forces.

Pyrrhos took 10 ltaly 23,000 infantry (including che advance force of 3,000), 2,000 archers, 500 slingers, 3,000 cavalry and 20 elepbants (wbich he is variously said 10 bave captured from Demetrios or bccn given, along with other troops, by Ptolcmy Keraunos). The comingems which made up this force can be identified in Dionysios of Halikarnassos' lis1 of units at Asculum. Best ofthe infantry were 5,000 Macedonians lene by Keraunos. (Tbere is no trace of an Epeiro1 infantry guard equivalent to che hypaspists of Alexander, tbougb a body of picked Chaonian infantry is later mentioned in Greece. When "hypaspis1s" are men1ioned in Pyrrbos' servicc be is sending 1bem 10 govern towns in Sicily, indicating that the term bere refers to a cadre of aides and staff officers ratber 1han a fighting unit.) In addition he had infantry from Ambrakia, the Greek city be bad made his capital, and from the tbree main Epeirot tribes, the Molossoi, Chaonai and Thesprotoi, plus a mercenary contingent of Ai1olians, Atbamanians and Akarnanians. These mercenaries will bave fougb1 as lightly armed javelinmen in Lheir national style, while the orhers fougbt as a Macedonian style phalanx. Ambrakia is unlikely to have provided more than 2,000 men, leaving the three Epeirot units and tbc mercenary force averaging 4,000 eacb. The bcst oftbe cavalry was an agema, 2,000 strong according 10 Dionysios, tbougb this seems unduly large. 2,000 may perhaps have been the tota! ofseveral native Epeiro1 units including tbc agema - who may perhaps be identified witb tbc Molossian cavalry at whose head Pyrrhos fought later in Greece. Keraunos also lent some Thessalian cavalry; 1be sources give larger figures for these 1han for Pyrrbos' entire cavalry strength, but there may in fact have been about 500. This would leave 500 for the Ambrakio1 borse and Macedonian, Ai1olian, Akarnanian and Athamanian mercenary cavalry.

The numbcrs oftroops provided by Pyrrbos' ltalian allies are difficult to determine. He fought his first battle, a1 Heraclea, with only bis own troops and tbc Tarantines; bui at Asculum be was belped by Taramines, ltalio1 Greek mercenaries in Tarantine pay, and the Oscan Samnites, Lucani and Brunii (aU tbese providing both infanrry and cavalry). The Tarantines had claimed tbe alliance (including the Messapii, wbo were in the alliance, but whose troops are not mentĂŹoned at Asculum, uoless they provided the unidentified javelinmen wbo appear in the skirmish 19

li ne) could raise a vast army of 20,000 horse and 350,000 foot. Nothing like this was of course ever done, and the highest figures for the army fielded at Asculum are 70,000 foot and perhaps I0,000 horse; a lower figure of 40,000 may be closer to the truth. Some ancient writers add the Campani and Apuli to Pyrrhos' allies, but it is clear that both in fact stood by Rome, unless a few Apulian towns were stili subject to Taras.

Soon afier this he campaigned in Sicily, and again initially received much locai support; he landed there with about 10,000 troops, was brought up by locai help to 30,000 foot and 2,500 horse, and when the enccrprise collapsed retumed to Italy with twicc as many troops as he had taken with him, 20,000 foot and 3,000 horse. The additional troops must mostly have been Siceliote Greek mercenaries.

In Greece Pyrrhos relied chiefly on Epcirot troops, as without thc resources ofTaras he could probably afford few mercenaries. Twice he was briefly king of Maccdon and Thessaly and could use their resources. He was at times allied with Aitolia, controlled Akarnania, and some sources say that in his last campaign against Sparta he was allied with Elis and Messene, though this is not certain. In 274-3 he hired at least 2,000 Gauls, who fought well bue lost Macedonian support by looting indiscriminately, plundering the Macedonian royal tombs. They formed pan of an army of25,000 foot, 2,000 horse and 24 elephants.

After Pyrrhos' death Epeiros revened 10 minor power status, fust under bis descendants and chen as a republic, no doubt mostly using ciùzen troops. However we do have an account of one singularly unfonunate experience with mercenaries. A force of800 Gauls, allegedly exiled by their own countrymen, had served Carthage in Sicily; they pillaged one city they were garrisoning, tried to betray another, and desened to Rome. In Roman service they sacked a tempie, and so were ejected from Italy as soon as possible. The Epeirots hired this band of desperadoes and set them to garrison the capitai Phoinike. True 10 form, the Gauls berrayed the city to a fleet ofpassing Illyrian pirates, and the entire population of Phoinike was enslaved! As Polybios says "no people, if wise, should ever admit a garri son stronger then their own forces, especially if composed ofbarbarians".

PTOLEMAIC EGYPT Aftcr Alcxander the Great's death Egypt fell 10 Ptolemy, son of Lagos (hence his dynasty is occasionally called "Lagid" rather than the more usual "Ptolcmaic") and his descendancs ruled the country till the death ofCleopatra in 30 BC. For much ofthe time the Ptolemaic kingdom also included Cyrenaica, Palestine, Cyprus, the south and west coascs of Asia Minor, and a foothold on the coast ofThrace. Ptolemy I's army must have been based on the small garrison of Macedonians and mercenaries Alexander lefl in the country. To increase his strength he hired more mercenaries - he could send 13,000 mercenaries to Cyprus in 306. Most ofthe 18,000 foot and 4,000 horse at Gaza in 312 were Macedonians or mercenaries, but there were also some naùve Egyptian troops, as well as Egyptian porters and camp servants. Ptolemy cleverly kept the loyalty ofhis mercenaries by retaining their baggage and belongings in Egypt; thus when Demetrios captured 3,000 of Ptolemy's men in Cyprus in 306, and incorporated them imo bis units in the accepted manncr, the mercenaries promptly deserted back 10 Ptolemy to regain their belongings!

Reliance on mercenaries was replaced by a regular army which can be reconstructed from surviving 3rd-2nd ccntury papyrus documencs. The soldiers were maiatained by grants ofland, kleroi, varying in size according to rank and arm of service. These klerouchoi or katoikoi would li ve on their land in peacetime and only be called out for war. Ptolemy I was probably responsible for starting this systcm himsclf, as he is recorded settling prisoners ofwar on the land in Egypt; prisoners were certainly recruited into the later klerouchoi. The sons ofthese military settlers received training in the corps of epigo11oi, "heirs". lt seems likely that the klerouchoi were equipped from state arsenals like other Hellenistic troops, but the cavalrymen's horses at !cast were their own, as they could be disposed of in a will. Klerouchoi are often referred to as Macedonians, and no doubt the aim was to create an army as Macedonian as possible; but the ethnic origins of soldiers recorded in the papyri suggest no more than a third were real Macedonians, the rest coming from ali over the Greek world, with even a few non·Greeks - Jews, Persians, a solitary Roman. The use of" Macedonian" to denote a class oftroops rather than a genuine nationality is typical of the Hellenistic world - "Tarantine" is another good example - and especially ofthe Ptolemies.

The royal guard units were recruited from klerouchoi. There was a household cavalry force, 700 strong at Raphia, and an infantry agema 3,000 strong at the same date, which may be indicated by the solitary reference to hypaspists in Egypt. Post cd next to the agema at Rapbia were 2,000 "peltasts", and the para Ilei with the Antigonids' agema and peltasts suggest this force too may bave been pan ofthe guard, but they could have been mercenaries. With the introduction ofnative Egyptians iato the army on a larger scale for Raphia and after picked Egyptians, maclzimoi epilektoi, wcre incorporated into the guard. Some mercenaries were also stationed at the capitai, Alexandria, and 20

tbere may also have been mercenaries wich guard status.

Aparc from the guards, the settler cavalry were organised into ltipparcltiai each under a ltipparchos; tbese were divided into i/ai, these into locltoi, and the locltoi apparently into dekades, tens. Sizc ofthe units above che dekas is unknown. The hipparcl1iai foll into two categories; at least live were known by numbers in the 3rd century, and the 4th and 5th ltipparcltiai are known to have still been in existence at the end of the 2nd. There were also four ltipparcltiai oflower status known by ethnic names - che Thessalians, Thracians, Mysians and Persians. Troopers ofthese units had landholdings ofonly 70 acres, wbereas cheir colleagues in the numbered units had 100. T his bas led Lesquier to suggest that thc "numbered" troopers were heavy, and the "ethnic" troopers light, cavalry. In che 2nd century the 70-acre lots were increased to 80, and the ethnic units acquire numbers to replace their names, so we hear ofthe l st and 2nd "ltipparchiai ofche 80-acre men". The cthnic names may referto the style ofthe units' equipment; they may possibly referto the majority nationality ofthe originai units, but if so they soon ceased to be an accurate description of the troopers' nationalities. One "Thracian" horseman is in fact a Pergamene, a " Persian" (with a Greek name) becomes a "Mysian" when he t ransfers to che appropriate unit, and so on.

Regardless ofwhether some men were genuinely ofthe same ancestry as the unit names they bore, in generai we bave anocher example of national names representing units or classes of soldier rather than genuine nationalities.

Tbe papyri indicate at least two ofthe senler ltipparchiai served in tbc Raphia campaign, but, confusingly, they are not mentioned there under any readily recognisable name. Polybios lists 2,000 "cavalry from Greece and mercenaries" who fought on one wing, while the guards plus 2,300 "cavalry from Libya and from the locai country" were on the other. Note that Polybiosdoes 1101 say "Libyan and native Egyptian cavalry", as he isso often translated; this is in interesting and perhaps deliberate contrast with his specilic use of"Libyan" and "Egyptian" forche native infantry. Native Egyptian cavalry are known from che papyri, but 11ot as early as Raphia; tbey do noi appear till the mid-2nd century. E Gal'ili convincingly points out that to raise a native cavalry force would be a very different proposition from training Egyptians as pikemen, which requires relatively linle training: "the virtual absence then ofhorses in Egypt 's agricultural economy - and bence, 100, che absence of a narural class of riders and horsemasters there - would obviously make it impossible fora serviceable cavalry force to be improvised out of Egypt's 'fellaheen' ". Greek accounts of the last independent Egyptian armies seem to suggest that they consisted entirely of infantry, so the macltimoi had apparently lost che tradition of cavalry lighting thai some probably possessed durìng the Saitic era.

So wbile it obviously was possible eventually to raise native cavalry, it would ha ve been difficult, if possible at ali, to raise any in the limited time available for the re-orgarusation ofthe army before Raphia. And ii is questionable whether Ptolemy IV's ministers would think it worch tbe effon. Tbe success oftbe new Egyptian infantry at Raphia must not obscure the fact thai when the army was being raised cheir performance was very much an unknown factor, and che wbole idea of enlisting natives something of a gamble. lt tbus seems unlikely that horscs would be wasted on uoops of sucb uncenain quality with no tradition of cavalry fighting. With this in mind the "locai" cavalry at Raphia seem very likely to be none other than the hipparcl1iai oftbe military settlers. The horse "from Libya" may be klerouchoi settled there, or alternatively citizens ofKyrene aod tbe otber Greek towns ofLibya, rather than native Libyans, who aga in had no tradition ofcavalry warfare. The distinction between "mercenary" cavalry and "cavalry from Greece" is probably a dis1inc1ion between 1roops already in Ptolemaic employ and recruits newly hired in Greece. (Alternatively, it is tempting to wonder if Polybios was misled by tbe names ofthe "ethnic" hipparchies and interpreted them as real "barbarian" mercenaries. Ifthese are his "mercenary'' cavalry, and the horse "from Greece" are all the Greek mercenaries, then the " locai" cavalry will have been only the numbercd hipparchiai.)

The bulk ofthe k/ero11cl1oi were infantry who formcd a Macedonian·style phalanx, 25,000 strong at Raphia. Some soldiers were called protoi, "first"; they may be "privates first class", paid more for long or distinguishcd service, or may be che file·leaders, who would form the first rank in batti e. Probable orgaoisation is that four files of 16 made up a pemekomarchia (equivalent IO thc tetrarchia of the theorists and che Antigonids), IWO of these fonned a ltekatomarcliia ("uait of I 00", though presumably of 128 rank and file), and rwo of these a symagma, which is known to bave bad its own standard·bearer and officers like the symagma ofthe t.beorists and che Antigonid speira, and was like them che basic tactical unit. Four syntagmata made a chiliarchia under a chiliarchos; these units were numbered as a reference 10 "tbe 7th clti/iarcliia" exiscs (hekatomarchiai, symagmata and probably other units were identified by their commander's name). As with the Ancigonids they may have been grouped into strategiai.

The native Egyptian hereditary warrior class, the 111acl1imoi, stili existed io che Pcolemak era and stili gave service in exchange for small land holdings. Some fought at Gaza, probably as lighc uoops, and chey served with the fleet and for internal policing, but ciii Raphia were not greatly relied on for land banles. But as pan of the reforms whereby Pcolemy IV's miniscers re-organised the outnumbered and inefficient Pcolemaic army before Raphia, native Egyptians wcre recruiced on a larger scale. The hereditary qualification ofthe macl1i111oi was abandoned and 21

Egyptians enlisted regardless of class - though ali seem to have received land and been incorporated into thc

machimoi class. 20,000 armed as Macedonian phalangites fought at Raphia, alongside the 25,000 GraecoMacedonian klero11cltoi. (lt has been argued that Polybios' figures ar e wrong and that 25,000 was the tota/ strengtb of the phalanx; the error is unlikely and unnecessary 10 assume, and Polybios' figures ha ve been convincingly upheld by the two latest writers on the battle, Bar-Kochva :md Gal ili.) Botb machimoi infantry and cavalry appear in units called laarchiai, known by the namcs oft hcir commanding officers; it would seem likely that the size and organisation of1heir uni1s in fact followed those of1hc regular army. While arming the machimoi won Raph ia, it proved a mixed blessing since, carried away by 1heir viccory, they proceeded co spearhead a seri es of native revolcs.

Ac che start of thc lst century BC a new and rather surprising class of troops appears in thc papyri , "Greek 111acl1i11101~'. Whethcr they represent a new stage in the evolu tion of tbe army, or simply indicate that 1he Greek lower class and the Egypti:ms were tending 10 merge, is unknown.

Along with the Egyptians, 3,000 Libyan infantry in Macedonian equipment fought at Raphia. They may bave been regarded as merccnaries rather than subjects, as Libyan mercenaries appear in 2nd-century records. There is no sign that Libyans were used in tbeir traditional national role as javelin ski rmishers. Nor is there any sign that the Ptolemies attempted 10 use 1he Nubian archers who had been an important pan of Pharaonic armies.

Greek mercenaries continu ed 10 be imporcant to the later Ptolemies, though not uscd to the same extent as under Ptolemy I. They were organised in the samc kind ofunits as the regulars. Tbc Greek mercenaries at Raphia were 2,000 horse, 3,000 Cretans, 8,000 infantry who fought and drillcd in one body with the senler phalanx, and must thus havc been armed in the same way, and perhaps the 2,000 pelcascs. As well as serving in tbe field army, mercenaries formed most oftbe garrisons ofEgypt 's overseas possessions such as Caria and Cyprus. They were of widely diverse origins. Ptolemy I's mercenaries included Rhodian, possibly the famous Rhodian slingers though we are not told so. In 221 tbere were l,000 Cretans and 3,000 Pcloponnesian mercenarics at Alcxandria. In 203 6,000 Aitolian foot and 500 horse were hired, and fought at Panion in 200. The mercenary garrison of Cyprus formed clubs or associations, koina, which reveal the predominant narionalicies. The koina of Acbaians, Ionians, Cilicians, Lykians, Cretans and Tbracians are known. The Lykians and Cilicians may be thureophoroi like tbe other Anatolians known from grave monuments at Alcxandria, and so may be the Achaians and Ionians. The Cretans and Thracians are more likely 10 bave fought in the their native sryles. Wc also bear of"mercenary kleroucl101~• settled in Egypt, presumably men wbo entered Egyptian service as mercenaries but then were granted land, and perhaps thercby incorporated in tbc regular army.

Galatians and Tbracians are known both as sc ttlers and as mercenaries. At Raphia 4,000 such senlers fougbt along¡ side 2,000 newly recruited mercenari es. It seems likely that these settlers were "mcrce nary kleroucho1". Tbracians could be recruited via the Ptolemaic foothold on the Thracian coast, andare known from thc garrison ofEpbesos in the 3rd century as well as Cyprus. Galatians wcre first hired as early as 274 when Ptolcmy II hired 4,000 10 campaign against his brother Magas who was in revolc in Kyrene. They in rurn proved rebellious, and were marooned on an uncuJtivated river-island 10 starve! .\siatic merccnaries were also uscd, but probably (with the exception of the Anatolian tlr11rcoplroro1) not in grca1 numbcrs. An associa tion of Idumacan Arab merccnarics is known from Memphis, but it is possible that thcsc men, likc other Arabs and jews sculcd in Egypt, were frontier guards or police rathcr than soldiers. By thc 2nd ccntu ry poi ice received small klero1 just likc soldiers. Thcre wcre severa! j ewish villages in Lower Egyp1 fulfilling this function, as well as Arabs on thc desert fronticrs. These men were armed, but not intended 10 fight in baule; thcy wcrc not cali ed up even for Raphia, though arming them would ha ve been a far lessdangerous step 1han arming thc clisa!Tcctcd subject Egyptians, and if thcy did not light at Raphia it is a safc bet they neverdid. Towards thc end ofthc Ptolemaic era, whcn che army of klcro11c/1oi sccms 10 have fallen to a ve ry low ebb if not disappeared, Caesar records an army of 18,000 foot :md 2,000 horsc including "brigands and pirates" from Ci licia :md Syria, exilcs and runaway sl:l\'es, in large numbcrs. Thcy may have provided the numcrous archers and slingers present. (At this ~tage the Ptolemaic army also includcd Roman legionaries who had bcen sc nt thcrc to bolster up che government; thc civil population of Alexandria, which had no formai milicary role but enjoycd thc occasionai riot, also played a part.)

Thc Ptolemies had artillery in large quant ities and builc up one of the best arsenals in the Mediterranean; thcir anifice rs did mucb to advance artillery design. Tbe kings had enough engines to lend some to allied cities, and there we re stili plenty to plague Caesar in the lst century.

Ptolemy I bad a few Indian elephants - be may have caprured a few from the regent Perdikkas, and he took Demetrios' 43 at Gaza - but his descendants were cut offby the hostile Seleucids from furtber lndian supplies.

About 270 Ptolemy II Philadelphos se nt expeditions to hunt the smaller north African elephant. Severa! pons and 22

hunting stations were es tablished on che coasts of modero Eritrea and Somalia, whence they were shipped down lhe Red Sea in specially designed boacs. Originally the elephants were trained by imported Indian mahouts, and tbus mabouts carne to be called Indians regardless oftbeir nationaliry, but at least some later mahouts were G reeks and others may bave been African, as at Canbage. T bere were 73 ofthese elephants at Raphia.

Ky rene Kyrene and lhe other G reek cit ies ofLibya had been left more or less alone by tbe last Persians and Alexander; but Ptolemy I seized Kyrene and hencefonb it remained closely tied to Egypt. lt was variously a province under a libyarches, a kingdom under a ruler ofthe Ptolemaic bouse, and interm ittently an independent state. Like ali Greek cit ies Kyrene relied on citizen hoplites, but one peculiarity was that she used four-horse chariots copied from her Libyan neighbours. The tactician Aineias says that in the mid-4th century Kyrene also used two- and four-horse wagons 10 transpon her hoplites over long distances. This would enable them to trave) relatively unafTected by the Libyan sun.

In 322 Kyrene raised 30,000 troops under elected generals, but thcsc included Libyan and Carthaginian allies.

Ophellas, Ptolemy I's govemor of Kyrene, marched in 308 to join Agathokles ofSyracuse with 10,000 foot, 600 borse, and I 00 chariots with 300 crew, plus I0,000 non-combatants. T he chariots were presumably Kyrenean; few ofthe other troops may have been, as many were Greek mercenaries, especially Atbenians who enlistcd in large numbers because ofOphellas' ties with tbeir city (be bad an Athenian wife). In 163 Kyrene mustered 8,000 foot and 500 horse, and since her Libyan allies had desened tbis probably reflects the citizen manpower. Note that the chariots seem 10 bave gone out ofuse - as in Cartbage, wbich also used them in the 4th century.

There is no specific evidence for Kyrenean troops being called up 10 fight for Ptolcmaic Egypt, unless the cavalry "from Libya" at Raphia were indeed Kyreneans, but it is likely they could be called on if wanted. Kyreneans however seem to have enlisted in thc Ptolcmaic regular army in some numbers.

THE SELEUCIDS After che fall of An첫gonos, Seleucus inherited most of Alexander's Asian empire. Later Seleucid armies wcrc bascd like thosc ofthe Ptolemies round the phalanx and beavy cavalry provided by Gracco-Macedonian military scttlers.

Though some of the senlements were cities founded by Alexander, most wc re the founda첫ons oflatcr Scleucid kings, so obviously Seleucus I had rclativcly fcw rcgular senlcr troops 10 cali on. This may cxplain why the army he took to Ipsos had relatively few infantry; there werc 20,000 foot, ofwhom some must have been Asia첫c light troops, 10 12,000 cavalry; probably no more than a third of the force at the outsidc would be settler infantry. At Ipsos Scleucus was allied to Lysimachos, who brought the allied force up to 64,000 foot and 15,000 horse. (Plutarch says "500 more" horse than Antigonos' 10,000, implying 10,500 forche allies; but Seleucus had more than this on his own! 500 is pcrhaps a corruption in thc texts for an originai 5,000.} The military senlements were concentratcd in Lydia, Phrygia, north Syria (around Antioch, wbicb became the capitai), the upper Euphrates (notably Dura-Europos) and Media. The soldiers held kleroi of land as in thc Ptolemaic army, but werc called katoikoi rather than klerouchoi. As with the Ptolemies the Seleucid scnlers were mostl y a mixture ofGrecks and M acedonians, but che Macedonian element seems 10 have been dominant, at !cast in Lydia, Phrygia and Syria. Thcre were also Iranian kacoikoi in Media, Thracians in Persis, Jews, M ysians and other Anatolians, wbile tbe Hyrkanians seni ed by Persia in Lydia seem to ha ve been assimilated to this Maccdonian military caste, as they were called "the Macedonian Hyrkanians" under the Roman Empire.

Seleucid gua rd troops included two elite cavalry units each 1,000 strong. The Companions or royal ile were recruited from the settlements ofSyria, Lydia and Phrygia. This was a standing unit, scrving in peace as well as war, and its t roopers seem to bave been recruited from the younge r generation of the se ulers, wbo served in thc Companions during their fatbers' lifetimes and then inherited the family k/eroi and reverted to tbe pan-time status ofmost ofthe settler troops. The other gua rds units probably served on che same terms. The second guard cavalry unit was che agema, recruited from the Medcs and their neighbours; though natives they probably held kleroi like their Macedonian counterpans. Either corps could escort che king in ba!lle, and they were some첫mes brigaded together undcr bis direct command, "2,000 cavalry who were accustomed to fight around him". After Media was lost to che Parthians about 160 a new agema of Macedonian colonists seems 10 havc replaced the Medcs. The elite infantry were che argyraspides, usually 10,000 strong. At Rapbia they were said to be "cbosen from the whole kingdom" which has led to their being traditionally ident ified as natives. At the Daphnai parade of 165, however, they are called Macedonians. Bar-Kochva suggcsts they were in fact picked from che military se!llements throughout the kingdom, and thus would be mostly Macedonians and G reeks. Therc are references to hypaspists in 23

the Seleucid army, and Bar-Kochva suggests they were an inner guard within tbe argyraspides; but since the two na mes nevcr occur in the same battle or description it seems more likely that they are simply synonymous and both signify the same corps. T he argyraspides are sometimes also referred to as peltasts, like the Antigonid guard infantry.

Most karoikoi, what one might cali "the troops ofthe li ne", were mobilised only in wartime and livcd on their lands in peacc, except for occasionai fonress garrison du ty. Ifthc mcthod of rcc ruitmcnt ofthe guards units set out above is correct, then most ofthe "line" troops will ha ve done a turn in the standing guards units in thcir youth, and BarKochva suggcsts t his helps account for their great loyalty to thc dynasty which is ofcen in evidence, and al so for their leve! ofmilitary efficiency whicb seems generally to havc been higher thao their Ptolemaic counterparts. Cavalry karoikoi carne mostly from tbc Median and Syrian settlements and tbe former may have included Iranians. There were 6,000 ofthese regular cavalry at Magnesia, providing the cataphracts. This is the highest strength found and must represent the maximum turoout ofthe settler cavalry. Tbe infantry karoikoi provided a phalanx of20,000 al Raphia and 16,000 at Magnesia. The Thracian settlers in Persis may have kept to their native style offighting and provided the 1,000 Thracian infantry at Raphia and thc 3,000 at the Daphnai parade. Alternatively thesc may perhaps have been mercenaries. Thracian tribal mercenaries were hired at least while the Seleucids had a foothold on the Thracian coast in the early 3rd century; Polyainos reco rds Antiochos II hiring some. Unit organisation of these regular troops is not unfortunately known but must have been similar 10 that ofother Hellenistic states. Tbe infantry had units called semaiai which may have bcen the equivalems ofthe Antigonid speirai. Cavalry were in units called 011/amoi, which also appear in Aitolian and Achaian organisation, where they are divisions of i/ai.

Semaia :md 011/amos are in some way treated as cquivalent, perhaps as the smallest tactical unit oftheir respective arms. lt seems likely from the size of che guard cavalry un its that units of 1,000 formed pan of che cavalry organisation, and it is also likely that organisation was in some sense te rritorial, so that men from the sa me settlemcnt would serve together.

There was alsoa militia, at least io Syria, recruited from those Greek cities with no role in the regular army, or where only pan ofthe population bad such a role (che capitai Amiocb may be in this class) and perhaps from the native cities. No sucb civic militia troops fought in thc well-documeoted major banles of Antiocbos Ili, but their importance may bave iocreased in the 2nd century as the kiogdom declined and many military settlements fell to Pe rgamon or Parthia. 3,000 citizen cavalry appear in the parade beld in 165 at Daphnai near Antiocb; tbey muse be from the north Syrian cities and perhaps from Antiocb itself. Antiochene citizens were certainly armed, as Demetrios II provoked discontent by disarming them. Not long afier very many of them were lost in Antiocbos VII's disastrous Parthian war of 129. The Seleucid army which fought at Azotos in 148 is called "the power ofthe cities", probably indicating that it consisted largcly ofthe civic milit ia ofthe coastal Syrian cities, tbe regular army being involved in civil war at the time. The terracottas of warriors from Seleuceia on the T igris, mostly of rl111reoplioroi infantry and light cavalry (figures 41 and 51) may represent that city's militia.

Mercenari es and irregulars Mercenaries were used, as ever, though in relatively small numbers. Many were Grceks, especially the u biquitous Cretan archers. Others may have been Anatolians; mercenary tl111reoplioroi ofvarious nationalities are known from grave monuments from Sidon, but it is not clear whether tbey belong to the 3rd century Ptolemaic or the 2nd century Seleucid garrison, or indeed both. There were 5,000 Greeks ac Raphia and 2, 700 mercenaries of mixed origin at Magnesia, perhaps boch tl111reophoroi. There may bave been as many as 10,000 mercenary infantry on Antiochos Ill's great eastern expedition; Bar-Kochva identifies the 10,000 peltasts menrioned as tbe argyraspides, but Polybios' description of che siege of Sirynx suggescs these peltasts are identica! to the merceoaries. Crecan archers appear in most armies, 1,500 plus 1,000 Neocretans (see figure 24) al Raphia, and tbe same numbers al Magnesia suggesting Antiochos III had a permanent corps of2,500 Cret:ms. Demetrios Il ruled after 147 by the power of his nume rous Cretan merccnarics, "the Cretan ryra nny'', till they were overthrown by a revolt of che regular army. Cretan rroops are sometimes termed "allies", and may havc becn rccruitcd unde r the terms ofcreaties with friendly Cretan cities. The Tarantine cavalry who appear in unspccified but probably small numbers ac Panion and Magnesia are perhaps also Greek mercenaries. ln addicion 400 cavalry, probably mercena ries, deserted from Ptolemy IV to Antiochos III in 218 and may pcrhaps havc fough t ac Raphia.

Once the Galatian cribes had settled in Asia tbey were bired by thc Selcucids, first appearing early in thc rcign of Seleucus II (247-226). Under Antiochos lii Galatian infantry fought at Apollonia and 3,000 foot and 2,500 horse fought at Magnesia. In 165 as many as 5,000 Galatian infantry appeared in the Daphnai pa rade, though chis contravened the treaty of 189 with Rome, which forbade the recruiting oftroops west ofthe Taurus. Tbc Galatians at Apollonia are called rhigosages. T his is nota known Galatian tribal namc, and the first element rhig- would seem 10 mcan " king". T hese may be a company ofmercenary warriors equivalent to thegaisarai oftbe west, taking their name from the Seleucid king they served.

24

Various Asia tic irregular troops were employed, whe1her as mercenaries, allies or subject levies. The peoples ofthe heart ofthe empire, Syria aod Mesopotamia, do not generally seem to ha ve been armed (the Late militia ofthe Syrian cities may be an exception) no doubt because 1he Seleucids feared revolts so oear the cemres ofpower - a danger hlghlighted by the risings ofthe Egyptian macl1i111oi. Despite the potentially large numbers of cavalry whlch could have been raised in Iran che only eastem irregular horsc che Sclcucids seem to have used were horse archers, 1,200 Dahai and others at Magnesia, unspecified numbers at other battlcs. lnfantry were used in larger numbers, and it is noteworthy that many were raised relatively locally to the theatre of war, whlch suggests that they were hastily raised lcvics probably lacking training, organisation and enthusiasm.

More appear at Raphia than any othcr baule; I 0,000 Arabs and neighbouring tribesmen, 5,000 Dahai, Carmanians, and CiJicians, 5,000 Medes, Kissians and Cadusians, 2,000 Pcrsian archers and "Agrianian" slingers, 500 " Lydian" javelinmen and 1,000 "Kardakes". Arabs also provided carnei croops at Magnesia, and auxiliaries for some ofthe wars against the Maccabees. They 1ended 10 switch allegiance becween Seleucids and Ptolemies as the control of PaJestine changed. The Persians were probably professional troops of higher qualiry than the rest; Persians are found on garrison duties in Asia Minor suggestiog they were full-time mercenaries. The "Agrianes" are unlikely to have bcen real Agrianes as the Selcucids had no access to their country; they may be locals armcd in Agrianian sryle, or conceivably military scttlers. Thc " Kardakes" are not to be identified with the Pcrsian kardakes but with Kurdish tribcsmcn (Kardaka to the Pcrsians). Similarly the "Lydians" may not be from Lydia but from another locai hill tribc with a similar name.

A1 Magnesia the Asiatic light infantry amounted to 3,000 Trallianoi, halfbrigaded with the Crecans, halfwith 1,500 Carians and Cilicians, 2,500 Mysian archcrs, 8,000 Kyrtii slingers and Elyrnaian archers, and 4,000 Pisidians, Pamphylians and Lykians, callcd pcltasts but perhaps ordinary light javelinmeo. T he Trallianoi are also found in Pergamene service, and there were some in the Roman-Pcrgamene army in this battle. They have been called Thracian, apparently on the assumption that they are the same as the (acruaJJy Illyrian!) Tralleis found fighting for the Antigonids. In fact Trallianoi indicates the inhabitants ofthe small Carian town ofTralles. lt would be very surprising to find the town producing very many troops, but it may conceivably bave given its name to a troop type. These Trallianoi may be slingers, as the slingers certainly present in the Roman-Pergamene army can best be identified with them. This would suit the brigading ofthe Seleucid Trallianoi with Cretans, as the mixing of equa! numbers of archers and slingers seems to bave been common Seleucid practice. " The slingers known as Kyrtii" were from Media or nearby, as they supported Molon of Media's revolt; the phrasc of Polybios quoted suggests the name indicates a class or unit of troops rather than a nationality. Also at Magnesia wcrc 2,000 Kappadokian infantry with Galatian equipment, who were allies sent by their king Ariarathes IV.

Elepha nts a nd cha riots The Seleucids were noted for their large elephant forces. Our sources say Seleucus I obtained 500 as part of a treaty with the lndian ruler Chandragupta Maurya, and that 400 ofthem fought at lpsos. Certainly he had a decisive superiority in elephants at lpsos, but 1he figure of 500 has been discarded by some modem writers as incredibly large. Tam pointcd out thai 500 was conventional for "very many" in lndian literature, and sugges1ed 1he figure should be rejected as rhe1orical. Howevcr there seems no good reason why an Indian convention should have crepi into Greek litera1ure, even dealing with lndian elephants, whlle the figures in the sources are ali murually consistent and may derive from a rcliablc originai. I sugges1 1hey may be accurate. By Antiochos I 's war against the Galatians in 273 few if any of this corps survived, and Antiochos hurriedJy ordered 20 to be sent west by his govemor of Bactria; 16 ofthem fought at the "elephant victory". Antiochos III only had 10 at the start ofhis reign, but built them up, presumably by trade, to 102 in 217. On bis great eastem expedition he exacted elephants from Bactrian and lndian kings 10 bring his force up to 150. In the mid 2nd century the Parthians cut off access to further Indian supplies. The la1er Sclcucids then actually obtained some African elephants, Demetrios II capturing some from Ptolcmy VI Philometor's army in Syria, in 145.

Asklepiodotos sc1s ou1 a thcoreticaJ organisation for elephams similar to that for infantry, in that units progrcssively double in sizc; hc givcs clephant units of I, 2, 4, 8, 16, 32 and 64. This is sometimes regardcd as irrelevant thcory, but the Hellenistic powers, and cspecially the Seleucids, maintaincd large elephant corps on a permanent footing, and must have evolvcd some organisational hierarchy if only for administrative purposcs. While tactical needs would obviously overrule any such formai organisation, it is noticeable that when elephants are used other than in one longline, theirdeployment often fits Asklepiodotos' framework. Thus Antiochos I's 16 fought in a unit of8 and two fours, while Antiochos III's at Magnesia were in two blocks of 16 and the rese in pairs.

lt was common to use light infantry in conjunction with the elephants. At Raphia thc 2,500 Crctans wcre probably csconing the 60 clcphants on one flank, about 40 men per elephant. If the 2,000 Persians and Agriancs were

25

providing the same service co rhe 42 elephancs on 1he ocher Oank, chis works out ac 48 each. At Thermopylai che elephancs have 1heir "usual guard", which may imply that a special unii was oow permanently attached to tbem; this would have obvious advamages in familiarising the men with the beasrs. This corps may be 1he "archers aod sliogers" who fought with the elephams a1 Paoion. Antiochos III may have instituted this force to support che new larger elephan1 corps he acquired in the eastern campaign, and previous practice would suggest chere were 40-50 men per elephant.

Tbe Seleucids were 1he only Hellenistic power (excep1 la1er Pontos) to copy the use of scythcd chario1s from tbe Persians. Seleucus I had over 100 at lpsos, Anciochos Ili s1ill had some at Magnesia and An1iochos IV took some on bis invasion ofEgypt. Again Asklepiodotos gives an organisation, on che same lines as that for elephants, going up to units of 64. Again this may bear some relation to rea l Seleucid administrative organisation.

PERGAMON The kingdom of Pergamon grew out oftbe treasure ches1sofLysimachos. He entrusted 9,0001alents to Philetairos, his governar of 1he Greek city of Pergamon io Mysia; Philetairos went over to Seleucus I. He never claimed independence, but kcp1 thc money which was the basis of the power of his heirs. His successor Eumenes I did assume indepcndcnce, and thereafter Pergamene kings fought sporadically witb Seleucid kings, Seleucid usurpers and Galatians, uncil the baule or Magnesia. Then in the peacc of 189 Pcrgamon was givcn the Seleucid tcrritory west oftbe Taurus. In 133 Attalos m bequeathed his kingdom 10 Rome, and it became the province of Asia.

Till 189 Pergamon controllcd no military settlements, and had no large Graeco-Macedonian popula1ion co provide a native reguJar army. Thcre was a citizen militia in Pergamon itself, but 1hough this served in dis1anc garrisons as well as defending the walls ofthe city it does not seem to have been relied on for baule. We do not know how the miliria was equipped. The bulk ofthe forces were mercenaries, ofthe usua l mixed Greek aod non-Greek origins.

They do not scem 10 ha ve formed a phalanx, as what accounts wc havc of Pergamene batties suggest che kings fought only with cavalry, peltasts, and lighter infan try, such as thc 800 cavalry, 3,000 peltasts, 500 Cretan archers and 500 Trallianoi skirmishers wi1h Rome at Magnesia. Thc "peltas1s" in 1his case were genuine lightly armed peltasts of che cradicional rype, ra1her 1han 1he cerro being used for heavier troops. 100 ofthe cavalry and 1,000 pcltasts were Acbafan allies, ve1eran troops who had disĂšnguished themselves under their leader Diophanes in skirmishcs with che Seleucids. Tbe rest may ali have been mercenaries. 6,000 foo1 and 1,000 horse took part in the Roman war with Macedoo io I 71. This seems to be che largest Pergamene force on record; the kingdom was never a major military power. Tbc cavalry included 200 Galacians; a further thousand had been sent to Greece but los1when1heir ships were captured by the Macedonians. The foot included :n leas1 150 Cre1an and Mysian skirmishers and 300 Kynii.

Ali the Pergamene forcc (excep1 2,000 foot decached for a siege) fough1 in the Kallinikos skirmish, and as chis involved cavalry and light iroops it seems all che Pergamene infan1ry were lighcly armed, many probably peltasts as at Magnesia.

Cre1ans were common in Pergamene service. Tbey servcd as carly ns 208, agains1 Macedoo. A 1rea1y of about 183 between Pergamon and a number of Cretan towns probably regulated their recruitment. Trallianoi (discussed under the Selcucids) also crop up on garrison duty. Mysians, whether as mercenaries or subjects, were often used; by the end ofthe kingdom's independence they appear as military settlers, perhaps haviog beeo inherited from the Seleucids in 189. Mysians are variously found figh Ăšng as javeLinmen, tl111reophoroi and archers; in which capacity they served Pergamon is not clear, but 100 light javelinmen mentioned in 171 could have been Mysian. Also memioned as settlers are Masdyenes, probably from Masdye in Paphlagonia. They may have started out as mercenaries, being given land after 189. How che mercenaries werc organised is not clear. The troops of one garrison forma strategia; aoo1her is divided into six hegemoniai, three of Mysians, two ofGreek mercenarics, onc of Pergamene citizeos. Perbaps every sizeable garrison farce was rated as a strategia, rather than it being a unit offixed size, and the strategia divided into hegemo11ia1: which could perhaps be 1he usual 256-man unit ofmost Hellenistic infantry. Al1ematively, hege111011iai occur in Ptolemaic organisation, and though 1he role ofthe unit thcre is not clear, Lesquier though1 it migh1 be an alternative name for the hekatomarchia.

In 189 Pergamon acquired thc Seleucid elcphant corps (the survivors of 1he 54 at Magnesia) and the militnry settlements of Lydia and Phrygia. Since elephants rarely breed in capcivity, the youngesc will bave been among those brought back from the cast by Antiochos just before 200; others may have been in Seleucid service longer, :md therefore bcen older. Few will havc lived after 150 BC, and Pergamon had no access to more. Tbe military however bccame a regular pan of the army. La1er Pergamene kings in fact increased this elcmenc by fouodiog military colonies oftheir own (see figure 40t), probably senliog mercenaries such as the Masdyenes. The result may bave bccn a dccline in the employment ofmercenaries.

26

Despite their reputation as the bulwark of Greek Asia against the Galatians, the Pergamene kings hired many Galatian troops. The cavalry in 171 have been mentioned. About 218 Attalos I hired a wbole GaUic tribe from Europe, thc Aigosagai. They proved a nuisance because their wagon loads offamilies and possessions encumbered them on the march, and they were superstitious - they refused to advaoce after ao eclipse ofthe moon, whicb they took as a bad omen. Attalos thus dismissed them.

In 183-179 Pergamoo fougbt in alliance witb Kappadokia against Pharnakes of Pontos, while in 155-4 Attalos li was joined by infantry and cavalry sent by both Ariarathes of Kappadokia and Mithridates of Pomos, under a Kappadokian prince, fora war against Prusias II of Bithynia. Once again Pergamon was revealed as a militarily weak power, unable to make headway agaiost the Bithyniaos until the allies arrived.

Finally Pergamon had a fine arsenal and an excellent store of artillery, providing catapults co Rome fora joint siege they fought in 200.

THE BACTRlAN AND lNDIAN GREEKS About 250 Diodotos, Seleucid governar ofBactria aod Sogdia, rebelled and became king ofan independent Bactria.

His successors expanded south into Aria, Arachosia and Drangiana, north·east to Ferghana, and south·east into the Paropamisadai mountains, Gandhara, the Punjab and the Indus vaUey, briefly penetrating as far as the old Mauryan capitai of Pataliputra on the Ganges. But after 170 the Bactrian Greeks were racked by civil war; they lost some lranian territory to the Parthians but were mostly overwhelmed by Saka and Yueh-Cbib nornads, who took Bactria itselfby 130. Tbc last Greek state in lndia fell about 55 BC. For much ofthis time the Bactrian Greeks (Yavanas or Yonakas to the Indians) were divided into severa( kingdoms.

Some rulers appointed sub-kings; the kingdoms were divided imo satrapies ruled by strategoi wbo probably combined civil and military functions.

Tbere were some Greeks living in Bactria before Alexander's conquest, exiles settled by the Persians; he settled nurnerous veterans and mercenaries there, and elsewhere in the cast. After his deatb Greek rebels io tbe eastern satrapies raised 20,000 infantry and 3,000 cavalry. Tbere may bave been new settlement even after Bactrian independence; certainly one king, Euthydemos ofMagnesia, seems to bave been a recent immigrant. Tbese Greeks' resentment of Macedonian rule helps explain tbeir support for Diodotos' rebellion and the subsequent maincenance of independence. We know that Greek colonies included two Cretan settlements, Asterusia in the Paropamisadai and Daidala in che Punjab; tbere was also ooe Pisidian city, Salagissa, also in the Punjab. This suggests Cretan archers and Pisidiao javelinrnen or peltast~ could be found in the Bactrian kings' armies, as these cities were very probably military colonies. Some ofthe Greeks too may have fought as peltasts, in which capacity their ancestors served Alexander, but some, perhaps most, formed a Macedonian·style pbalanx (see figure 39).

Cavalry were an importane pare ofthe armies. Euthydemos fielded 10,000 in 208, trying to stop Antiochos III's attempted Seleucid reconquest witb tbese cavalry alone. Tbc large number raised rneaos most of Euthydemos' rroops must bave been raised from the local Iraoian nobility. Judging from the troops provided earlier to Persian armies, tbc Bactrians tbemselves would field a mixrure of light and heavy cavalry, the Arachosians and otbers mostly lights. lt is probable that in the 60 years ofSeleucid rule in Bactria prior to Diodotos' revolt the locai nobles had been organised into regular cavalry units, holding their land in rerum for military service, as in Seleucid Media.

Eutbydemos' cavalry fought in units Polybios calls hipparchiai. Antiochos III's 2,000 guard cavalry could handle one of these hipparchiai in dose combat with ease, but had trouble with three; this may indicate the Bactrian hipparchiai were about 1,000 strong. A body of 500 Yonakas associated with Menander may represent a Greek bodyguard unit.

Euthydemos certainly had elepbants, and it is likely most oftbe other kings did too. Tbere are references to kings riding in cbariots, but on joumeys not in battle, so tbese were probably not war vehicles.

In India the Bactrian Greeks' armies muse have been mostly lndian, especially once the Greek rulers in India had become separated from Bactria itself, since there was little Greek settlement in India. Tbey enjoyed suppon from Indian Buddhists, since the Sunga kings who had replaced the Mauryas sponsored a Brahman reaction against Buddhism; tbc Greek king Menander was considered a Buddhist sage. The hill·tribes of Gandhara and the Paropamisadai, on the nortb·west frontier oflndia, rnay also have been supporters ofthe Greeks. Tribes such as the Karnboja, Asvaka and Bhallas were regarded by the lndians as barbarous Bahlikkas - Bactrians - because they included Iranian elemeots related to the Greeks' apparently loyal Bactrian subjects. The Paropamisadai was indeed the last stronghold ofGreek rule in India. The Greeks may have stationed lndian garrisons in provinces remote

27

from their homes; a group ofSibi from the lower Punjab are found in Madhyamika, 400 miles south-east oftheir home, and Tam suggested they were a garrison placed by the Greeks - though whether the Greeks ever reached as far as Madhyamika is uncenain. They cenainly fought in alliance with locai lndian kings, 1he anack on Pataliputra (about 150?) being undertakeo io alliance wi1h 1he kings of Mathura and Panchala, though figh1ing broke out among the allies and obliged them IO withdraw.

Saka mercenaries may also have been used, Euthydemos threatening to cali them in against Antiochos III. T he last Greek king in India, Hermaios, rnay have been allied with the Kushans (a tribe of Lhe Yueh-Chih, and later the ruling one); Tam connecrs this with a known Chinese embassy, suggesting the Kushans aided the Greeks at the ambassador's instiga1ion, and that Hermaios held thc Paropamisadai as a vassal of Han China! Kushan troops would mos1ly havc bcen horse-archers and armourcd lancers likc 1he Saka. The Han ambassador had no 1roops with him, so there is no case for Chinese aod Greek troops actually fighting together.

THE MACCABEES Judas Maccabaeus bcgan his revoll against the Seleucids in 166 BC with a small guerilla force, at first only "with nine men or thereabou1s 11 • Though his forces grew and wc hcarof armiesof3,000 and 10,000 in the early days, our sources appear to understate Jewish oumbers, to glorify Judas' successes aod excuse his defeats. Afler the liberation ofJerusalem in 164 he had enough mento seod 3,000 on :m expcdition IO Galilee, 8,000 Lo Gilead, and stili field a third army large enough to lose 2,000 in one battle, implying a tota! of20,000 at least. The manpowcr of Judaea itselfwas augmented by Jews brought in from outlying districts and possibly by others who had lived abroad and perhaps served in Hellenistic arrnies. After Judas' death at Elasa in 160 his brothers Jonathan and Simon were aga in reduced to guerilla leaders in the wilderoess; whcn Jonathan rccovcred Judaea and bccame king he could cali on enough men for the Seleucids to consider hiring 30,000 from hi m in 152, and in 142 he fielded 40,000. His brother Simon "chose ou1 of1he country" 20,000, implying this was not 1he 1mal levy, io 134.

The early forccs werc guerilla light infan try aod such probably continued to be an imponant pan ofla1cr armies.

Judas orga nised units of 1,000 subdivided into hundreds, fifties and teos, followiog biblica! precedent. But although sources rarely referto them, in pursuit ofthe policy ofunderstating Jewish strengths, thcre are hints that the army carne to include both steadier infantry and some cavalry. Joscphus calls Judas' infantry a phalanx, while the course of severa! baules shows them putting up a long and sometimes successful opposition to the Seleucid phalanx. There is a passing reference to a horseman, Dositheus, in li Maccabees' account of one of Judas' baules, Judas' pursuit ofBacchidcs at Elasa seems to have been carried out at cavalry speed, while Simon's cavalry in 134 aod Alexaoder Yannai's at the end ofthe 2nd century are specifically menciooed. The military expertise needed 10 form these forces may have been supplied by expatria1e Jews who had served in Hellen istic armies, especially thc Ptolemaic, aod retumed to help their country. Bar-Kochva specifically idencifies Dositheus as coming from an cxPtolemaic military seulemcnt in Trans-Jordao, though this seems to dcpcnd on an uncertain rcading ofthe text.

Other factors leading to the professionalisation of the army were the need to maintain standing garrisons, and the payment and cquipping oftroops by the scate, which was done by Simon's time at least.

A possible organisation for regular Judaean troops is given in 1he Rule o/ lt7ar fouod among the Dead Sea scrolls.

T his is of post-M accabean date, perhaps l st century AD, and describes an apocalyptic war with the Romans appareotly cast as the forces of darkness. Despite this cootext, the organisation, equipment and tactics of the Jewish forces show sufficient resemblance to known Maccabean practicc, and to the Hellenistic and specifically Seleucid practice which must have influenced the Maccabees, 10 suggest that Judas' or his successors' armies inspired the author ofthe scroll, and that with due caurion we can use it to reconstrucl Maccabean warfare.

The anny is based on units of 1,000 infantry equipped as Hellenistic tlrureophoroi. The biblica! and Maccabean organisation into thousands, hundreds, fifties and tens is also mentioned in the R11/eso probably reOects the internal infantry organisation. Cavalry organisa1ion is a I inie obscure but it seems rhat each 1,000 infaotry are supported by 200 light cavalry, perhaps in squadrons of50. On each Oank ofrhe army are 700 armoured "cavalry ofthe army" recruited from older men, in seven troops of 100. T he total ofthe army is 28,000 infantry and 6,000 cavalry; the infantry tota! may include slingers who are briefly mentioned. The proportion of cavalry seerns too high for historical Maccabean armi es who seem to have suffered from a shortage of cavalry, being appareotly outnumbered in horsemen at Elasa and Azotos, wh ile Simon's 20,000 men had fewer horse than a Seleucid force they could nonetheless quitc easily beat.

No foreign allies or mercenaries seem to have been used, due no doubt to the religious-nationalistic nature ofthe revolt, till the reigo of Alexander Yannai (or Jannaeus, 103-76 BC) who was at odds with rnany ofhis subjects.

28

Againsc a revolt backed by the Seleucids he fielded 1,000 cavalry, 8,000 Pisidian, Cilician and Greek infamry, and 10,000 loyaJ Jews. Note the cavalry proponion is stili low.

THE THRACIANS Thradans fearure in this era more as auxiliaries in other armies than fighting on their own account. Their numerous tribes were rarely uniced, most having their own kings though some ofthe wilder hill tribes were kingless. Before our era the Odrysai had domioated much ofThrace but their kingdom collapsed on the assassioatioo ofKotys I io 359, 10 be followed by Macedonian overlordship. After Alexander the Great's death Lysimachos ruled mosc of Thrace, dominaùng the Odrysai under Seuthes III, but his attempts 10 expand northwards were chwarted when his army was beaten by king Dromichaites ofthe Getai, a rising power. After 279 the Galatians sec up the kingdom of Tylis in the south-east, which raided the surrounding areas till its overthrow by the Odrysai about 200. Thereafter the Odrysai were again the dominane power, sometimes allied and sometimes hostile 10 Macedon, though most of the other tribes retained some iodependence.

Armies would stili be essemially the same as those describcd in 5th and early 4th century sources, made up of cavalry, peltasts (or equivaJent, though larger shields replaced the pelte), javelinmen and archers, the firsc two types being predominant. Polyainos has a reference 10 Thracian slingers in the early 4th century. The early accounts suggest 25-40% cavalry; in our era the Getai fielded 4,000 horse and 10,000 foot against Alexander, the Odrysai 8,000 horse and 20,000 foot against Lysimachos, which fits these proportions. The Getai and Odrysai had the best and most numerous cavalry, while the wild hill tribes like the Dioi and Bessoi relied on their fierce infantry.

Thucydides in the 5ch century indeed says the Getai ali fought as Scythian-style horse·archers, but though they may have produced all-cavalry forces for raids or distant expeditions, they clearly had many infantry for home defence.

The Odrysian king Kotys in 171 had 1,000 elite cavalry, probably a royal bodyguard of nobles, with an equal force oflight infantry anached 10 them.

Bithynia The Bithynians were a Thracian people of oorth-west Asia Minor, whose kings were practically independeot of Persia since about 435; they kept their independence throughout our period, becoming steadily more Hellenised.

When Xenophon fought them in 400 their army was stili tbc traditional Thracian cavalry and peltasts. There is no evidence to say for certain whether or not the Hellenised kings tried to build up a professional Hellenistic army.

There is no trace of Greek mercenaries or military settlers, the usual sources of manpower for such an army, in Bithynia. However che Greek inscription ofa Bithynian officer Menas, killed (probably at Korupedion in 281, perhaps later) leading a force ofinfantry fightiog in suppon ofadvanced cavalry, does suggest a degree ofregular organisation. Nikomedes Il1io88 BC fielded allegedly 50,000 foot and 6,000 horse, but what kind oftroops they were is not known. Nikomedes I was responsible for bringing the Galatians into Asia in 278, to use them in a civil war. Prusias I hired European Thracian troops in 219, Prusias Il had 500 ofthem as his bodyguard in about 150 BC.

THE ILLYRIANS The Illyrian tribes of modem Jugoslavia and Albania iocluded the Dardanoi, Aucariaces and Taulantoi. Many kings are simply called "ofthe Illyrians" and the term seems 10 indicate one specific tribe or group of smaller tribes as well as being a generai name. They were great raiders, both by land and by sea, using small 50-man ships. The importance ofraiding and plunder is shown by the career ofBardylis, who rose from a humble charcoal-burner to being a powerful warleader and then king, because ofhis faimess in distributing loot as much as his successes in battle.

Most ofiheir warriors were infancry with spear or javelios, supported by smaller numbers ofskirmishers with sling or javelin and light cavalry presumably provided by the tribal nobles. These types were never numerous; a Dardanian army in 200 had no cavalry or light infantry at ali, while Bardylis in 358 only had 500 cavalry 10 10,000 foot. Higher proportions ofhorse appear in small detached forces, such as the 200 cavalry and 1,000 foot serving Perseus ofMacedon in 169, or the 2,000 foot and 2,000 horse with Rome the next yea r, but these were not typical.

Polybios describes 5,000 Illyrian infantry forming up in their usual battle order and advancing in speirar~· this need not indicate that thcir units corresponded exactly 10 the 256-man Antigonid speira, but must at least indicate the use ofrelatively small compact companies rather than ao uodifferemiated tribal mass.

One pcculiarity of Illyrian warfare was that their slaves fought alongside them, some rich individuals being able 10 field severa! hundred slaves. The Illyrians' regular raiding of their neighbours must have provided numcrous slaves; in 230 for instance the enti re population of Phoinike in Epeiros was enslaved. Slaves may ha ve provided 29

some of the skirmishers, but Alexander the Great's use of lllyrian javelinmen would indicate some native skirmishers existcd.

THESTEPPES In our period the great Eurasian steppes from the Black Sea to centrai Asia were inhabited by relaced Iranian· speaking nomad and semi-nomadic tribes. As Herodotos poincs out the Greeks called them ali indiscriminately Scythians (Skutliioi) while 1he Persians called them ali Saka. lt is convenicnt to reserve che 1erm Scythian for those north ofthe Black Sea, and Saka for 1he groups north-eas1 oflran. In be1ween and moving west during our era were the Sarmatians.

Scythians In che 4th cencury the Scythians ruled north of che Black Sea from the Don co the Danube, lording ic over subject peasants who grew the com which che Scythians exchanged for Greek luxury goods, including armour. In the late 3rd cenrury Sarmatian tribes migrated across the Don and broke Scythian power. Scythian remnants survived in "Scythia minor" (che Dobrudja at the mouth ofthe Danube) and in the Crimea, where a new capitai, Neapolis, was founded and fortified. These Crimean Scythian were under heavy Sarmatian influence and may even have had a Sarmatian ruling class; Sulimirski even calls them "the Scytho-Sarmatian kingdom."

The Scythians were famous for their light horse archers, who were supplemented by better armed noble cavalry. By the late 4th century at least these were grouped in special bodies rather than being scaucred as lcadcrs ofbands of horse archers. In che 5th century Herodotos speaks of severaI nomad tribes, including the Sarmatians, as fighting alongside the Scythians, and they may havc been tributary. Most armies were ali cavalry, rclying on mobility, but in 310 an army of 10,000 horsc had over 20,000 foot. Armies fighting near Scythian bascs might include poorer Scythians fighting as infantry, but numbers like this probably include hordes ofsubject peasants.

The Bosporan Greeks The Greek colonies of Pantikapaion (Kcrch) in the Crimea and nearby Phanagoria in the Taman peninsula were the basis ofthe Bosporan state. This was ruled by the Hellenised barbarian Spartokid dynasty, with the title "governor (archon) of Pantikapaion and king of the Sindians and Maiotians". The Spartokids nevcr uscd a largc army; the Scythian peace guaramccd no foreign invasions, as Scythian kings and Greeks both profited from 1heir trade rclations, while a small force of mercenaries ensured locai order. Aincias rcmarks on the steps taken by Leukon, archon 393·353, 10 cnsurc the loyalty of his mercenaries. In the civ il war of 31 O the only "Bosporan" troops involved were 2,000 Greek and 2,000 Thracian mercenaries, che citizenry playing no pan. The usual militia organisation in the Greek cities probably never disappeared, but seems to have been fairly irrelevant.

This changed in che 3rd century when Sarmatian pressure weakened the Scythian peace. A larger mercenary force was probably beyond the state's financial means, and a national force was created, embracing both Greeks and native peoples. The nobility carne under heavy Sarma1ian influence and provided Sarmatian·style cavalry. The infantry include the nobles' retainers, probably mostly indigenous tribesmen, and Greek militias. This army may ha ve been the work of Eumelos, who ruled from 309 to 303 with the title of king. H e certainly buih a Oeet and conquered neighbouring native tribes. Iri thc civil war whereby he carne to power he had been opposed by the Scythian king, so probably felt he could not rcly on thc Scythian alliance and needed another form ofpro1cction.

The new army was not enough to keep ofTSarmatian pressure and at the end ofthe 2nd century Mithridates VI of Pontos had 10 be called in to fight ofTthe Rhoxolani.

Sarmatians and Sa~a In the 4th cenrury the Sarmatians lived east of1he Don and nonh ofthe Caspian, but in the 3rd groups such as the Rhoxolani and Iazyges crossed the Don and defeated the Scythians. These were nomads living in felt tents carried on wagons. Like ali nomads they relied heavily on cavalry, Strabo saying they regarded i1 as contemptible to go on foot. However even they probably had some infantry, who may bave guarded the wagons in battle, as Strabo includes slings among their weapons, which muse ha ve been used by footmen. Other groups settled down and were no longer entirely nomadic, such as the Siracae in the Kuban valley; these lived in hillforts, and domin::ited corngrowing peasants as well as forest areas. In 31 O thcy fielded 22,000 foot 10 20,000 borse. However Sarmatian infantry were probably as ine!Tective now as later, when Tacitus says the cavalry were the only useful pan of1he army.

East ofthe Sarma1ians were the Saka tribes ofwhom the M assagetai were most powerful; these, whose name may

30

mean "grea1 Saka horde" , were a large confederacy who domina1ed 1he eas1ern Sarma1ians. Their army comained armoured cavalry as well as horse arcbers, and infan1ry. Saka cavalry CO¡Opera1ed with Bactrian and Sogdian rebels againsl Alexander. About 135 Saka !ribes overran Greek Bactria and eastern Parthia, some invading India. When subdued by the Parthians, some remained in Drangiana, called Sakastan (Seis1an) afler 1hem, where they founded a mixed Panho-Saka kingdom under Panhian overlordship.

Parthia Ano1her Saka confederacy was 1hat of1he Dahai. One of1hcir tribcs, 1hc Parni, occupied che Seleucid sauapy of Parthyaea about 250, founding thc Parthian kingdom which cxpandcd 10 overrun mos1 of 1he Seleucid eastern provinces, imposing Panhian overlordship on locai princes in Persis, Elymais and Media Atropatene. When the Parthian army is first clearly described in the first century BC it consists almost cntirely oflight horse archers and cataphract lancers. There are hints that the early armies may have bcen more diverse. The alJ.cavalry organisation may have been established in imi1ation of the successful Saka invaders of 1he I 30's, who were nm defeated and absorbed tĂŹll the early I st cenrury BC. Originally the Parni 1hcmsclves will ha ve been similar cavalry, bui can never have been numerous. In t 29 their army was relatively weak in cavalry, 100 weak 10 face Amiochos Vll's Seleucid cavalry in a straigh1 figh1, but had a strong force ofinfantry capable ofacting in rough 1errain - this is clear from 1hcir successful ambush ofthe Seleucid horse. The Panhians cenainly called upon 1he service of1he militias ofthe Greek cities they conqu ered, and also it seems from ex-Seleucid katoikoi. In 129 after beating Antiochos Vll they incorporated his 1roops wholesale iato their own service, bui 1hese promptly desened to 1he Saka; and in generai, though che ex-Scleucid croops may ha ve scemcd a good source of infantry they were 100 likely 10 be loyal 10 the Seteucids to be used against them. As the Parthians bad no ready access 10 other Hellenistic mercenaries, i1 is likely 1ha1 most of their infantry, like 1hose used to ambush An1iochos VII, were Asia1ic. Locai hill 1ribes sometimes fough1 for the Panhians, as 1he Tapurians did agains1 Antiochos III; 01her 1roops may have been provided by the 1ributary rulers or even by peasant levies.

IN DIA lt is difficult 10 write with cenainty about Indian armies of this era because, except for the invaluable Greek testimony, 1he wrinen evidence is mostly of uncenain da1e. The main single source is 1he Arthasasrra, a manual on government and warfare auributed 10 Kautilya, the minis1er ofChandragup1a Maurya (321-297 BC). But there are difficuhies wilh this traditional anribution; the Arthasasrra does noi much use Mauryan officiai terminology, but does use many later terms, while its account ofmilitary administration differs in some respects from thc eyewitness testimony of Megasthenes, Seleucid ambassador to Chandragupta. lt is probably a mixture of Mauryan and later materiai, in A. L. Basham's words "the elaboration ofa Mauryan originai which was perhaps the work ofK autilya himsetr'. The epic poems, notably the Mahabliarata, may al so contain relevant information. The Mahabliarata describes a war which took piace long before our era, if it is historical at ali, and overall reflects the heroic, aris1ocratic spirit of an earlier era's warfare; but i1 cook roughly i1s presem form gradually between about 200 BC and 200 AD, so may have information from our period, and can, iftreated with cau1ion, be used.

In the 4th century India was divided into innumerable small tribal states, in the north descended from thc 1ribcs of the Aryan invadcrs, and mostly monarchies. The largest was Magadha in 1he Ganges valley, ruled by 1he Nanda dynasty. About 321 a M agadhan exile, Chandragupta Maurya, partly inspired by Alexander 1hc Grea1, ovenhrew 1he Nanda at 1he head of an alliance, 1he most powerful of wbose members was a king Parva1aka, possibly Alexander's erstwhile ally Poros, or conccivably bis successor. The Mauryan kingdom expanded under Chandragupta, his son Bindusara and gra ndson Asoka, 10 include ali India excep1 1he far sou1h, which was allied.

Afler Asoka's dea1h in 232 the empire began to fragment and in 180 Magadha itselffell 10 Pusyami1ra Sunga, a generai ofthe last M aurya, whom he murdered; he was a Brahmin and sponsored an orthodox reac1ion agains1 1he Buddhism of 1hc 1:11e Mauryas. Meanwhile 1he nonh-west fell to the Greeks, followed by the Sakas, and locai dynas1ies in 1he Deccan grew in power. These include tbe Andhra or Satavahana kings, who were responsiblc for onc of our major sources on 1he appearence of Indian troops, the reliefs a1 Sanchi.

Indian armies were "fourfold", comprising infantry, cavalry, chario1s and elephants. Infantry were mos1 numerous, chariots and elephants, especially the former, the most prestigious. One theoretical arrangemeni in the Mahabliarata gives proponions of onc elephant 10 one chario1, three cavalry and fivc infantry; or, assuming two men per chario1 and 1hree per elcphant (which is far from cenain as crew sizes varied; see figures 177 and 181 ), 23% elephant crew, 15% charioteers, 23% cavalry and 38% infantry. The Arthasastra 's unii compositions imply similar proportions. Recorded armies include Poros', which (following Arrian, and including the advance force bca1en before the ma in battle) was 200 clephants (a mere I W1/o with a crew of3), 420 chariots (61h% with a crew of6, which is anes1ed by Curtius), 6,000 cavalry ( 15%) and 30,000 foot (77%). The las1 Nanda king is credi1cd (taking 1he

31

lowest figures) with 3,000 elephants, 2,000 chariots, 20,000 cavalry and 200,000 infantry. Note he has far more elephants compared with chariots than does Poros, which fits the Greek authors' testimony that elephants were far more numerous (as well as fiercer) in his realm than in the west. Chandragupta is said to have conquered India with 600,000 men and to have ruled with a standing army of9,000 elephants, 30,000 horse and 600,000 foot. Proportions vary quite widely, but in generai there might beone cavalryman to every 5-20 foot, whlle adding chariots and elephants together, rhere could be one per 30-50 infantry. The two prestigious arms are thus markedly less numerous than theory would suggest as the ideai; not surprisingly, since it would be relatively easy to levy more infantry and perhaps even cavalry, less so to acquire extra chariots or elephants.

Accounts ofunit organisation are comradictory and mostly seem theoretical rather than realistic. The Mahabharaca gives a basic unit or paui of one chariot, one elephant, three cavalry and five infamry; ebree paccis make a se11amukha, three of these agulma, three ofthese a gana, three ofthese a vahi11i, three vahinis a prca11a, three prca11as a camu, three ofthese an a11iki11i, and ten a11ikinis an aksauhini(a vasc horde of2 l,870 chariots, and the same number of elephants, far larger than any real army!). The A rchasascra claims each chariot (or elephant) is accompanied by five cavalry and 15 infancry; these chariot-units are usually to be arrayed in groups of nine, five of these groups making up a chariot-division, rachavyuha, of 45 chariots, 225 cavalry and 6 75 infantry; the number of chariot-units in a division could however vary. Elephant-units were organised in the same way as charioc-units, with the same numbers of cavalry and infamry. Leftover troops which did not fit into this neat system were to be fitted in whereever appropriate. Both these organisational theories mix infantry, cavalry, and chariots or elephants (in the former case, both) in the same unit. This may reflect earlier miltary organisation when the chariot-mounted hero would be accompanied by his retainers on horse and foot, but ifit had any significance now must be purely administrative, as infantry, cavalry and so on formed separate unics on the battlefield. Other sources indicate a simple decimai organisation, with units of IO, 100 and 1,000, which may be closer to reality; while the Archasascra elsewhere groups ics units of one chariot, 5 cavalry and 15 infantry in cens and hundreds.

The army was usually led in battle by che king himself - che Archasascra assumes he will lead his own forces, leaving a regenc to govern the kingdom in his absence, while Poros and Ambhi ofTaxila both led their own arm ies in Alexander's day- or by a son or grandson ifhe were too old. The more advanced states at least had a sophisticated military bureaucracy. Megaschenes describes Mauryan adminiscration involving six boards offive members each, responsi bi e respeccively for the fleet, commissariat, infantry, cavalry, chariots and elephants. The Archasascra does not follow Mauryan practice here, but gives individuai superintendents ofhorses, elephants, chariots and infantry, and an overall army commander with responsibility for ali military affairs under che king. Horses, weapons and other equipment (at least those of che regular core of che army) were property of che state not of the individuai soldier, and were retumed to the arsenal at the end ofa campaign. Armourers, bowyers and so fonh were directly employed by the state.

The peasantry were not called upon to fighe, were in fact not allowed arms, and were left unmolested by both sides in an internal lndian war. The backbone of the army was the hereditary troops, desoended from the kshacriya warrior-nobility. These provided ali four arms; there is an account of tbe Buddha visiting four villages ofhereditary infantry, cavalry, chariot and elephant warriors. By the Mauryan era at least these were a paid professional standing force. Megasthenes reports the Mauryan army was a considerable drain on the treasury, and "when they are not on service they spend their time in idleness and drinking bouts". Despite che emphasis on training and sophisticated formations in the Indian sources, note that Megasthenes does not mention peacetime training. No doubt there was some, bue the very practice ofhanding in their weapons and horses at the end of a war indicates that training was aot constane, and this does not speak well ofthe hereditary troops' standard ofdrill and discipline. The second principal source ofmanpower was mercenaries, who may have been troops ofmuch higher quality; "Now the best fighters amoag the Indians were mercenaries, whose custom it was to trave! from one city to another as they were needed; they defended their clients with vigour and caused Alexander heavy losses." Many mercenaries will also have been of kshacriya stock bue they included advencurers of diverse origins, there being references co fugicive nobles, brahmins (the priestly class) and weavers taking service as archers. In one case a brahmin even becomes an archer's servant. Heredicary troops and mercenaries ac least usually had servants, and an lndian army on campaign would be accompanied by a vast tail ofnoa-combatants. Tbese archers were infantry, as were those hired by the Assakenoi to fighe Alexander, some ofwhom were javelinmen. Ocber mercenaries may bave been cavalry; chariots and elephants are unlikely. There is little trace of non-Indian mercenaries in our period, though it has been suggested chat Chandragupca may bave seized power with the help oftroops provided by Alexander's govemors in the Punjab.

Shonly after our era, as the Greek states in India fell, Greek mercenary guardsmen, anillerists and engineers were in demand. A third source of manpower was "guild" troops, said to be slowly mobilised and used for short expeditions. They bave been variously identified. They may possibly be guilds ofbodyguards and caravan guards, professioaal fighting men but not regular troops, employed by merchanrs but who could be called up in wartime.

32

Perhaps more likely is the suggestion that they were an urban militia ofthe artisan classes, organised on the basis of craft guilds like some ofthe militias ofmediaeval Europe. Either way, they would not be especially well disciplined or dependable. Auxiliary troops from the wild forest tribcs were sometimes employed, especially for scouting and diversionary attacks. T hey are noted as being under their own officers, as were allied troops, the standing forces of dependent kings. Deserrers from the enemy army would also occasionally be employed.

Regional variations The nonh-western mountain tribes included Iranian elements. These are che Indians whose customs Herodotos said were like those of the Bactrians, and the "mountain Indians" and Paropamisadai whose cavalry fought brigaded wich Arachosians at Gaugamela and elsewhere. The nature of their home terrain meant they had few cavalry and no chariots. Elephants were also scarce, partly no doubt because ofthe nature ofthe country, partly because they would have to be imported from further south, and the hill tribes were none 100 rich. The Assakenoi fought Alexander witb 30,000 infantry, plus 7,000 lndian mercenaries, but only 2,000 cavalry and 30 elephants.

Some ofthe lndian tribal states had never adopted monarchy but remained oligarchie republics, ruled by a citizen class claiming common descent and caste status, with an elected head of state and officers including an army commander. Alexander me1 severa! such kingless peoples in the Punjab, such as the Malloi, Kathaioi and Oxydrakai. Some kept their identity and autonomy throughout the Mauryan empire, tribes like the Arjunayanas and Yaudheyas emerged as independent powcrs aftcr the fall ofthe Mauryas. An interesting peculiarity about the republican armies Alexander met is that none contain any elephants. Thus the Malloi had 80-90,000 infantry, 10,000 cavalry and 700-900 chariots; the Agalasseis had 3,000 cavalry and 40,000 infantry; and the Sambastai had 6,000 cavalry, 60,000 infantry and 500 chariots. Thcre are no detailed figures for the Kathaioi but the account of their battle with Alexander mentions cavalry, chariots and infantry, but no elephants. Since the Greek writers were interested in elephants, one ofthe most striking fearures to them oflndian warfare, and men첫oned even small forces of 15 or 30, ic is not likely that they have merely overlooked the elephants in these cases. Obviously the upkeep of war-elephants was a costly business; kings could afford it, but the republics had no such rich centrai govemment, and while some of 1beir rieber citizens could afford to maintain chariots, few ifany could afford elephants. Most of the republican tribes claimed to be of ksliatriya status; this does not by any means indicate that they were all professional soldiers, but probably does mean tbat the republics could mobilise most oftheir cirizen manpower, a bigber proporti on oftbe population tban tbe kingdoms could cali on, while altbougb tbese warriors would not be as well disciplined or experienced as tbe best professional troops, they probably had a greater overall esprir de corps than che rather beterogeneous royal armies.

Tbe extreme soutb of India was tbe bome of che Tamils, who emerge in the Mauryan era organised into tbe kingdoms of che Cbolas, Pandyas, Satiyaputras and Keralaputras. Megastbenes credits che Pandyas witb 500 elepbants, 4,000 cavalry and 130,000 infantry. Chariots were not unknown, Tamil poetry ofthe early centuries AD mentioning armed nobles in cbariots, but were clearly rare; nor are cavalry very plentiful. This reflects tbc perennial scarcity ofborses in the soutb, wbere tbey do not breed well. Later, Arab and then Portuguese borseimporters were 10 bave an economie stranglebold over southern kings. The Tamils in this period were fierce and warlike, apparently mucb more bloodtbirsty and rutbless th~n che rela첫vely resrrained and cbivalrous nortbem warriors. Tbe kingdoms and tbeir tesser feudatory lords were engaged in almost continuai warfare; che kings were primarily war-leadcrs. Tbere are bints ofmassacres and even cannibal feasts after victory. Late in our era or soon after, tbe Pandyas ac least bad inner palace guards of armed Greek women, and outer guards of male Greek mercenaries.

CARTHAGE Canbage was 1be larges1oftbe Pboenician cities in Africa, domina첫ng the other Phoenician settlements of Africa, Sicily and Spaio. Rome called ber people Poe11i, "Phoenicians", whence "Punic wars". Sbe was essentially a commerciai and naval, nota military, power, and her armies usually consisted mostly of mercenaries. Cartbaginian generals were elected, probably by the se nate bue some첫mes confirmed by che popular assembly, with no time limit setto 1beir tenure of command; this allowed che emergence of professional experienced generals. Tbey were drawn from che Canhaginian nobility, bue probably only from a few military families witbin tbat class. Tbe Canhaginians, like ocber powers, often punished unsuccessful generals, sometimes going to tbe extent of crucifying tbem pour encourager /es aurres. Mercenary troops were generally under officers of tbeir own nationalities, but only in the case ofXanthlppos the Spartan does a foreigner seem to bave risento bigb command.

Carthaginian citizens, as well as providing officers, would fighe themselves in emergencies. Some 4th century armies also bada core of citizen troops who would even serve overseas. A "Sacred Band" of2,500 citizen infantry 33

"distinguished for their valour, reputation and wealth" and also "by the brilliance oftheir armour aad the slow pace and strict discipline of their ad van ce" fougbt in Sicily in 341, wbere they were wiped out. 2,000 citizen troops were in Sicily in 311, perhaps a similar unit, and apparently the last use of citizen troops overseas. The Sacred Band was reformed, beaten again at Tunis in 310, and is not heard ofthereafter. Larger citizen forces rurned out in emergencies, but these seem to ha ve been modest fora city ofthe size and wealth ofCarthage. At the Bagradas in 255 there were 12,000 foot and 4,000 horse; 10,000 troops mobilised against che rebel mercenaries in 240. Both these armies did contain some mercenaries as well as the citizen levy, which implies a citizen turnout ofperhaps 8,000 foot and 2,000 horse at the most. The reason is that Carthage lacked che class ofsmall farmers working land oucside the city who provided the bulk ofGreek or Italian citizen infantry; most ofthe land around Carthage formed large estates owned by the nobles and worked by harshly treated subject Libyans. These estates also provided che fine horses which contributed to Cartbage's cavalry strengtb. As ever, the civic cavalry was levied from che rich, "some of them belonging ro rhe wealthy and noble families". Probably the Sacred Band was the only citizen infanrry force with che training to fight as a phalanx, rhe generai citizen levy fighting in more individuai sryle and in looser formations (see figure 98). Although usually inexperienced, Cartbaginian citizen infantry usually fought very weU, "when ar bay, they showed frantic and extraordinary courage".

On what terms, ifat all, the other Punic cities had to supply troops is not cenain, bue chey seem occasionally co have pur their citizen infantry into the field. The " Poeni" at Ibera and the "Carthaginians" at Ilipa were probably the civic levies ofthe Punic towns ofSpain.

The mercenaries Carthage could afford co employ large armies mostly or entirely com posed of mercenaries; 70,000 infantry and 10,000 cavalry and chariotry served in Sicily in 341, while Hannibal sraned his march from Spaia co ltaly with 90,000 foot and 12,000 horse. Cavalry usually provided I 0-25% of an army, no vast proportioo but usually enough ro give a decisive su periority over Greek or Roman opponents. The mercenary forces were extremely diverse; men from almost every nati on of the western Mediterranean are found at some stage in Carthaginian service.

The largest concingencs were raised from Carthage's African subjects. These troops are usually called Libyans, chough Livy calls them Africani and Libyphoenicians are occasionally mencioned (Polybios memions 450 "Libyphoenician and Libyan" cavalry Hannibal left in Spaio; Livy menrions ooe ofHannibal's officers, Munines of Hippacra, called a Libyphoenician and "a half-breed African": Libyphoenician was originally used simply to mean Phoenician settlers in Libya, but carne to be used for half-breeds, or Libyans who had adopted Punic culture). lt has been debated whether the Libyans were technically mercenaries or subject, conscript troops, but however they were raised they certainly behaved like mercenaries; Libyans formed the majority of che mercenaries who rebelled in 240, because their pay was overdue, even before thc rebel force was augmented by disaffected Libyan peasams. Canhage had provision fora large garrison -Appian says there were barracks and stables for 24,000 men, 4,000 horses and 300 elephants built into the walls. The garrison, whether or nor it is was this large, probably constiru ted a small standing army ofLibyans (perhaps including Libyphoenicians and even professional soldiers of other races) who would be reiaforced in time of war by newly recruited Libyan units which would be outfitted from Carthage's vasi arsenals - 200,000 infantry panoplies were said to have been surrendcred to Rome in 149. Wben, after rhe fĂŹrst war with Rome, Hamilcar Barca began 10 consolidate a large province in Spain, a second standing army developed there, since warfare was virtually consram. This roo was largely Libyan, represenred by che 11,850 Libyan infantry and 450 horse left by Hannibal in Spain and the perhaps 20,000 African infamry be rook to ltaly ( 12,000 reached Italy). He seems to have had no Carthaginian or African cavalry with him, except perhaps a small bodyguard, as only Spaniards, Gauls and Numidians are listed among che cavalry at Cannae.

The other major African pan of the army was the famous Numidian light cavalry. These were used in the 4tb century, though they are nor then prominent. By che late 3rd they were raised from allied kings rather than as pure mercenarics, and their kings could command them, as Syphax did at Zama. Cavalry were levied from severa! tribes - Massyli, Massaesyli, Maccoei, and also tbe Moors and Gaeruli who are sometimes not reckooed strictly to be Numidians. A few Numidian light infantry may also have been used, and wheo Syphax trained an infantry force he provided allegedly 50,000 foot and 10,000 borse co an allied army in 204. Although che Numidians after this date mostly sided with Rome, there were a few Numidian and Moorish cavalry at Zama, cwo exiled leaders with 6,000 men joioed Carthage against their own king Masinissa in 150, and a few fought on during che last war with Rome.

The light javelio infantry common in Carthaginiao armies are ofren noi identified by nation; some are called Moors, bue they probably included Numidians aod Libyans. Moors also occasionally provided archers.

Spanish troops fought as mercenaries in most early armies, but were only used in very large numbers when Carthage's Spanish province was eolarged after the first Roman war. le was 1hen possible to raise large armies from subject Spanish tribes; these mass levies were unreliable and unenthusiastic as they showed at Ibera and Ilipa, and 34

many ofthe tribes were quick to desert to Rome. Others were ofbetter qualiry. The army Hannibal took to Italy originally included many unenrhusiastic Spanish levies, some ofwhom were clismissed to their homes before he crossed the Pyrenees, while other expendable troops were left to garrison northern Spaio. Both Spanish infantry and cavalry remained, and 8,000 infantry reached Italy; perhaps a third ofthe 6,000 cavalry were also Spanish.

These were troops of high quality, loyal to their generai, experienced and well disciplined. Carthage's Spanish troops were mostly Iberians but included Celtiberians and Lusitanians. 4,000 good Celtiberian mercenaries crossed to Africa in 204. These were close-fighting infantry (scutan"1) as were most of Hannibal's men, though skirmishers (cae1ra11) may also have been used. Polybios describes Spanish infantry organised into speirai, Livy in cohorces; neither need imply their organisation exactly followed Hellenistic or Roman patterns, but probably give an indicarion ofunit sizes. Balearic slingers were effective but never numerous; Hannibal left 500 in Spaio, sem 870 to Africa, and probably had less than I ,000 himself. As archers were rarely used, the Balearics were often the only long-range missile troops in a Carthaginian army.

Sarclinian troops were used in 393 but perhaps not in our era. A few Ligurians were bired; s.-0me were sent to Sicily in 342 and presumably fought at tbe Krimisos, Hannibal left 300 in Spain; many joined his brother Hasdrubal's attempt to reach Italy; some were at Zama. A Carthaginian officer, one Hamilcar, was stili stirring up the Gauls and Ligurians against Rome in 200, after the war was over. Mercenary Gauls also served at Krimisos, wbile in tbe fìrst war with Rome Hamilcar Barca bad about 3,000 in Sicily, of whom 2,000 joined the mercenary revolc. Large numbers ofGauls were only used when Hannibal persuaded the Boii and Insubres to support his invasion ofltaly; they provided perhaps 9,000 foot and 5,000 cavalry at the Trebia, nearly a Ihird ofhis infantry and halfhis horse.

Gauls also joined later Carthaginian ventures such as Hamilcar's expeclition.

Greek mercenaries, accorcling to Plurarcb, were fìrsc bired in 340, because after the Krimisos the Carthaginians "had come to admire them as irresistable troops and tbe most warlike ro be found anywhere" . Occasionai early references may ali be to mercenaries in tbe bire of Sicilian Greek tyrants, who sornetimes sided with Cartbage againsr Syracuse. Most Greek rnercenaries were hoplites but in 309 Greek cavalry are found fighting bravely against Agathokles ofSyracuse in Africa; tbeir determination is explained wben we learn that oftbe 1,000 Greek prisoners taken, 500 were exiled Syracusans, political enemies of Agathokles. In the firsI Roman war Greek mercenaries were stilli used, notably tbe Spartan officer Xantbippos, and in tbe mercenaries' revolt whicb foUowed many assorted Greeks, "mostly deserters and slaves'', took part. Livy claims rhat a "legion" of Macedonians fougbt ac Zama, buI rhis is patene nonsense and regarded as such by most modem writers. These Macedonians are not mentioned in Polybios' more reliable account; Macedon was now at peace with Rome, and since king Pbilip V bad not venrured to support Cartbage even wben she was winning be was bardly likely Io start now. In any case, be could nor have transported troops to Africa as Rome controlled tbe sea, and bis own fleeI was fully engaged elsewhere.

The whole story is a Roman annalisr's fiction (calling the Macedonians a "legion" bims at that, as Ihe annalisrs tended IO misuse anachronistic Roman terminology) inspired by Rome's later bostiliry to Pbilip.

A few Italian mercenariesare found in early arrnies; 800 Carnpanian cavalry in 410, 200 Erruscans in 311, wbile one of tbe rebel mercenary leaders was a Campanian. Larger Italian forces were used wben Hannibal invaded Italy; Capua in Campania, most of tbe Sarnnites, the Lucani and the Bruttii joined him. He also had some Roman deserters in bis army, ar one point putting rbem ar rhe bead ofhis column ro bluffhis way inro a rown. Some armies under Carthaginian command were almosr entirely Italian; one at Beneventum in 214, under a Canhaginian generai, had 17 ,000 infanrry, mostly Bruttii and Lucani, and 1,200 cavalry, mostly Numidians and Moors witb a few Italians. The Brunii, perbaps the mosc derermined of che allies, mobilised 15,000 men on their own. Some followed Hannibal to Italy and foughr in tbe rbird line at Zama along witb che orher veterans (the impression given

by Livy, tbac this line was ali halian, is noc correct.)

Ch ariots and Elephants In tbe 4tb century tbe Cartbaginians made considerable use offour-horse chariots. It is not clear wbetber the crews were Libyans, operating in native style, or Carrhaginians, tbe predecessors oftbe citizen cavalry. There were 300 cbariots in Sicily in 343, allegedly 2,000 aI Tunis in 310. Thereafrer chey drop ouc ofuse, probably because rbey were relatively ineffective and elepbants were a promising replacement.

Tbe Carrhaginians first rnet war-elepbants when they fought Pyrrbos in Sicily in 278. Tbey must bave made a great impression as by 262 Canhage bad an elephanr force ofher own. Carthage mosrly used the small north African elephant. Ptolemaic Egypt may have provided assistance and tbe fìrst mahouts, and indeed a few Inclian elepbants seem to bave found rheir way imo Cartbaginian employ, which can only bave come via Egypr. Tbe mahouts are called Inclians, but this seems to have become a generai term for elepbant drivers ,·egardless of nationality. A few true Indians may ha ve come, via Egypt, to scart the corps. Later mahouts included Negroes and Moors. Appian says

35

Canhage had stables for 300 elephants; this number was never fielded, the 140 used in Sicily about 252 being the largest number known. None were used after Zama.

THE NUMIDlANS The Numidians relied chicfly on thcir cavalry, Livy stating that before thc arrivai ofa Roman cmbassy co Syphax in 213 BC they "had no experience ofinfantry warfare and were uselcss except on horseback 11 • These cavalry were provided by those nobles, herdsmen and hunters who could provide their own horses and arms. In 203 Syphax armed and mounted peasants at his own expense, as an emergency measure, witb the avowed intention of calling up every man of military age in the country; tbis was unprecedented, though Jugunha was to be reduced to the sa me straits after a defeat in 109 BC. The only pan ofhis army that had stayed loyal had been a cavalry bodyguard; no doubt earlier leaders had similar guards.

In 213 the Roman centurion Quintus Statorius organiscd an infantry farce for Syphax "very much after the Roman pattern, gave them instruction in forming up, manocuvring, following the standards, and keeping formation, and accustomed them to thc various military duties, including fortification, and ali so succcssfully that the prince soon carne to rely on his infantry no less than his cavalry". This turned out to be a bad move on the Roman pan as Syphax then opted forche Canhaginian side! He apparently took the lesson to hean, since when raising a new army some years later he organised the foot into cohons and the cavalry into turmae after the Roman example. We are not told what unit organisation was used before tbis, but it was probably fairly rudimentary. After 213, infantry usually (though not always) exceeded cavalry in Numidian armies; Syphax had only 10,000 cavalry to 50,000 infantry in 204, though 60-70% in fantry is more typical. A few infantry may in fact bave been used before the Roman mission, Polybios implying that ali Mago's ambush force at the Trebia, cavalry and infantry alike, was Numidian. The Moors, much of whose territory was mounta ins, used numerous light infantry. The majority ofbotb nations' foot was light javelinmen but there were smaller numbers of archers and slingers. Appian describes the Moors in Hannibal's first line at Zama as archers, while che Numidian king Micipsa sent clephancs to the Romans in Spain in 134 "wich tbe archers and slingers wbo were usually brigaded with them".

Early Numidian princes had no elephants - Carthage would no doubt bave not encouraged it - but by 191 Massinissa could lend 20 to Rome, and they continued to be an important part of Numidian annies, Jugurtha having 44 in tbe late 2nd century, Juba I 120 in the 1st. By Jugunha's day bolt-shooting artillery was available 100, at least for the dcfencc oftowns. These engincs probably carne from the Libyphocnician towns now in the kingdom.

TRE SPANlARDS Spain was inbabited by severa! peoplcs, but except the coastal Greek and Punic colonists, they ali fought in broadly thc samc manner. Thc south and east was the count ry ofthe lberians, tough and warlike though the Turdetani, who lived in the prosperous south whicb had once been the kingdom ofTartessos, were relatively quiet and orderly.

Severa! invasions bad settled Celtic tribes in the cou ntry, especially in the west; the Celtici ofsouthern Ponugal were civilised by their Turdetanian neigbbours, and other Celts seem 10 bave adopted Spanisb weapons. In the centrai Meseta Celts and Iberians merged to fonn the four Celtiberian tribes, probably tbe most forrnidable fighters in Spaia. l t is not clear whether the tribes ofthe western highlands - Lusitani, Cantabri, Astures, Vascones and others - were l berians or not; Diodoros certainly regarded the Lusitani as Iberian. AIJ tbese mountain tribes were a rough, hardy and simple people, regularly practising with weapons and living "a life of brigandage and uninterrupted war" as Strabo says.

The tribes were never united, though alliances of two or more were frequent, and mercenaries wcrc hired to augment tribes' strengths. The Celtiberians in particular served as mcrccnaries, the Turdetani hiring 10,000 of them once. One tribal alliance was said to field over 100,000 men; but more realistically, the 35,000 men the Celtiberians raised in 181 was larger than their usual armies. Similarly, Indibilis in 205 raised 34,000 from the Ilcrgetes, Ausetani and others.

4,000 ofthese were cavalry, or about 12%; in the I st cemury BC the Lusitani gave tbe Roman re bel Senorius 4,000 foot and 700 borse, about I 5% cavalry. These seem to be typical proportions. T he cavalry were no doubt raised from the tribal nobles, :md would provide thc commanders' bodyguards. These guards were sworn not to survive tbeir master, but to die with him as a sacrifice if he fell, a practice thc Romans knew as devotio. The result, as Plutarcb drily remarks, was that few Spanish commanders bad largc retinues.

There were two types ofinfantry, j.avelin skirmisbers called caerrati and heavier scutarii: who werc the majority type. One third oflndibitis' army was cae1rat1: probably typical. Lusitanian infantry were ali caerrat1: but this may be misleading as some of them were heavily equipped. Skirmishers included slingers, and conceivably a few 36

archers, as arrowheads have been found; but they are never memioned in baule, and tbe bow may have been solely for hunting.

Silius ltalicus claims Saguntum, a Greek-influenced Iberian town, used catapults when besieged by Hannibal; be calls rhe engines Phocaean, indicating they were supplied by Massilia (Marseilles) a Phocaean Greek uading colony.

THE C ELTS Some Celtic tribes were ruled and led in war by their kings; othcrs wcre dominated by the nobility from among whom a war leader would be elected. In either case, the nobles and their servants and retainers were an important military force, provicling the cavalry and chariots. The power and the unruliness ofthese nobles and tbeir retinues seems to bave increased with time, at least in Gaul itself, as commoners became the clients of no bi es. Caesar records tbat tbc retinues ofthe Aquicanian nobles were particularly loyal to the poiot offanaticism, bound by oath not to survive tbeir leader. These men were called soldurii; one chiefhad 600 ofthem. This version ofthe retinue follows Iberian practice, and the Aquitani were a mixture of Celtic and lberian stocks; it was thus probably not found among the other Celtic nations. Chariots are last mentioned in baule in 225, but they are still shown on Roman coins in the late 2nd century, when some tribes may have continued to use them. By this time the great majority of the Celtic rnounted 1roops will however have been cavalry.

The noble moumed 1roops could be a major pan of1he army, up to 30% or so; tbus at Telamoo tbere were 50,000 foot and 20,000 cavalry and chario1eers. There seems to bave been a disùnct warrior class, tbe poorer free men not being expected to bear arms, 1hough perhaps they might in a naùonal emergency. This probably explains the relatively high proportion of mounted troops, the nobles' retinues forming a higher proponion of the warrior classes than of1he 1ribe as a whole. In baule each re1inue probably acted in some seose as a unit. Tbe infantry seem to bave formed one undiffcrentia1ed mass, or in 1he case ofan allied army, eacb tribe's infaotry seems 10 bave fough1 in one body, with liule or no internal organisation. Peter Connolly bas suggested tba1 tbc Celts' use ofuumpets and standards proves they fought in distinc1 companies, 001 ao unorganised mass, but we do oot know that lhesc had a definite milicary function. The instruments may have been chiefly intended to boost morale by their "dreadful din"; they ali seem to have blared fonh a1 once ratber 1ban have been used to signal orders. Similarly the s1andards may ha ve had rcligious ra1her 1h:m mili1ary functions, or ha ve served 10 dis1inguish the nobles' re1inues ra1her 1ban the infantry. Livy docs indccd have a rcference 10 Celtic 111a11ip11li, bui bis annalistic sources use Roman terminology indiscrimina1ely for foreign armies, so little weigb1 should be placed on this wilhou1 supporting evidence. However mos1 of 1he larger 1ribes were federa1ions of many srnaller tribes or clans, whose men will naturally ha ve fough11ogc1hcr, and 1his s1ructure could be the basis of at least a rudimentary milicary organisa1ion.

Polybios records 1hc cxis1cnce ofa community ofmercenary warriors living in the Alps near the Rhone, under 1heir own cwo "kings", who joincd thc invasion ofltaly in 225 for pay and loot "witb a richly equipped and formidable force". They wcre callcd gaisatai, which Polybios says means "mercenaries" but actually means "spearmen" (gaesum is uscd in Latin for javelins, cspecially Celtic ones; gae is a type oflrisb spear). Tbey represem a warrior order ou1side normai 1ribal socie1y, providing merceoary infantry of notable courage and ferocity. They are comparable to the Irish mercenary bands, 1hefia11a, orto a sort ofCeltic Jomsburg. As noted under the Seleucids, 1he Galatian rhigosages may be a similar arder. So may the 10,000 Bastarnai cavalry and 10,000 supporting ligh1 infantry Perseus of Macedon 1ried to hire; 1hey were certainly professional mercenaries who "had no other trade bui war".

Thc Cehs were ncver unifìed, but often tribal coalitions would form, such as that whch invaded ltaly in 225. These could some1imes include Ligurian 1ribes; in 200 the Gallic Boii, lnsubres and Cenomani, and the Ligurian Cclines and llvates, fìelded 40,000 men against Rome.

The Galatia ns Some Celtic tribes ofthe Balkans invaded Macedon, Greece aod Thrace in 280, wi1h an army allegedly of 152,000 infantry and 61,200 cavalry. After they were repelled from Greece, some founded the kingdom ofTylis in Thrace, while three tribes, the Tectosages, Trocmi and Tolistobogii, crossed to Asia. Thcy evenrually settlcd in the plateau of Phrygia, which ca me to be called Galatia, organised iato a loose tribal federation, lording it over subject Phrygian peasants, and raiding the surrounding countries. Their origina] strength of20,000 warriors increased to allegedly 50,000 foot and 10,000 horsc for the two tribes ofthe Trocmi and Tectosages alone in 189. Originally at least they also had chariots; in 273 they had 160 convemional Celùc two-borse chario1s and 80 scythed vehicles, probably captured from the Seleucids. A cbario1 yoke appears amoog tbe trophies carved at Pergamon in 1he early 2nd

37

cenrury, so their absence in the battles against Rame in 189 may be due only 10 the mountainous terrain where the fighùng took piace.

In 280 some Greeks, Thessalian and Ainianian nobles and their followers, joined the Galatian invasion. Some Paiones too seem 10 have been swept up in the migration and crossed to Asia, as a Lykian city recordcd its delivcrance "from the Galatians and the Paiones". In 189, 4,000 Kappadokians and mercenaries of the king of Paphlagonia supported the Galatians against Rame.

ROME The Roman army was based on legions of heavy infantry supplied by property·owning citizens. These citizcns carne not only from the immediate area ofRome, but from various cities which had received Roman citizenship as a result of conquest or alliance. l n some cases this was civitas sine suffragio, involving all the privileges and duties of citizenship, including military service, bui not the voce. This pania! citizenship was first granted to Etruscan Caere in 353. After the Latin War of340 some Latin towns and the Campanian nobility received full citizenship, other Latins and the bulk ofthe Campanians the civitas sine suffragio. Ali these served in the regular legions (thus Livy mcntions Caeritans in the army in 302) though there were occasionally extra Campanian units - cavalry at Scntinum and a legion sent 10 garrison Rhegion in the Pyrrhic war. Rorne's citizen manpower was thus far grcater than that ofany single city; Polybios says the census rolls for 225 listed " Romans and Campanians", i.e. Roman citi· zens, at 250,000 foot and 23,000 horse. Casualties and Campanian revolt reduced this during the second Punic war; thereafter the figures rise ùll 164 but decline slowly thereafter.

As well as ciùzens Rame called on aJlied troops, at first from her Latin ncighbours and then from ali ltaly. Except in thc Laùn War, a Roman legion would be accompanicd by an equa! sized or larger farce ofallies. Polybios' figures for available allied manpowcr in 225 tota! 340,000 foot and 41,000 borse. This immense reservoir of manpower must be the chief reason for Rome's rise 10 dominate the Mediterranean after the unification ofltaly. Earlier, the absorption of Campania seems 10 havc been tbe crucial step which gave Rame thc rcsources 10 take on any combination ofothcr ltalian powers.

Ofcourse notali this vast manpower could be mobilised at once. In the 4th ccntury Rome expected 10 ficld only four legions for most wars, about 20,000 citizens and 20,000 allies, eacb consul commanding two legions. But in 225 the equivalent of 14 legions wcre ficlded, and in the war with Hann ibal severa! years saw over 20 legions mobilised 25 in 221, which meant about a quaner ofa million men in arms. The two-legion armies which were adequate 10 defeat the Hellenistic kingdoms seem paltry by comparison.

The result ofexpanding military needs, howevcr, was a graduai but fundamental change in the nature oftbe army.

Initially it was recruitcd from the adsidui, those classes able ro afford their own equipment; the prolc1arii below the minimum propeny qualification played no pan except in dire emergencies. Esscntially it was a part·time farmers' army, like any Greek hoplite citizen force. But the largcr and longer serving armies needed for the war with Hanni bal led to a drastic cut in 1he propeny qualification, from 11,000 asses at the start ofthe war 10 4,000. Later, about 130, it was cu t again to 1,500 asses. M any oftbe poorer men thus recruited could not afford their equipment, and wherc necessary it was now provided by the sme and the cost reclaimed from pay. The theory of a part·time citizen army was increasingly becoming a fiction, and givi ng way to a professional proletarian force, virtually a standing army as it was necessary 10 keep semi-permanent forces abroad. In 107 Marius began openly 10 recruit without regard to the property qualification; hc did liule more than forma lise an existing situation, especially as the more prosperous were increasingly reluctant to serve abroad.

The lower classes could be called up in emergencies, by a wmulws or mass levy; thus noi only old men bui ex·slaves were mobilised in 296, and the profeta rii were armed by the state to defend the ciry against Pyrrhos. In 216, 8,000 slave volunteers (volones) and 6,000 convicts and debtors were recruited.

Command was assigned 10 Rome's elected magistrates. The two consuls usually led an army each; when more than two forces were operating other magistrates such as praetors would be called on. In national emergencies a dictator would be chosen to take supreme command, with a magis1er equitum as his subordinate. Unfortunately, the rwo consuls or the dictator and his lieutenant could often be personal or politica! rivals, which did not help unity of command. One compromise sometimes reached was for two consuls to command on alternate days. Since the term of office of all these magistrates was one year, it was sometimes necessary 10 prolong a general's command after his term of ci vii office had expired.

38

The legion The legion, legio, had no overall commander, being officered by six 1ribunes, tribuni militum, responsible to 1he overall commander ofthe army - who would in many cases have only 1wo legions.

We ha ve two accounts oflegionary organisa1ion in this period; Livy describes 1he mid-4th century legion, Polybios the mid-2nd. The transition between the two is obscure. Livy's legion has 5,000 infantry. Whereas in earlier armies the troops were divided up according to 1he equipment 1hey could afford, organisation is now on the basis ofage and experience. The legion forms up in three lines, the first of 15 111a11ip11/i of liastat1~ "1he flower of the young men", with 20 skirmishers called leves attached to each mani pie. The second line is of 15 ma11ip11/i ofpri11cipes, older men.

Probable strength ofthe ma11ip11/11s is 60 men and 1wo officers, but as 1he liastati seem to be a picked force while the pri11cipes were the mainstay ofthe army, 1he former may perhaps ha ve originally been less numerous. The third li ne is of 15 ordi11es, each ordo divided into three 'l/exilla, one cach of triarii, who are the veterans, younger men called rorari1~ and accensi, "the least dependable". Each 'l/txillum has 60 rankers, 1wo ce11111rio11es in command, and a standard-bearer. Livy's account has been queried on severa! grounds; it seems generally sound but the rorarii and accensi are problems. Livy makes no menti on of either in his account of 1ac1ics, and there says 1here are no reserves behind the triarii, though earlier saying 1he rorarii and accensi deploy behind them. His association ofthese three classes would imply rorarii and acce11si are hcavy infantry like 1he rriarii~ bui other sources (Varro on tbe Latin language) say rorarii are light troops; they are called swift, and cven Livy has them running imo action. Il seems clear that rorarii are ligh1 infan1ry, anached 10 the triarii in thc same way as 1he le-ves are 10 1he lrastati - unless one is a mere duplication ofthe otber, wbich the different numbers make unlikely. Thc word accensi means anendants or supernumeraries; Varro interprets this as mili1ary scrvants, and tbcy clearly cannot be a regular combatant part ofthe legion. Livy only records them in action once, in 340, and 1hcre 1hey are used in a ruse wbicb may be 1he key to their iden첫cy; it depends on their presence being a surprise 10 their Latin adversaries, who, though organised in Roman style, bave no accensi of1heir own. Thc La1in War was a dire emergency for Rome, witb ali her allies turned enemy, and it seems likely that the acce11si werc levi ed to met 1his cmcrgcncy from thc poorcr classcs wbo would not usually fight. This explains their name, as they were supcrnumerary to the usual strength ofthe legion, explains why they were unreliable, and why the Latins did not expcct them. Thc Latin war may thus bave been tbe only occasion that they fought.

In the later Polybian organisa첫on the legion normally has 4,200 infantry - whicb, ifLivy's 5,000 includes over 900

accensi who were additional to the usual strength, may be no real cha ngc. Thcrc are 1,200 ltastati in ten manipuli of 120, 1,200 pri11cipes organised in tbe same way, and 600 tria rii also in ten ma11ip11li, bui of only 60 men like che 4tbcenrury maniple or vexi/111111. Each 111a11ip11/11s has two ce11t11rio11es, two optio11es as their subordinates, and two standard-bearers. Tbe remaining 1,200 men are ligbt infantry called veliw, divided for adminis1rative purposes among che ma11ip11/i. The youngest and poorest men serve as ve/ites; as before, the lrastati are also young, tbe principes older, and tbe criarii are veterans. When necessary legions could be larger, up to 5,000 or even (forche Pydaa campaign) 6,000 infantry; if so tbe tria rii remained at 600, wbile the othcr classes incrcased in number, so a 6,000 man legion would bave 1,800 of each class. As the complement of officers shows, each mani pie was divided into rwo centuries, no longer oftbe 100 men [beir name indicaces.

lt is noc clear wben tbc Livian organisatioa gave way [O 1he Polybian, or whether the change was dras1ic or graduai.

However, a reform oftbe ligb[ infantry seems to bave taken piace in 21 1. In that year, after a successful use ofligh1 infantry in co-operatioa witb cavalry, Livy claims that velites were incorporated in the legions for the fim time. In fact he uses ve/ices for skirmisbers earlier than this date, and skirmishers, by whatcver name, had been included in the legions fora long time. Bue tbis is cenainly tbc first time wc hear ofthe equipment later standard for the 'lltliw, and may well reflect a real reform. About this time the minimum property qualification for military service was reduced from 11,000 to 4,000 asses. Tbis would produce many recruits who could not afford hcavy infantry armour, so it is likely tbat steps were taken to ensure tbat the poorer recruits served in the unarmoured velites. The equipment was no doubt reformed at tbe same 첫me. However, ifthe earlier lroes and rorarii wcre bo1h ligh1 troops, tbe number ofskirmishers in a legioa would have remained at abou1 1,200.

One mani pie each of liascac1~ principes and criarii could be combined in a cobort (colrors). These seem to have been first uscd about 210 as a tactical expedient in Spain. Fora century or so maniple and cohort co-existed, used accordiag to tactical circumstances, bue by Marius' day che cohort was the basic tactical and administrativc unit.

Thc different roles of hastati, principes and criarii thus disappeared. Marius is often also given credit by modem writers for abolishiag the veliees, bue Frontinus says Sulla s첫ll had them in 86 BC, so they may not havc vanished till soon af!er this date.

Each legion had 300 cavalry in ten curmae of 30. The t11m1a elected three decuriones, leaders of ten, of whom 1he

39

senior commanded the cumza with the rank of praefecrus. Each decuria chose an oprio as his second-in-cornmand.

This organisation probably applied throughout our period.

Italian allies The beginning of our period saw Rome engaged in a series of wars to restore ber contro! over the Latin League, which had collapsed after the Gallic invasion of386. In 360, the Latin cities Tibur and Veli trae even called in Gauls 10 help against Rome. Roman authority was restored by 354. T he Latin League provided forccs cqual in number, organisation and equipment co tbc Romans, and subject to Roman command. After the Latin rcvolt in 340 most Latins were incorporated as Roman citizens and the League was dissolved, replaced by a series ofbilateral treaties between Rome and the individuai cities. Latin colonies were founded throughout ltaly, however, and "aliics ofthe Latin name'', socii Latini 11omi11is, had greacer autonomy than ocher allies. They seem to have provided the bulk of the allied forces, Livy sometimes speaking as if all the allies were Latins, bue as Rome expanded throughout ltaly more states were obliged 10 provide allied troops - Sabini, Umbri, Volsci, Marrucini, Paeligni, Ferentani, Dauni and others, beside Latins, at Asculum, for instance, while by the mid-2nd century aU ltaly was allied. One special case was the Greek cities ofthe south, who were socii 11avales, generally obliged 10 provide ships rather than soldiers.

Tarantine nobles are found fighting against Hannibal, though, and Rome may have made a specific request for the famous T arantine light cavalry. Another exceptiooal case was the Bruttii, who after tbeir suppon for Hannibal were completely demilitarised, and were not recruited for 1he army.

Polybios describes allied organisation. Each Roman legion was accompanied by a La1in ala, wing, theore1ically with tbc same number ofinfantry as the Romans and three times as many cavalry. In practice there were often more allies 1han Romans - 20,000 allied infanrry to 16,000 Romans at the Trebia, or 15,000 to two lcgions in Spaio in 180. The allied force was cornmanded by officers called praefecti sociorum, apparently 1hree to an ala (Polybios envisages twelve when four legions are raised). An elice force of extraordinarii was selected from 1he best allied 1roops, one·third of the cavalry and one-fifth of 1he infantry. Apart presumably from 1hese, each allied unii was composed of men from one town or nation. Cavalry were in turmae probably the same strength as the Roman.

Although Polybios assumes the infantry camped in 111a11ipuli, the cohors was the standard unit both for recruitmen1 and tactics a1 least as early as the second Punic war, before Roman citizen troops used it. lt may originally ha ve been a Samnite unit, so some allies may have used it before their incorporation into the Roman army. Allied cohortes of 460, 500 and 600 men are recorded. Manipuli probably existed as sub·units within tbc cohors.

Allied armament is not specified - which may simply be because 1he Romans laid down no regulations about it.

Certainly the Latins were equipped in Roman s1yle, but many other ltalian peoples fough1 in their own armies in ligh1er equipment. Livy believed that in 217 Rome traditionally used only citizen and La1in heavy infantry, which may imply that other allies retained their native lighter sryles. Perhaps a graduai adoption ofRoman armament took piace. Plutarch seems to envisage the Paeligni at Pydna equipped as Romans, but they had then been Roman allies for nearly 150 years, and even ifthey had adopted Roman arms that need not mean ali allies had done so. Allied alae seem to bave included some light infantry skirmishers, but perhaps in smaller proponion to the legions; a1 the Trebia there were 6,000 light infantry, ofwhom che four Roman legions should alone have provided over 4,000; but losses al the earlier Ticinus skirmish may explain this.

Beyond Italy Outside ltaly, Roman armies used a wide variety offoreign auxiliaries, usually locai 10 the 1hea1re ofwar. The first non·ltalian allies were Gauls; 800 deserted from Carthage to Rome during che firs1 Punic war. The Gallic 1ribes of what is now nonh ltaly were mostly hostile, but the Cenomani were pro-Roman, and fought in the Telamon campaign and at the Trebia, where they s1ayed loyal when 01her Gallic units deserted. In che same year, Gallic cavalry in 1he service of Rome's ally Massalia fough1 for Rome on the Rhone. In the Telamon campaign, the Cenomani were brigaded with Veneti, a people possibly oflllyrian origin living in the area stili called Venetia who were long standing allies ofRome. U nlike their Gallic neighbours, the Ligurians are rarely found in Roman service, though 2,000 were in the army sent to Greece in 171.

In Spain locai auxiliaries were used on a large scale, s1arting with 20,000 Cehiberian mercenaries hired by the Scipio brothers about 212. They deserted, contributing bo1h 10 a Roman defeat and to the Spanish reputation for treachery. Thereafter most Roman armies in Spaio used native troops, 20,000 at Ilipa for instance and 6,000 against the Celtiberians in 181.

Numidian cavalry were widely used, andare an e.xception to the generai principle 1ha1 foreign troops were recruited locally; they were apparenùy considered valuable enough to be used far from home. A few Spanish and Numidian cavalry deserted 10 Rome in 215, but 1he firsl large scale use was when king Masinissa joined Sci pio in Africa with

40

6,000 foot and 4,000 horse. Thereafter Numidians were used outside Africa quite often; 800 cavalry in Liguria in 193, 500 in the war with Antiochos III, 2,200 cavalry and 1,000 infantry in Greece io 171. Numidia also provided most ofRome's elephants; the first few used against Macedon in 200 had been captured by Rome from Canhage, but they were soon joined by tcn sent by Masinissa, and Numidia sent elephants to fight at Magnesia, Pydna, in the third Punic war and at Nurnantia in 134, when they were escorted by archers and slingers.

Rome was also supplied with troops by several friendly Greek states. Syracuse sent 500 Cretan archers and 500 or 1,000 other infantry, called variously slingers and peltophoroi (thus slingers with the pelte?) who fought at the Trebia. Aitolia and Pergamon co-operated with Rome in the first Macedonian War, and in the second Aitolia seot 6,000 infantry and 400 e.xcellent cavalry to Kynoskephalai (where 500 Cretan archers, 300 Apollonians with the same equipment, and 1,200 Athamanian infaotry also scrved). Aitolia was hostile during the war with Antiochos IIl, but Eumenes II of Pergamon supplied an invaluable force of3,000 peltasts, 500 Cretans, 500 Trallians and 800 cavalry (including some Achaian 1roops already serving with the Pergamene army). In the war with Perseus, Pergamon sent 6,000 infantry (including Cretan, Mysian aod Cyrtii skirmishers) and 1,000 cavalry (including 200 Galatians) to Greece in 171 , though most went home before Pydna. Also serviog in the army in 171 were Aitolian cavalry, 1,200 Achaian archers, 300 cavalry and 100 foot from Apollonia, and 300 Thessalian cavalry. Achaia was a frequent ally; the League's army campaigned with Rome against Sparta in 192, and 100 slingers helped at the siege of Ambracia in 189. Even Philip V of Macedon became a Roman ally after Kynoskephala i, helping against the Seleucids, and allowing the recruitmcnt of2,000 Macedonian and Thracian voluntccrs who fought at Magnesia.

In 229 Romc acquired a protectorate on the Illyrian coast. Troops were raised in this area against Perseus of Macedon and Genthios oflllyria in 168 - 2,000 cavalry aod 2,000 infantry from the Parthini, plus troops from the Bullini and the coastal Greek cities of Apollonia and Dyrrhachium.

Rome 's artillery was also mostly supplied by allies, as she made no attempt 10 build up a large state arsenal. Thus Greek cities in Sicily supplied engines for Scipio in Africa in 204, and Pergamon provided engines in 200.

THEOSCANS The Samnites and their relatives in centrai and south ltaly are k:nown as Oscans, from the name oftheir language.

The name is actually sHghtly mislcading, as ic derives from the Osci or Opici who were the Samoites' predecessors in che centrai highlands. Thus some modem writers prefer che name Sabellian, from the Latin Sabellm which carne into use in the l st century BC to denote Oscan speakers other than the Samnites. Alternatively, some ancient writers call che whole lot Samnites. As Sabellian is striccly anachronistic, and the generai use ofSamnite misleading, I bave sruck to Oscan. Apart from the Samnites, Oscan peoples include smaller centrai ltalian hill-peoples such as the Sidicini aod Frentani, and tribes io the south formed by Oscan migrants conquering and absorbing the natives, and taking some oftbe Greek cities, in the 5tb century - the Campani, Lucani and Brunii. Some other cemral ltalians such as the Vestini and their allies spoke related dialects Professor Salmon refers to as "Sabellic".

Most ofthcse states were organised as leagues oftowns or tribes, with elected officials probably iocluding elected geoerals. The noble families dominated politics, especially in Campania, no doubt providing most ofthe generals, but their power was limited and tbc Oscans were regarded as democratic.

Samnites The Samnites were tbe largest and most importane Oscan nation, a league of four tribes, the Peneri, Caudini, Hirpini and Caraceoi. Polybios' summary ofltalian manpower allots them a potential strength of70,000 foot and 7,000 horse in 225, when much oftheir territory had been annexed by Rome; Livy records 40,000 men in one field army in the 4th century. They fought a series of wars with Rome from 343 to 272, and were by far Rome's most formidable ltalian foe. M any later joined Hannibal. Uoder a generai appointed for one year their army was organised into legiones (as were those ofthe Lucani, Brunii, aod probably most ltalians) which may ha ve been larger than 1he Roman legio, as 1he eli te Lioeo Legion of293 had 16,000 men. They were divided into cohortes of 400, as were the Hernici in 362. Thc references are from Livy, who sometimes misuses Roman military terminology; but the strength of 400 does not conform 10 later Roman cohorts. This suggests it is accurate, and that the cohors was first used by centrai ltalian hill peoples, and was introduced by such allies imo Roman service. Samni1e officers were close enough to Roman tribuni militum and ce11111rio11es for Li vy 10 use tbose words for them. Despi1e obvious similarities to che Roman organisation, there is no trace of the Roman triple-line system, nor of special1s1 light infantry (pcrhaps because the light Samnite equipmem made them unnecessary).

41

Elite infantry bodies existed. Livy's account of magnificent "gold" and "silver" forces in 308 is cenainly exaggeratcd and may belong to a totally fictitious campaign; but in 293 the Samnites raised an elite legio limeaca, 'Linen Legion' (the reason for the name is obscure) from their nobility, bound never to flee by special oaths which were administered by armed (and combatant?) priests.

The proponion of cavalry actually fielded is never mentioned, but may well have been low. Samnium's wooded mountains are not good cavalry country, and in severa! battles with Rome the Samnite cavalry played no pan worth mentioning, perhaps because oftheir small numbers. Like the Romans they were organised imo cum1ae.

Being a fairly poor people, the Samnites rarely if ever hired mercenaries; Livy records a rumour that in 322 they were levying mercenaries from their neighbours, but none are mentioned in the subsequent battles. They fought in alliance, however, with various other ltalian peoples - even with Rome in 340. In the second Samnite-Roman war (326-304) Nola in Campania was an ally, as were the Vestini and their associa ees and probably the Apuli, while some Volsci and Hernici carne over from Rome to Samnium and some Etruscan cities joined the alliance from 31 1. In the third war (298-290) the Vestini were pro-Roman, but some Etruscans were allied with Samnium from the sran, and a grand alliance ofSamnites, Etruscans, Umbri, and Senonian Gauls was formed- and defeated at Sentinum. Livy describes the Apuli as allied with the Samnites, violating their treaty with Rome. Later the Samnites joined Pyrrhos' coalition, bue stayed in arms longer than Pyrrhos did (284-272).

Campania The Oscan Campani took over the Greek and Etruscan cities ofthe Campanian plain in the late 5th cemury. Their league, based on Capua and Cumae, covered only the north ofthe plain. Further south Nola was an independent Oscan city aligncd with Samnium. Neapolis (Naples) stayed Greek, though with some Oscan infiltration; in 327 Nola forced Neapolis imo alliance and installed a garrison of2,000 Nolani and 4,000 Samnites - which was soon ejected under Roman pressure.

The Campani were influenced by the Greek civilisation they took ovcr; to defend their new wealth against Samnite pressure they allied with Rome in 343, sparking offthe first Roman-Samnite war. When the two great powers made peace and agrced spheres of influence, the Campani joined the Lat.ins in revolt against this dornination. The Campanian nobility however stood aloof, and their excellem cavalry seems to have played no pan in the war.

Thereafter the Campani were made Roman citizens (without the suifrage, except for the nobles), and fought for Rome till Capua joined Hannibal. As citizens, they are lumped in with the Romans in Polybios' figures for Italian resources; Livy estimates that in 216 Campania could raise 30,000 foot and 4,000 horse, but he may only be referring to the area ofCapua, as the figure seems too small for the whole country.

The Campanian plain was good horse-breeding country, and Cumae and other western Greek cities had continued the reliance on aristocratic cavalry long obsolete back home. The Oscan nobility took over the tradition, and Campanian cavalry became famous as mercenaries and distinguished themselves in Roman service. Some oftheir infantry, too, seem to have adopced Greek hoplite equipment. After 340, as Roman citizens serving in the legiones, they presumably adopted Roman organisation and arms, evenrually ifnot at once.

Lucani and Brunii The Lucani formed a league which according to Polybios could cali on 30,000 foot and 3,000 horse in 225. The plain round Paestum (formerly Greek Poseidonia) was good cavalry country, producing horsemen very like the Campanians. But most of Lucania was wooded hills, which explains why Polybios' cavalry proportion is comparable to Samnium rather than Campania. Livy says the Lucani allied with Rome in 326, but this is almost cenainly confusion with the Volscian town of Luca. Lucania did cali in Roman aid against Samnium in 299, probably for the first time. Later they joined Pyrrhos, and some joined Hannibal.

The Brunii in the "toe" ofltaly were an indigenous tribe who had been conquered and Oscanised by the Lucani.

In 356 they rebelled and set up their own independent league, which conquered some ofthe coastal Greek cities.

They fought Taras, sometimes allied with che Lucani, and sided successively with Pyrrhos and Hannibal. Little is known ofcheir armies except that they relied on the usual mobile javelin-armed infanrry organised in legiones and coliorces. They raised 15,000 men in 215.

42

Apulia Apulia in rhe south-east ofltaly was inbabited by three tribes known as Iapyges or Iapygii; from nonh to south tbese were the Dauni, the Peucctii, and tbc Mcssapii in the "heel" ofltaly. Tbc Iapyges contained a strong and probably dominant Illyrian element. By tbc mid-4th cenrury the Dauni and Peucetii are found grouped togetber as Apuli, whĂŹch may derive from the name Iapygii. They are sometimes regarded as an Oscan nation and may have been dorninated by Oscan immigrants. The Messapii, pcrhaps the st rongest of the three tribes, kcpt thcir distinctivc language and idcntity, and aligned tbemselves with Samnium while tbe Apuli allied with Rome in fear offunber Oscan pressure from the centrai higblands. The Mcssapii tbus aligned themselves with Pyrrhos, while tbe Apuli stayed loyal to Rorne, and tbe Dauni attacked Pyrrhos' camp at Asculum with 4,000 foot and 400 horse. In 225 Polybios gives the whole country, " Iapygcs and Mcssapii", by now ali allied with Romc, 50,000 foot and 16,000 horsc. Much of Apulia, especially Messapia, was good pasture-land, famous for ics horses, hence che high proponion of cavalry.

T b e "Sabellic League" The Vestini, Marsi, Paeligni and Marrucini, "Sabellic" hill-tribes nonb of Samnium, formed a confederarion Salmon calls the "Sabellic League". The neighbouring Oscan Fremani may bave been associated witb tbem. Livy sometimes calls the wbole league Vestini, suggesting chese were dominant. He says they were as strong as Samnium, probably exaggerating since Polybios only allots tbem 20,000 foot and 4,000 horse in 225. At firsr aligned with the Samnites, they joined Rome in 304-2. Frentanian cavalry are mentioned serving against Pyrrbos, Paeligni and Marrucini infantry cohortes fighting bravely but vainly at Pydna. Thcy were probably armed and organised much like the Oscans.

THE ETRUSCANS Etruria was a league, t raditionally of 12 cities. This was mainly a religious association, with its federai centre at thc tempie ofVoltumna in Volsinii. Politically thc various cities mostly pursued their own courses, but occasionally managed to unite for war under tbe auspices oftbe league; Livy records in 357 "a cali to arms to ali those wbo bore the name Etruscan". By tbc start of our period Etruscan powe r bad already been decisively weakened by Gauls and Romans. Tbc Gauls overran the Erruscan cities in the Po valley about 400, though Felsina (Bologna) held out till 350 or so: Rome had already destroyed Veii in tbc soutb, while Caere was her long-standing ally, absorbed into the Roman state in 353. The remaining Etruscan cities fought a series ofwars with Rome, sometimes in alliance with Samnites or Gauls, till Rome crushcd an alliance ofEtruscans and tbc Gallic Boii in two banles at Lake Vadirnon, in 282 and 28 1. Tbereafter Etruria srayed quiet under Roman rule, supplying allied troops against tbe Gauls in 225 while Etruscan cavalry are mentioned in the war against Hannibal.

Livy says the Etruscans bad more manpower than any orber 4th-century ltalian nation; in 225 Etruscans and Sabines together raised 4,000 horse and over 50,000 foot, but this may not have been the toral available strengrh.

Class divisions in Etruria were sharp, the bulk ofthe rural populace being oppressed serfs. These would not be able to alTord adequate armament, nor would they be trusted by tbc ruling aristocracy. Etruscan cities may thus have been able to mobilise smaller proponions oftheir manpower tban others. The citizenry fougbt as beavy infantry armed originally as Greek boplites. By the 4tb cenrw y some used pila instead, and since tbe long spear did not go out of use Etruscan armies probably contained a mixture of pila and spears likc Roman legions oftbe period. Livy indeed calls tbem by Roman terms, mentioning legiones and ma11ipuli. The legio and cemuria at Rome go back to the period ofEtruscan dominion, so probably Etruscan organisation was very similar to Roman, except that therc is no sign ofany division ofthe heavy infantry into lioes like the Roman hastat1~ pri11cipes and triarii. Ligbt infantry are rarcly mcntioned so may not have been numcrous or imponant. H owever some peasant militia system, which no doubt provided lightly armed infantry, seems ro have cxisred as " hasrily levied cohortes of Elruscan peasanrs" turned out to fight Roman raiders in 310.

43

T ACTICAL METHODS THE PERSIANS A Persian army involving the full levy ofthe empire would be led by Lhe G reat King himself; in baule his 1raditional position was in or behind the cemre ofthe baule li ne, surrounded by his bousebold troops. The last king, Darius III, was able, ingenious, and personally b rave, but nota bold tacĂšcian; he was cautious aod deliberate, :md would cut his losses and run when things went wrong, rather than eitber make a desperate last stand or improvise a new pian.

When thc ki ng did not command in person be was traditionally reluctant to entrust too much power to a single generai, for fea r of revolt. With limited au thority, Pe rsian generals tended to proceed slowly and with caution, " Iodeed, it is the usual custom for Persian comma nders, not being iodependent in the generai conduct ofwar, to refer ali matters to the king and awai t his replies concerning every detail" says Diodoros. For the Granikos campaign command was delegated to a committee of courtiers, locai satraps and Greek mercenary generals.

Rejecting the Greek Memnon's proposed cautious scorcbed-earth policy, which would devastate their own estates, lhese nobles chose 10 risk a baule, ahhough they were outnumbered. T ypical aristocratic rashness, perhaps, but not typical oftbe Persian approach to war, wbich tended towards the cautious and defensive.

T he best and most important part of the army was lhe cavalry, of whom thc armou red javclinman on an unarmoured horse, provided by the Persians themselves amoog others, was commoncst. Asklepiodo1os says the typical formation was a squadron drawn up witb as maoy raoks as files. As organisation was decimai, a 100-man squadron IO ranks deep may have been tbc basic unii, but 10 seems 100 deep to be useful; Polybios for instance says cavalry are no use deeper than eight ranks. A 50-man squadron, drawn up approximately seven by seven, perbaps with the officer outside the ranks, seems more likely. Diodoros says the royal guard at Gaugamela formed one outsize 1,000-man body, and Plutarch notes tbese were io a deep fonnation. Cavalry would normally form up on the wings. At Gaugamela, with few good infantry, cavalry fonned almost ali the main li ne, and a1 the Granikos the front line was exclusively cavalry. Severa! reasons bave been advaoced for the curious decision to hold a riverbank with cavalry, when there were good infamry avaiJable. I favour baste ratber tban planning, but it is worth noting that an ea rl ier satrap of Armenia had tried to bold the river Keotrites against Xenophon with cavalry, deploying his foot behind (though these were not particuJarly good troops). Cavalry would often use numbers and mobility to outOaok the enemy, but fewer now had the bow than in earlier armies, so they were less likely to skirmisb and more Likely to charge home, throwing javelins on the way in and closing with hand-weapons. Units on armoured horses would often lead the charge.

T he best light infant ry were Persian arcbers and slingers, wbo were probably used mixed together, as was later done

by Macedonian generals who had access to Persian troops. T bey would screen dcployment along with the cavaJry, and outOank an enemy or harrass his march in hilly terrai n. At Gaugamela the native Persian contingent mixed cavalry and infan try togetber, perhaps imitating tbe M acedonian use oflight infantry at the Granikos. A Persian army could produce a frightfu l missile barrage, but as lhe tradit ional bow-and-spear heavy infantry units seem to ha ve disappeared, archery was now less important than ea rl ier. N onetheless the desirc to let archers shoot as long as possible at an advancing enemy may have contributed to the preference for s1atic and defens ive tactics.

Although some of the unreliable subject levy contingents were similar light infamry missilemen they were generally drawn up behind tbe main line ofbanJe, wbere ifthey contributed nothing to victory they wouJd at least not provoke defeat by an early lligbt. T beir only effect on the course oflssos or Gaugamela seems to bave been to hinder the Persian retreat.

Apan from the small royaJ guard, Persia's only good heavy infantry were Greek mercenaries, who were the backbonc ofthc army and displaycd remarkable loyalty to their paymaster, 2,000 oftbcm following Darius to his death cven after thrce Pcrsian disasters. T he mercenaries were largely responsible for many victories over rebel satraps and Egyptians, and put up a brave and desperate fight against Alcxandcr. There was only a limited numbe r ofGrceks availablc evento so ricb a power as Persia, and they wcre often supplemented by heavy infantry oflesser enthusiasm or experience, wbo wouJd usuaJly form up on thcir nanks, as the kardakesdid at Issos. Tbe army which invaded Egypt in 351 was divided into four divisions of which three werc of mixed Greeks and "barbarians".

Scythed chariots, wben used, we re spaced out at intervals in front ofthe li ne to disrupt the enemy by an earlycbarge.

T hey were initially used against G reek infantry, witb little success; Darius at Gaugamela may bave meaot to Lry them out against the M acedonian cavalry, as most were in front ofhis wings, but in the cvent they were beaten by light infantry. Even when they succeeded in charging home, the chariots needcd to be closely supported by otber troops to exploit the gaps they made. Such suppon was rarely forthcoming, but the scythed chariots' ooly two

44

successes, against Greek hoplites in 395, and in Pontic hands against a Bithynian army aftcr our era, occurred wben their charge was supported. The elephants at Gaugamela may ha ve been inrended for such an early charge, as tbc Persian battleplan placed them in front ofthe centre; but there is no memion ofthem in the actual fighting, while elephants were later captured in the camp. lt seems likely they did not fight, perhaps being withdrawn at the last minute !est they frightened the Persian cavalry horses, who were not used to them.

The Persians showed a good eye for terrain, exploiting thc defensive values of rivers and mountain passes, using the hills fora flank move at Issos, and taking care to choose :md clear a plain for the chariots at Gaugamela. At Issos, Darius streng1hened with a palisade (presumably hastily improvised) those sections ofthe river bank which were less steep; this the Macedonians saw as proof of his cowardly and slavisb methods of thought. This is a harsh judgment on a sensible move, but one which is rypical ofthe cautious, defensive streak which runs through the Achaemenid approach to warfare.

THE HE LLEN1STIC TRADITION The various Greek, Macedonian and Hellcnistic armies drew their tacrics from a common basis of docrrine and experience, and can best be considered together. Organisational and tactical theory was summarised in severa! manuals ofwhich that by Asklepiodotos, written in the lst century BC, is the oldest surviving. Later manuals by Ailian and Arrian contain much the same information, Arrian adding Imperia! Roman materiai. Tbc manuals are a useful outline, but their relation to practice is sometimes limited.

Despite considerable development from the Pcloponnesian War on, to the Greek city¡states ofthe 4th cenrury war was still largely seasonal, amateur and tactically simple. Tactics were based on a solid phalanx ofhoplites (the term phalanx can refer to any solid formation of infantry, or to a whole battle-line, and is even uscd occasionally for cavalry; howcver phalangite (Greek phala11ggires) is usuaJly rcstricted to Macedonian-style pikemen), witb tbe best troops on the right, the position ofhonour. This would advance and engage the enemy band to band. Reserves, outflanking moves, and such variations were rare. The Boiotians had long favoured deep phaJanxes, and had jus1 before our era developed this into an attack in echelon, led by a deep hoplite mass on 1he left wing. CavaJry and Light troops were subordinate arms in most armies.

A general's role was limited to deploying 1he army and then fighting in 1he front rank; there was little question of him actually directing the baule. The 3rd-century Achaian generaJ Aratos ofSikyon went funher, drawing up the army and then retiring 10 a safe distance 10 await the outcome ofthe bartle. He was ridiculed for tbis, bowever, and throughout this perioda generai was cxpcctcd to fight in pcrson.

Philip II of Macedon developed his tactics largely from the Boiotian model. Their deep formations probably inspired his deep phalanx, and he and his son Alexander used the oblique advancc with one wing refused; at Chaironcia Philip developed this into a feigned withdrawaJ by one wing. Cavalry was more important in the new Macedonian army, used todeliver a decisive chargc at a weak point in the enemy linc. Alexander also made good use oflight infantry, offensively in difficult terrain or in dose conjuncrion with cavalry in both attack and defence. Hc also introduced a second infantry line to dea! with breakthroughs or outflanking moves by a mobile enemy. His hallmarks were co-operation berween thc various arms, careful timing ofthe attack, and a dose persona! contro! over ali the units at least ofthe offensive wing wbich be commanded in person. H owever Alexander stili fought in person, leading the decisive heavy cavalry charge, and most later generals followed his example. Philip and Alexander were also innovators in their emphasis on a vigorous pursuit of a beatcn army, and in waging war ali year round, not stopping for winter or harvest. After the generation of AJexander's Succcssors, however, warfare tended to becomc a scasonal business again.

The Successors, educated in Alexandcr's school, continued his tactics. It was stili common to attack with one strong wing, refusing tbe other. Thus at Paraitakcne both sides attacked with the right and refused the left, while at Gabiene and Gaza the two strong wings mct. Against Krateros in 32 1 Eumenes attacked witb both wings, refusing his weak centre. Victory on che wings could now decide the baule, whatever happened to the two phalanxes. The Successors' cbief innovation was the use of elephants, mainly to open the attack ali aJong the line. Pyrrhos, with fewerelephants to risk, tended to hold both these and bis best cavalry in reservc fora decisive blow. Generals by now had effective contro! over the whole battlc, not just their own wing; Eumenes for instance could summon cavalry from one wing to the other in mid-battle, which not even Alexander seems to have achieved.

In the 270's some Greek states replaced the traditionaJ hoplite with thureophoroi, who were used primarily as skirmishers, though equipped for dose combat too. Later, these states and others took up Macedonian-srylc pike phalanxes, and thus Macedonian tactics. Thureophoroi and othcr light troops, mostly mercenary, stili played a largc

45

part. Cavalry never became as important in Greece as in Macedon.

After the generation of che Successors, thc major Hellenistic armies underwent a decline in tactical flair and flexibility. By che scandards ofche ancient world as a whole, the leve I of generalship was by no means poor; but afler Alexander, perhaps the greatest gene rai in history, and then a generation ofable commanders who had leamcd their trade under him, mere compctcnce was bound to be an antidimax. Cavalry dedined in numbers and importance, at least in Macedon and Ptolemaic Egypt, che Seleucids perhaps being an exception. Instead ofthe decisive cavalry charge there was more reliance on che phalanx, which formed a higher proportion ofthe army and was the main offensive arm. The use of offensive and defensive wings also dedined, as the aim was to get your centre into contact wirh rhe enemy and defeat him in a phalanx fight. Reserves were stili rarely used, and Antiochos Ill's offensive employment of them at Apollonia may be unique. However, Antigonos Doson's carefully planned assauh at Sellasia, Antiochos III's unorthodox cavalry anack ac Magnesia, and Eumenes of Pergamon's decisive use oflighc troops at the same banle, ali show thai latcr Hellenistic generals could stili produce some effective surprises. The use oflighc troops on forced marches and in rough country was also well developed, especially by Philip V, who was noted for his speedy offensives, sometimes surprising the enemy outside the usual campaigning season.

Hoplites Hoplites formed up in dose formacion on a frontage ofabout 3ft/90cm per man, and probably the same dep1h per rank, each man 's right protccted by his neighbour's projecting shield. They most commonly fough1 in muhiples of four ranks deep, usually eigh1, 1hough four was known, 12 and 16 common. Asklepiodocos says early infantry orgaoisation was bascd on a file of eight men, bue there is linlc evidence that hoplites were organised at ali down to this level, except in Sparta; and Spartan organisation was bener adapred co deprhs of six or twelve, twelve being usual. Theban offensive wings could be up to 50 ranks deep.

Normally two hoplice phalanxes would simply head straight for each other, break into a run for the last few yards, and shove and stab till one li ne broke. The Spartans were noted for their slow and ordered advance, marching in step to the rhythm offlute-players; others were more excitable, the Thebans at the Nemeia in 394 breaking into a run when che armies were 200 yards/180m apart, and some Spartan allies and mercenaries start.ing to run at halfthac distance. While Spanans advanced quietly, to hear che flutes and officers' orders, and the allied army at Thermopylai in 279 advanced "silently and in order", other armies would shout warcrics.

As a hoplite li ne advanced it tended to edge to the righr; the righe-band man would drift right in fcar ofbcing out¡ flanked, and rhe rest ofthe li ne would follow suit, each hoplite try.iog to keep under the procection ofhis right-hand neighbour's shicld. Thus each righr wiog might overlap and beat rhe opposing left. Planned flank attacks were a different matter, and less common, though che deep Theban columns were particularly vulnerablc to them.

Against light croops the aim was to dose as fast as possible co minimise missile casuahies. Either the whole linc would charge at a run, or, cspccially in a Spartan army, the younger hoplitcs would rush out ofthe ranks to catch the flect-footed enemy, the rest ofthe line following at a steadier pace. Similar tactics might be used to support cavalry outnumbercd by encmy horse, but ifunsupported by cavalry ofrhcir own hoplitcs were reluctant to advance in the foce of enemy cavalry because of their vulnerability if outflanked or dispersed. lf surrounded or vulnerable to surprise attacks, hoplitcs would forma hollow square, with baggage, camp-followers and light troops in the centre.

Tbe Macedonian phalanx Asklepiodotos specifies three densities offormation for the phalanx; open order, which "seems to be thc natural one and has, thereforc, no special name", allowing both frontage and depth of four cubits (6 feet/l.8m) per man; pyknosis, with frontage and depth of two cubits, which Polybios confirms was che usual baule formatioo; and synaspismos, "locked shields", al so sometimes calledpyknosis, allowing only one cubit frontage and depth per man.

Diodoros gives Philip Il crcdi1 for inventing Lhis very dose formation, which is probably 1rue since it could not ha ve been used by earlier hoplites with two-cubit shields. lt was only used 10 resist an anack, being 100 dosely packed 10 allow much movement. Diodoros says it was used by Alexander's phalanx to deter chariots at Gaugamela, while Polybios describes Philip V's Macedonian pellasts crossing a river and forming up in synaspismos on the oLher bank ro keep off A.itolian cavalry. At Sellasia Lbe Macedonian left, ini1ially repulsed, rcformed into an epallellos, "interlocked'', phalanx, perhaps aoother oame for tbe same very dense formation.

Under Philip 11 the usual depth may bave been ten ranks, as a file was called dekas, ten, but Lhereafter it was 16.

Thus che speira and its equivalents, the basic 256-man unit, formed up 16 men wide by 16 deep. However at Issos the phalanx attacked in half-files, eight deep, because the impetus which was thc ma in advaotage of tbc deeper formation would have been broken by the river obstade. The Seleucid phalanx at Magnesia fought 32 deep, wbich 46

was regarded with hindsight as roo deep for the rear ranks to be ofany use. At Kynoskephalai Philip V doubled the depth ofhis phalanx on the right wing, so they too may have attacked 32 deep. But he initially deployed only half the phalanx, and then concracted it to make room for the other balf, so he may well have staned eight deep to cover the ground, then deepened to tbe normai 16 to attack.

In the 4tb century the phalanx was cbiefly used to pin tbe enemy centre while heavy cavalry delivered the decisive blow; Alexander seems to have regarded it as relatively unimportant, so much so that be only took two taxeis in bis main force at tbe Hydaspes. Latera pbalanx attack became the main means to victory. A cbarging pbalanx was universally agreed to be a terrifying sight, and Polybios thougbt tbat on leve! ground it would beat even the Roman legions in a straight frontal figbt. Pyrrhos' crack Macedonians proved at Asculum that this could be done, but at Heraclea the honours were much more even. Usually the phalanx would form one solid body, presenting an unbroken hedge ofpikes. Alternatively it might advance in ecbelon as at Gaugamela, wbile Antigonos Doson split bis phalanx imo rwo bodies at Sellasia and Antiocbos II1 deployed cavalry between che pbalanx and che similarly armed argyraspides at Magnesia. An outflanked phalanx could forma hollow square to withdraw from tbe field, as at Gabiene and Magnesia. Asklepiodotos also mentions columns and convex or concave lines. The pbalanx was very vulnerable to rough ground which would disrupt its solid ranks. One solution was to intersperse its units with lighter infantry who would be less affected and could more easily adjust their forma첫on to cope with gaps.

Alexander seems to bave done this at Issos when he attached rnercenaries, probably peltasts, to each Macedonian unit; Pyrrhos alternated the speirai ofhis phalanx with ltalians, Antigonos Doson with Illyrians. If Antiochos III was trying to produce a similar result by putting elephants, perbaps escorted by ligbt infantry, berween his phalanx units at Magnesia, then be failed disastrously!

While drill was available to face the phalanx about, so that the file-leaders stili formed the new front rank, there was no similar technique to faceto a flank, except by wbeeling tbe wbole formation; in addition, any turn was obviously difficult when pikes were Iowered and the phalanx was engaged. lt was tbis weakness wbicb the Romans, with tbeir more flexible units, exploited. Nonetbeless, the uomanoeuvrability oftbe later pbalanx sbould not be exaggerated.

Note that at Kynoskepbalai and Corintb tbe Romans won by attackiog tbe pbalanx from an uoexpected direction while ir was already engaged fronrally, a difficult situation for any type oftroops, wbile at Magnesia and Pydna other causes of disorder applied. Philopoimen at Man첫neia, wheeling one unit of his phalanx on to the Spartan flank, proved that be/ore combat had been joined, the separate units even ofthe later phalanx could manoeuvre independently and effec첫vely. That they rarely did so in practice was due as much to the rarity of flexible generalsbip and mobile tactics as much as to tbe inherent clumsiness ofthe formation .

Philip's and Alexander's hypaspists generally formed on the right ofthe phalanx; this has led to the idea that they were a fast-moving " h inge" between the phalanx and the heavy cavalry, and thus that they must have been lighter armed than the phalanx. However there is no evidence for the rapid battlefield movement that such a role would imply, and the position is equally compatible with their being an elite unit armed the same as the phalanx, deployed on the pres첫gious right wing. Cenainly as argyraspides tbey fougbt in the centre of Eumenes' pbalanx, and equivalent guard uoits in some later armies at least were pikemen. Tbe feigned retreat at Chaironeia, however, may argue for light and mobile equipment. Both hypaspists and phalanx raxeis are found on Alexander's forced marches.

Peltasts and thureophoroi Peltasts were traditionally skirmishers, evading when charged, and wearing their enemy down with a rain of javelins. Thus they won severa I notable victories, the most striking being Iphikrates' destruction of a Spanan mora in 390. They were better equipped for close combat than psiloi, so were used in ambushes orto dr ive offskirmishers.

Peltasts could be used along with hoplites to support cavalry against superi or enemy borse, but ifcaught alone were likely to be ridden down. Iphikrates is credited with reforming tbeir equipment to make it better for close combat, but it is not clear if the new weaponry was used widely, or at ali (see note to figure 20). The rltureophoroi who succeeded peltasts io tbc 3rd cenrury were stili rated among the light troops, euzono1~ and were similarly used as skirmishers. Plutarch says that when the Achaians were armed in this way they did not fight in formed units, skirmished from a distance but were ineffective at close quarters, and that their tactics were peltastikes, peltast style.

Tbey were used to support skirmishers, and along with tbese were used on forced marches, screening heavier troops in difficult country, orto open a baule. Thureopltoroi and similar troops, such as T hracians and Illyrians, might also be dep loyed in baule to protect the vulnerable flanks ofa phalanx (though not infrequently even good commanders like Pyrrhos and Pbilopoimen seem to bave covered their phalanx flanks only with cavalry). Though often found fighting hand to hand with their own kind, they could not stand up to a pike-phalanx. Asklepiodotos bases their organisation on files of eight, and this was probably their typical depth in batti e. lt seems likely they would skirmish in open order but dose up to pyknosis for close combat.

47

Light infantry Tbe psiloi or gym11etes were javelinmen, archers and slingers. In Greece and Macedon javelinmen were mucb the most imponant, bui the Seleucids and other eastern armies bad access 10 many locai archers. They usually fought in open order and probably eight deep, Asklepiodotos basing their organisation on files of eight, though some may ha ve had very !inie formai organisation at ali. H owever their formations were obviously less rigid thai those of otber troops, and would be adapted 10 locai fearures ofterrain or the need to screen a given fromage ofphalanx.

They were chiefly skirmishers, deploying in from ofthe phalanx to open the combat and then fall back, allowing the heavy troops 10 close. Asklepiodotos suggests they could draw up behind an eight-deep phalanx 10 shoot overhead, as Xenophon advocated earlier in the Kyropaidia. Onasander sets out 1he rivai view thatpsi/oi should deploy in front so they can see the target, javelinmen can take a run up 10 increase tbeir range, and arcbers shoot ai a lower trajectory with more force. This was the course usually followed, except in the special circumstances of shooting down from, or up a1, a hill; thus Doson's Cretans al Sellasia were behind the troops attacking mount Euas. Occasionally psiloi would instead deploy on the wings when there was suitable high ground for rhem 10 shoo1 from. Generally the skirmishing was a mere preliminary to the clash of phalanxes, but the poorer Greek hill peoples relied heavily on skirmishing, and the Galatians in 279 were beaten almost solely by Aitolian and Phokian javelinmen. Where possible, skirmishers would outflank the enemy 10 shoot from directions where their vict ims could nor use their shields.

Psiloi were often used 10 suppon cavalry, chiefly by shooting al enemy cavalry; the Theban cavalry victory over the Athenians al Mantineia in 362 was anributed to the numbers and qualiry of the supporting light infantry.

Ale.xander used Agrianian javelinmen joining in cavalry melees at the Granikos and the Jaxanes, while al Gaugamela Grecks (probably peltasts) were apparently concealed behind horsemen, charging out in their suppon (as Xe nophon had ea rlier recommended). This co-operation is rarer after Alexander because the psiloi have a new rote guar ding elephants, bui al the Kallinikos skirmisb in 171 Perseus mixed Cretans and others with his cavalry.

Elephants were closely supponed by light infantry from tbeir first appearance in Hellenistic armies, usually abour 50 men per beast. They servcd to protect the elephants' vulnerable flanks and 10 give missile suppon.

In mountains psiloi were used to defend passes, often fortilìed, orto attack similar positions. Such positions were hard to take frontally but could usually be outflanked (or bypasscd; Kassandros, lìnding che Aitolians holding the coastal pass ofThermopylai against him, built rafcs and sai led his army pas1). Alexander relied beavily on archers and Agrianes in these condiùons, sometimes using arcbery to Iure barbarians out of strong wooded positions, in the hope of driving ofTthe bowmen; he would then hit their flanks witb heavy infantry or cavalry in the open. He also favoured night attacks on their camps.

Other roles for psiloi were scouting, protecting heavy troops on the march, keeping ofT skirmishing cavalry, shoo1ing down chariots and elephants, guarding flanks, forced marches, outflanking moves, feigned retreats, and ambushes. Antiochos III at Raphia pur light As iatic infantry in his ma in line ofbanle, with a marked Jack ofsuccess.

Cavalry Polybios, calculating frontages for the battle oflssos, allows each cavalryman a frontage of3 feet/90 cm within tbc unir. Halfthe frontage of a cavalry force, bowever, was taken up by gaps between the i/ai, of equa! length to the from ofan ile, so that over a longer distance there would beone cavalryman 10 eacb 6 feet offrontage. The gaps were 10 allow the units to turn or wheel to the flanks. We sometimes hear of 4th-century Greck cavalry drawn up "Likc a pbalanx of hopli1cs11 which sugges1s a solid li ne witbout the usual intervals. Tbe late Scleucid cataphracts 100 probably used a solid line; they were too unwie ldy for much more than a direc1 charge straight abead anyway, so there was !inie point in making provision for 1hem to manoeuv re; the gaps would only expose 1heir vulnerable flanks.

Asklepiodotos records the basic Grcck cavalry unit to be an ile of 128, with 8 ranks of 16 men, wbicb, because a horse is longer than i1 is wide, would appear 10 be a square. Polybios regards eight as the grcarest useful depth for cavalry, and Greek cavalry are often recorded four or six deep. Instead of the usual square, Tbessalian cavalry uscd rhomboid formaùon with the squadron commander, ilarcl1es, riding at the point towards the enemy and subordinate officers at other angles, an ouragos or squadron-closer at the rear and plagiophylakes, flank guards, at the sides.

Asklepiodotos' example is for an i/eof61 men. The rhomboid manoeuvred more casily than tbe square, as it could wheel by keeping station on the appropriate officer, or a simple turo would leave the ile facing in the right direction, stili in rhomboid formation, with a difTerent poin1 leading. lt was also bener for breaking an enemy line as the best horsemen could be concentrated at the front point, and the narrow formation was ideai for exploiting a gap.

M acedonian cavalry used a wedge, equivalent to the front half of 1he rhomboid, which Philip II is said to have 48

adopted from the Thracians and Scythians. lt bad the same shock value as the rhomboid, but dispensed with che waste ofmen in the rear balfofcbe formation, at che cose ofslightly reduced manoeuvrability.

Greek cavalry were javelin skirmisbers, used co scout, harrass che enemy, drive off enemy cavalry and ride down ligbt infantry. Wben oumumbered by hostile cavalry, as when fighting Persians, they would charge to pin che enemy and hope to be reinforced by their own infantry before superior Perisan numbers told. They could safely charge down peltasts on open ground, and scored several successes against them, but could rarely hope 10 damage boplites; at Kynoskepbalai in 364 Pelopidas' cavalry were bloodily repulsed by Alexander of Pherai's infantry. But they later cbarged Alexander's boplites while chese were engaged frontally, and routed them. Similarly, Timoleon at tbe Krimisos tried to bit the Canhaginians simultaneously in front with his infantry and in nank with the cavalry.

Tbis combined assault became che basic tactic ofthe new Macedonian cavalry; as early as 358 Philip used cavalry to anack IUyrians in nank while his infantry engaged them fron tally. For Alexander the charge ofthe Companions was decisive but was only launched after careful preparation with other cavalry and light infantry, which at Issos safeguarded bis nank and at Gaugamela also stretcbed the Persian line dangerously thin, creating a weak point at wbich the Companionscharged. His Thessalians were the mainstay ofthe defensive left wing; it has been suggested they held offlarger numbers of Persian cavalry by evading their charges, but at Issos at least tbey countercbarged.

For scouting Alexander bad many ligbt cavalry, including Asiatic troops such as horse archers, who were later found in Seleucid and otber eascern H ellenistic armies, riding round the nank ofa less mobile enemy and sbowering them with arrows - as at the Hydaspes where they shot down Indian chariots, Paraitakene where Lhey harrassed Eumenes' elepbants, Ipsos and Azotos.

During tbc Successors' wars in Asia the victory ofone side's cavalry could decide the battle; tbus wben Eumenes' cavalry beat Krateros' che laner's superior infantry surrendered rather than risk being outflanked, wbile Antigonos saved the day at Paraicakene with one cavalry wing despite the defcat ofhis foot. Most ofthe bcavy cavalry would be on one wing for the offensive, the other being mostly light horse to occupy the enemy bue refuse close combat.

Cavalry are al so found in reserve; Eumenes had a picked force of 300 behind his right at Paraitakene, Pyrrhos his agema in reserve at Asculum. The main cavalry wings were often supported by small detached squadrons on their flanks or rear, as a protection agamst encirclement, or else thrown out in front, probably to break up the enemy formation before the main anack. The flank guards could be thc ncw Tarantine javelin skirmishers. Tbcir other tactical innovation was, according to Arrian, that they rode in a cirdc to shoot. This is presumably the formation that che Romans later called "Cantabrian cirde"; the cavalry would gallop round in a cirde witb tbeir sbields on the outside, sbowering che enemy with javelins while presenting a fast-moving, sbielded target to any rerum fire. lt would allow greater concentration offire on one point than the earlier rectangular squadron.

In the 3rd century cavalry tactics dedined again, limited chiefly to scouting, skirmishing and defeating the enemy cavalry. The Antigonids reverted to skirmishing even with hostile cavalry, and wc re bighly surprised to be cbarged

by Roman cavalry; Livy implies the whole idea ofshockcavalry was newto them, so we are not dealing merely with specialist skirmishers who formed only pan of the Antigonid cavalry force. An exception 10 the dedine is the Seleucids, who had more and better cavalry to use. Antiochos III had few at Raphia and only used them against enemy horse, but at Magnesia he deployed some ofhis cataphracts berween rwo forces ofinfantry and used them to ride down Roman foot; at Beith Zacharia in 162 Seleucid cavalry were again deployed in dose support ofinfantry.

The Bactrian Greeks too relied heavily on cavalry, but we bave no account oftbeir tactics, though they must bave been influenced both by Macedonian shock tactics and by steppe borse-arcbers.

Elephants Elephants are not naturally warlike, and panie easily. I Maccabees records the Seleucids giving their elephants wine before batti e, in the hope of overcoming this. Panie was particularly likely when the driver was killed or disabled, as the bond between man and beast was dose, or when the leading elephant was hurt; an army's elepbancs seem to ha ve established a hierarchy much like a wild herd. These considerat ions meanc tbeir use in battle was always risky.

Their usual deployment was strung out in a line in front of the army, the gaps between them filled witb light infantry. The Seleucids ai Pani on in 200 used Tarantine cavalry coo. Ifenough elephants were available they would extend before the whole army. Ifnot, they would be in front ofcbe wings as cbeir greacest effect was against enemy cavalry; horses, especially ifunused to elephants, were cerrified by che sigbt ofthem and tbeir trumpeting. Thus Demetrios' cavalry at lpsos would not approacb Seleucus' elepbants, and the "Elepbant viccory" was won by their effect on horses. Perseus of Macedon, wbose Roman opponents bad elepbants wbile be bad not, allegedly trained his horses to ignore their sight, sound and smell by concealing a trumpeter inside a dummy elephant smeared with foul-smclling paste. Thc bcasts' psychological effect on men who were unused to them was also considerable, but

49

stcady infaotry, cspecially pikcmen, could sometimcs beat them off, as Alexandcr provcd at tbc Hydaspes.

Aotigooos io Pisidia and Pyrrhos at Benevcntum, however, did use them successfully to break tbc line. As ao alternative to the advanced screen, Pyrrhos used his few elephants more cautiously, employiog tbem ooly io a late stage ofHcraclea aod deploying them in reserve at Asculum. Antiochos III at Magnesia used 22 ofhis 54 elephants in pairs between the units ofhis phalanx, to break up che Roman formation, and had che rese in reservc behind tbe wings. At Beith Zacharia che co-operation of elephants and infantry was taken further, each beast accompanied by 1,000 iofantry (iocluding argyraspides armed in Roman style) and 500 horse.

Elepbants were also used in sieges, to batter down gates or uproot palisades. At Megalopolis io 318, Polyperchon's elephants chargcd in dose formation through a breach in the walls. But they were too big and clumsy for use in cities; Pyrrhos' night attack on Argos was held up because the bcasts' towers had to be removed to get them througb the gates, and tbe confusion which led to his death was partly caused by one wounded elephant getting stuck in a gatcway, and by anotber which ran amok through the streets.

Deployiog elephants on tbc wings meant they were often opposed to each other. Wheo cwo clephaots fought, thcy would ioterlock their tusks and push till one forced che other's head aside and gored him (or her; both cow and bull clephants werc used). Mcanwhile thc crews would fight with sarissai and javelins. Otberwise the maio measure agaiost clephants was to sboot at them. They wc re rarely killed by archery, but a barrage of mĂŹssilcs would wouod thc clcpbants and disablc many drivcrs. This could cause thc clephaots to panie aod flee, ofteo stampeding through their own troops. As well as light infantry, Antigonos used horse-archers for this at Paraitakene and the Megalopolitans uscd bolt-shooting catapults against Polyperchon. But in tbis case the artillery could be set up in advance as the elephants' line ofattack was known, through a brcach in the wall; io tbe field artillery would rarely be io tbc righe piace at che righe time. Spikcd obstacles like those oftbc Megalopolitaos or Ptolcmy at Gaza were really auxiliary to tbc missiles, as their main effec1 was lO halt the elepbants' advance, giviog a safely distane aod stationary target. Thc Mcgalopolitan dcviccs werc spikcs hiddco io shalJow trenches; at Gaza the obstacles were portable, some kind of network of iron spikes linked by cbains (Diodoros calJs it charax, palisade). Perseus of Macedon is said to havc raised an anti-clephant corps whose helmets and shields were covered witb nails; tbey werc wholly ineffective.

Aootber trick was to release squealing pigs, since the sound apparently terrified elepbants. The Romans did this against Pyrrhos, and the Megarans later against Antigonos Gonatas, witb tbe added refioemeo1 tbat the poor pigs were covered in tar and set alight. This no doubt ensured they squealed adequately, and tbe flamcs wcrc an added fright for the elephants. After this Gonatas stabled pigs with the elcphants, to gcc them used to thc squeal.

C hariots Kyrenean chariots skirmished with javelins, in what Xenophon regarded as tbc ancieot Trojan manner. 01herwise the only Hellenistic chariots werc che scythed vebicles uscd by Seleucids aod Pootos, used in che Persian style.

Against chariots Hellenistic armies relied mostly on shooting them down with ligbt infantry or cavalry. Chariots wbich survived this barrage might panie and turn back against the.ir own troops, which was disastrous at Magnesia.

Ifthey carried on they would usually veer away from solid infantry formatioos, and lanes would be lcft between che iofantry units down which che chariots would be carried, probably against the will ofthcir drivers, to be dealt with by reservc.s when their momenrum ran out.

Artillery Apan from its primary use in sieges, artillery was often used io che defcnce of mountain passes, often mounted on ficld fortifications. Philip V at thc Aous valley, Antiochos III at Thermopylai, and Perseus in the Elpeus valley ali used it against Romans. lo tbc field it was usually onJy practicable to use artillery where it could be set up in advancc, as at Onomarchos' ambush of Philip orto cover river crossings. Machanidas' use ofbolt-throwing cngincs at Mantincia seems to bave been che only instance of an army marching on to the field and setting up catapults in sight ofthe cncmy; and these engines seem to bave been koocked out by the first Achaian charge, as we hear nothing ofthem in the baule.

THE MACCABEES The guerilla armies of Judas' early campaigns and Jooathan's later stay in che wilderness relied on raids and ambushcs, attacking Seleucid forces u they passed tbrougb defiles in the hills or assaulting their camps at night.

Most early attempts to fighe a pitched batĂše io tbe open led to defeat - mos1 decisively at Elasa whcre Judas was killed. Despitc che account oflarge-scale desertioos before Elasa, morale seems gencrally lO have been high, tbc men

50

"ready either to live orco die valianely". Religious enehusiasm was importane, the scripcures being read before baule, che assault being signalled by blowing sacred crumpecs, and some priescs fighcing in che ranks. Religion led co one early disascer when a rebe l group which refused co fighc on che Sabbach was attacked and massacred wichout resistance, but thereafter the Sabbath laws were relaxed to allow fighting on the holy day ..

As time wenc on Maccabean armies became capable of standing up to the Seleucids in the open, and normai Hellenistic cactics as well as armamene were apparently used. For instance the main body seems to have fought the enemy phalanx at dose quaners, while the generai may have commaoded one cavalry wing. Simon's oucnumbered cavalry in 134, however, were deployed for safety between two infantry forces. This is unusual and especially inceresting because it seems to reflect Seleucid practice and is itself reflecced in the Rute o/ War, which has light cavalry deployed with the infantry and heavier cavalry on the wings. This may be inspired by the Seleucid formation on the march to Beith Zacharia against Judas in 162, when each body of 1,000 fooc was supported by 500 horse with more cavalry, apparently cataphracts, on the wings.

The Rute describes skirmishing slingers, who after shooting fall back on the wings ofthe main formation. l nfantry then advance, deliver seven volleys ofjavelins, and then fall back. T hese seem not to be separate skirmishers but the rhureophoroi who are the army's main strength. After shooting for some time they dose with the enemy, presumably in the hope that these have been weakened by the missile bombardment; this seems to reflect Hellenistic rhureophoroi tactics. The infantry are said to be formed seven ranks deep, but boch this and the seven javelin volleys probably reflect che magical-religious significance ofthe number rather than real practice. The Rute specifies that the army should be able to form square, concave line with che wings advanced, or convex Line with che wings refused, in a passage reminiscenc of Asklepiodotos. T hese formations may be preceded by "towers", apparently three sides of a hollow square with I 00 infantry on each side; since the men in the cowers are allotted larger shields chan usual, and the shields appear co be imporcant, this may reflect che Roman resrudo rather than historical Maccabean practice. The Rute also stresses the role of priests blowing sacred trumpets, as in real Maccabean armies.

THE THRACIANS Thracians were regarded as warlike, ferocious, and savagely bloodthirsty, bue ill-disciplined and sometimes unreliable, with a weakness for loot. Macedonian generals often used them for executions or massacres. Livy describes Kotys' elite cavalry and their supporting infantry charging "like wild beasts who had long been kept caged", and singing as they retumed from battle with severed heads as trophies. They often seem to have sung war¡ songs, Tacicus later saying they "capered and chanced in front oftheir lines according to their nationaJ custom".

Thc Thracians had been thc originators ofthe pcltast style offighcing which che Greeks had adopted, and cheir infantry stili relied heavily on javelin fire. Thracian auxiliaries in Macedonian and Roma n armies were often used for raids and skirmishes. In pitched battles, however, Thracians who attempted to skirmish would now often find themselves ridden down by enemy caval ry, and against Romans and Macedonians Thracia:ns often found that after an initially successful exchange ofmissiJes they would be forced to fight hand to hand and soundly beaten by their heavier armed opponents. On their own heavily wooded and hilly home ground they relied on ambushes wich some success, seizing commanding hilltop positions and delivering surprise attacks from the cover of forests. One Thracian army awaiced Alexander che Great atop a hill behind a barricade ofwagons, which they then rolled down on the Macedonian infantry.

The cavalry who charged so ferociously were probably armoured nobles, and the light equipment ofthe bulk ofthe cavalry suggests they were primarily ski rmishers. T heir traditional formation howeve r was the wedge, which as the Macedonians found was well suited to a charge. Kotys' cavalry was supporced by light infantry who stood between the ranks, charged in suppon ofthem and hamstrung the enemy horses. Bithynian cavalry too seem to have been dosely supporced by attached infan try. This is not recorded before the early 3rd cencury (ifthat is the date ofche Bichynian example) and is probably a result ofHellenistic influence, as Greek and Macedonian generals were using light infanery in dose support oftheir cavalry long before.

THE ILLYRIANS The lllyrians were "a nation offighters both on land and sea", warlike but rather unprediccable. They sometimes charged rashly into siruations they could not really handJe, as at the Roman siege of Skodra in 168 when the defenders unwisely sallied out oftheir strongly fortified position "and on leve! ground joined battle wich greater spirit than they were able to maintain", being soundly beacen. At che other extreme lllyrians could panie easily, as when Alexander's troops in Illyria, bemusing the enemy with elaborate dr ili manoeuvres, suddenly switched to the

51

anack, and che Illyrians lled tbeir bili posicion without attempting to fìgbt. In addition they kept no proper securiry when they camped, with no sentries, no ditch or palìsade, and could easily be surprised. All this corresponds with che stereotyped Graeco-Roman picrure ofthe "barbarian"; but it may not be che whole story. The Oard:moi in 200 are described as much steadier warriors - "these troops do not leave tbeir ranks impulsively but keep dose order both in combat and withdrawal".

This may be connected with a shift in Illyrian tactics from rather disorderly skirmishing towards dose combat. ln 1he 51h cenrury Tbucydides records that "as 1hey fìght in no son of order, they have no sensc ofshame about giving up a position undcr pressure. To run forwards and to run backwards are equally honourab lc in their eyes ... every man is fìghting on his own ... ". They would noi press home an attack against enemy who stood fìrm, and would retire ai speed from vigorous charges, operaring in fac1 like the typical peltas1 and only dosing with encmy weakened by missiles or by their fìerce warcries. Alexander's opponents in 335 dcarly siill had no shame about abandoning positions in che face of a deiermined enemy, while che speed wbich characterised 1he early infantry was stili in evidence in 200, when ii was expected "1ha11he spced oftbe lllyrians would be effective in quick dashes and sudden charges", but ai least these Illyrians (Tralleis in Macedonian service) are expected to charge. Similarly 5,000 Illyrians in 233 had no compunction abou1 charging a larger Aitolian army (lhe League's whole levy), brushing aside their light infantry holding hilltops and following up with a fasi downhill charge which swept away 1he Ai1olian cavalry and heavier troops. The good order :md the willingness 10 hold fìrm and fìgh1 hand to hand, displayed by the Dardanoi in 200, is seen as early as 358 when Bardylis' men formed square when OU!llanked by a Macedonian army with more cavalry, and though eventually beaten put up a prolonged resistance. Tbe contrast between 5th-century Illyrians fìgh1ing individually in no order and their 3rd·century counterpans in fonned speirai suggcscs a marked improvemeor in bartlefìeld organisation probably related to che new tactics. The Macedonian example probably bada lot 10 do with these changes, and improved organisation and more resolute t:ictics no doubt contributed to the successes of3rd century Illyrian raiders.

The Illyrians' shonage of cavalry and light infantry meanr they could be vulnerable to harrassing and outflanking attacks by such troops. They made a habi1 ofseizing commanding bills, often wooded, which gave some security against 1hese cac1ics as weU as enablìng che Illyrians in tum to threaten tbe llanks or the lì ne of march of an enemy less at home in thls sort ofterrain. The forested mountains oftheir own territory made them difficult to conquer, especially as tbeir biJ11ops were dotted with fortifìed towns; they "were given confìdence by che nature of 1heir country and by their fortifìcations."

THE STEPPE PEOPLES

The Scythians and Saka used the tradi1ional skirmishing cacùcs ofhorse archers; oucllanking an enemy, evading when cbarged and shooùng bebind them as they galloped away, and feigning llight to Iure che enemy into a trap.

Thus the Saka surrounded and shot down detached forces of Alexandcr's troops, and at his crossing of the Jaxarces, tried similarly to surround a small force, only to be outflanked themselves. Thc tradiùonal Scythian cavalry formation was a wedge, probably like che squadron wedges ofthe Macedonians uscd primarily for shock action, rather than the more typical skirmishing 1actics. A mucb Later Byzancine taccic called "che Scythian" may rellec1 traditional Scythlan skirmishlng tactics. A unii splits inco cwo sub-units, which ride in opposile directions across the enemy front, showering them with arrows. Some Scytbian cavalry, no doubt the armoured nobles, were dearly prepared for an immediate charge as the baule ofthe Thates shows, and it was probably these wbo made most use ofthe wedge. At che Thates thcy wcrc deployed in che cenrre, which was the traditional post ofthe Sqrthian king, under thc persona! command of the commander in chief. Sarrnatian cavalry of course placed most reliance on a similar immediate charge to contact wi1h che lance.

The rote ofinfantry on the baulefìeld is nor dear, tbough no doubt very much subordinate to that ofthe cavalry.

Otherwise Sarmatian infantry at )cast are seen sniping and harassing enemy columns in forests and defending the hillforts ofthe semi-senled tribes. Thcy might also have been used to defend wagon laagers ofthe nomad tribes, as in tacer steppe warfare.

The Parthians used these same basic cechniques, softening the enemy up with horse archers before charging with their heavier cavalry; the heavy equipmem oftheir cacaphraccs made che charge more formidable than the usual Scythian charge. However, their tactics were affected by their gr~ter proportion of infantry; thus against Antiochos VTI, who had a superiority in cavalry, tbey lured the Seleucid cavalry imo rough terrain by a typical feigned flight, and then ambushcd them apparently with infanti]'.

52

INDIA

lndian warfare was a rathcr styliscd business, conductcd at !cast in thcory according to a chivalrous code ofrulcs, and rarely aiming at the destruction ofthe enemy. The Arrhasas1ra recognises three types of conquest. " Righteous conquest" is undertaken to extract tribute and homage from a weaker neighbour. The defeated king keeps his tbrone as a subject ally, the conqueror ultimately seeking recognition as chakravar1i11, emperor or world-ruler. The acceptance ofsubmission to a powerful conqueror's overlordship explains why so many lndian states were prepared to submit to Alexander. The second form of conquest, "conquest for greed", sought booty and the annexation of part ofthe loser's territory. Third was "demonic conquest", aiming at the politica I extinction ofthe enemy state and its incorporation into the victor's. This was generally disapproved ofin lndian thought, but was advocated by the Ar1hasas1ra. Tbe kings of Magadha bad long pursued demonic conquest, absorbing severa! neighbouring states.

The Mauryas built their empire largely by demonic conquest, though some ofthe outlying areas may bave been merely tributary.

Given the prevalence of"righteous conquest", where only prestige and tribute were at issue, warfare was often a sport ofkings and professional soldiers, irrelevant to the common people. These were left untouched by it, farmers ploughing their fields in the sight of contending armies. Warfare was conducted according to a restrained, chivalrous code which disapproved ofruses or surprise anacks - "Open warfare, however, in which the piace and time for fighting are indicated, is most righteous" - tbough th.is was not always adbered to. Poi son ed weapons were abhorred, non-combatams, wounded or surrendered enemy were nor to be harmed. In theory chariot·drivers and their horses were inviolate, too, tbough the use of armed chariot drivers suggests this was Unle observed. Also in theory, a warrior would only fighe foes of equa! status - chariot·riders against chariot-riders, infantry against infantry, and so forth. In fact this rather unrealistic limitation is not observed even by t he heroes ofepic poetry, let alone in reality.

The army would assemble for war in a large encampment, ska11dhavara, which may be where pan at least ofthe standing forces are garrisoned in peacetime. It is elaborately fortified with ditch, rampart, palisade aod towers, with an inner compound for the king and the headquarters buildings, protected by the palace guards. The army would march slowly ro seek the enemy, slowly because of its vast train of non·combatants and because it would usually carry supplies wich it rather than foraging. The line ofretreat would be secured by occupying key forts. A camp, presumably less elaborate chan the ska11dhavara, would be set up at nights. Before the battle with Alexander the Kathaioi, who had assembled a large army from neighbouring tribes, encamped in a wagon-laager. Before battle, priests and astrologers would encourage the troops. After battle, in keeping with the Umited nature of most warfare, a vigorous pursuit was positively discouraged, " The vehemence ofone rerurning again to the fight and despairing ofhis !ife becomes irresistible; therefore, one should not harass a broken enemy".

In bactle, Poros' deployment at the Hydaspes is probably typical. His elephants were spaced out in front of his centre, with infantry behind them, opposite the gaps between tbe beasts, and also extending the line on either flank.

Beyond these on each wing were cavalry behind a screen of chariocs. The Ar1hasas1ra cices severa! otber possible formations, some of which ignore the infantry entirely, an indication of their subordinate status. The "mixed" array had cbariots in the centre, cavalry on their flanks, elephants on the wings, which seems to have the elephants dangerously exposed. The "array breaking with the centre" has elephants in the cenere, chariots on their flanks, cavalry on the wiogs. The final deployment is wich chariocs in the centre front, infantry on cheir flanks, cavalry on the wings, and elephants in the rear. Generally these deployments ignore reserves, though they are given passing mention. An army could be organised in severa! divisions, which could operate in li ne, "the staff', in column, "the snake", in "cirde" which seems intended to surround an advancing enemy, or could be deployed to refuse the flanks or the centre. The best troops were usually in the centre, but could be deployed elsewhere, to attack the enemy's weak points or mass against his strong points.

The Arthasastra gives the frontage ofvarious troop types. Infantry should beone sama, about afoot/30cm apan, whicb adding tbe distance taken up by the man himselfshould be dose to the cwo-cubit frontage ofHellenistic dose order. Cavalry were tbree samas apan, elephants or chariots five. Double or treble spacing could be used but was apparently not standard. Poros' elephants at the Hydaspes were actually 100 feet/30.5m apan. The numbers of ranks, or the depth aJlowed per rank, are not giveo, but both iofaotry and cavalry seem to use dense mass formations rather than shallow lines.

Elephants were the main shock arm, intended to win by breaking the enemy li ne. Their psychological impact was also considerable, especially against men and above ali horses which were not used to them, though this was probably oflittle imponance in an internal lndian war. They could also serve as a rallying point, Poros' cavalry twice falling back on their own elephants. "Marching in the van ... securing victory single-handed, restoring

53

broken ranks, breaking up unbroken ranks ... causing terror, showing magnificence ... these are the functions of elephants." A generai on his elephant was a clearly visible inspiration and rallying point for his army, but was vulnerable lo anack; thus "one appearing as che king should be stationed al the head ofthe army".

Chariots were similarly used to break the enemy line, though more oftcn in support of cavalry than against sol id infantry formations. Many lndian chariots were big heavy vehicles with large crews, which gave them considerable imperus and fircpower, but meant they were slow, clumsy, and liable to bog down in muddy ground; "(ground) not causing jolting, withoul fields under water and enabling a rurn around, is excellent for chariots". Again much of their effecl was psychological; " ... reuniting broken ranks, breaking up unbroken ranks, causing tcrror, showing magnificence, and making a frightful din, these are the functions of chariots". lndian cavalry do nol seem to have been very good. They were both outmanoeuvred and outfought by the Macedonians, and seem to have paid lirtle attention to scouting. They were used to seize suitable banle sites or ocher key poincs, to raid and destroy enemy supplies, exploit gaps in che line, arcack the flank and rear, pursue a beaten enemy or fend offpursuit after a defeat, and to rally broken troops.

Infantry are so much a subsidiary arm that theArrltasasrra never describes their rote in great detail - "Bearing arms in ali places and seasons, and military activity are che functions ofinfantrymen". Only infantry were much use in wooded or broken ground, but it is clear tha t even they were not well adapted to this and preferred the open.

"Without thorns, not very uneven, with room for reueat, this is excellent ground for infantrymen". Neither their closc formations, nor their long and unwicldy bows, swords and shields were well suited to woods or scrub. They were primarily missile troops, Nearchos observing that despite their powerful cwo-handed swords infantry did not often advance to fighl hand to hand in an internal Indian war. The Artltasasrra notes that armoured infantry should be placed in front of che archers. These are presumably the javelinmen, though they were far from always armoured. Some infantry could be attachcd to chariots and elephants to suppon them at dose quaners, guard their Oanks and prevent them from being hamstrung. One early source allots rwo infantry lo each chariot, eight lo an elephanl. The practice is also menĂšoned in che Mahabltarata.

Despite the preference for open warfare as mosl honourable, various ruses are suggested in the Arrhasasrra. They include a liberal use of spies and bribery, feigned Oights or diversionary anacks (usually using forest tribesmcn, allied troops or deserters rather than one's own men), Oank or rear auacks, tempting the enemy lo raid a cattle herd and then ambushing them, attacking at night, or keeping che enemy awake ali night by feint auacks and then auacking next day when they are exhausted. A generai could try to scare his enemy's elephants and horses by stampcding herds of cattle, buffaloes and camels towards them, laden with sacks of pebbles to produce a frightful noise. Ifa king was short ofchariot horses, he is advised co harness bullocks to some ofhis chariots - ifthis was an acceptablc substitutc, one wonders what speed Indian chariots were normally capable o( Not much, ifbullocks could produce a turn ofspeed that was even approximately comparable. Similarly, "one with few elephants should makc the ccntre consist ofcarts pulled by donkeys and camels". While the camels might upsct che enemy's horses, and the cans could presumably concain archers, it is difficuh to see these cans being of great military value.

CARTHAGE Carthaginian armics ofthe 4th century were esscntially hoplite armies; the cquipment ofthe citizcn and Libyan infantry did not correspond exactly to hoplite weaponry, but they fought at dose quarters in similar closely packed phalanxes, while the other arms were very much subordinate to the heavy infantry. An army in Sicily in 309 fought in two separate phalanxes, one ofGreeks and one of"barbarians", no doubt the Libyans. The various lighter armed mercenary troops seem to havc bcen regarded as oflittle importance, "a mixed crowd ofrabblc also followed along for the sake of booty, men who are of no use whatever to an army". There is little sign of any attcmpt to take advancage ofthe different weapons and capabilities ofthe various nationalities, except the Balearic slingers who are admiringly mentioned. Nor did Carthage's superiority in cavalry make much mark; the Krimisos defeat was a consequence of a complete lack of cavalry scouts, while che best che cavalry could do at Tunis in 31 Owas a forlom frontal attackon Greek hoplites. The Numidians howevcr are found in their traditional harassing rote by 309. Thc most disĂšnctive Carthaginian arm at this stage was che chariot. At the Krimisos thcse effectively protected their own infantry from tbc Greek cavalry's ancmpts to prevent them deploying; "the chariots drove up and down and protected the front. Tbc horsemen were compelled to wheel about concinually so as to prevent their own formation bcing broken up". Bue at Tunis che chariots too were wasted in a frontal anack on hoplites. Since chariots seem only to havc been effecrive against cavalry, in which arm the Carthaginians were generally superior anyway, it is not surprising that they dropped out ofuse.

54

The fĂŹrst major tactical change was the introduction of elephants. At Akragas in 262 the Carthaginians seem unsure how best to use their new weapon; the elephants were drawn up in the second line, and swept away when the first line ofmercenaries broke. Xanthippos at Bagradas used his 100 elephants in a straightforward charge to open the battle, trampling down the Roman line. Although at Panormos in 250 the Carthaginians learm the vulnerability of elephants thus used to open the attack, this became the standard tactic. Ha milcar Barca used bis elephants to charge down the rebel mercenaries in open baule, as well as to trample prisoners. Hannibal at the Trebia initially used bis to disorder the Roman cavalry, whose horses were not used to them, but when they were driven offby Roman light infantry he regrouped them and reverted to the traditional role, sending them in to charge and break Rome's Gallic allies. At Zama he also used them to open the baule with a charge against the enemy infantry, wbich mauled the Roman velites badly but made liule impressioa on the heavy infantry. Another use ofelepbants was 10 force a way iato an enemy camp; Hanno did this against the rebel mercenaries, Hannibal near Capua in 211.

Xantbippos at the Bagradas also employed another tactic which was to become standard, a double envelopment.

The superior Carthaginian cavalry would chase offtheir Roman opponents and then fai! on the flanks and rear of the infantry. This was Hannibal's tactic at the Trebia and Cannae, but was already established practice by the time he left Spain, as Hasdrubal tried the same thing at Ibera. The Cannae envelopment shows the unusually high degree of discipline and control Hannibal had instilled into his cavalry. The Gauls and Spaniards on his left charged and broke their Roman opponents, but did not get out ofhand and pursue; they swung across the Roman rear, charged the Italia a horse on the other flank, regrouped aga in, and delivered yet a third charge on the rear ofthe infantry. The skirmishing Numidians also made a considerable impressioa; they are a good example ofhow Hannibal used the different capabilities ofbis troops to the best advantage, a principle markedly absent in Carthaginian tactics ofthe previous century. In pitched battles the Numidians would usually skirmish and refusc contact with the enemy.

Tbeir scouting was invaluable, and they could raid and forage at will, dcvastating enemy territory and gathering in supplies. lt was thanks to the desertion ofmost ofthe Numidians to Rome thae Hannibal was decisively outnumbered in cavalry at Zama and could not try an envelopment, but fought a straight frontal banle.

The danger with tbe double envelopment was that ehc enemy might cui their way through the cenere before the flanks closed their trap. This happened at ehe Trebia :md, at greater cost, at Ibera. In che centre Hannibal usually poseed bis ex penda bi e troops - bis Gauls - bue at Cannae he stiffened the Gauls with alternate units ofhis vereran Spaniards. Rather than refusing bis centre - the obvious course when trying to envelop the enemy, bue wbich might give warning ofthe trap - he acrually threw ehe centre forward, relying on the Gauls and Spaniards to give ground only slowly, holding up the Roman advance while the flanks closed in. Again, it is a tribute to Hannibal's leadership that the notoriously fickle and short-winded Gauls actually stood up to this prolonged punishment. Tbe Libyan vetera ns were refused on the flanks, to advance later and anack the Roman flanks once they were engaged.

However good African infantry could be stationed in the centre on other occasions. The light troops, javelinmen and slingers, were now used more advenrurously and offensively than before. They carried out all the classic light infantry tasks - screening deployment, supponing cavalry, joining in the envelopment of thc flanks and rear, seizing and holding hills, ambushes - where necessary carefully supporting them with heavier troops, such as Spanish scuran'i. Hannibal's light troops, at least in the early stages ofthe war, displayed a marked superiority over Roman light infantry.

Hannibal's success did not depend solely on his organisation and inspiration ofthe troops, or his handling ofthem on the battlefield. He did bis best 10 Iure tbc enemy into battle on his terms - most obviously by ambushes as at Lake Trasimene, bue Hannibal was constantly using his Numidians or light infantry 10 harass the enemy and provoke them imo battle, by raiding allied territory before tbc Ticinus or by harassing the Roman camp before the Trebia. With the famous exception ofFabius "Cu11ctaror '~ tbc Romans usually obliged.

THE NUMIDIANS Numidian cavalry were famous s.kirmishers. Operating "in haphazard groups" (Sallust on Jugurtha's men) they would harass the enemy with a rain ofjavelins, out:flanking and surrounding them, but scattering to avoid enemy charges. Heavier cavalry such as Romans could make short work ofNumidians ifthey ever managed to catch them, but usually "could not cope witb the enemy, as the Numidians easily scattered and retreated, but afterwards wheeled round and auacked with great daring - ehese being their peculiar tactics" (Polybios). They favoured ambushes concealed by dips in the ground or oeher terrain fearures. Though normally reluctant eo fight hand eo hand because oftheir small moums and light armamene, they would attack outflanked enemy, while Massinissa's cavalry seized the opportuniry ae Zama eo chase offSyphax's men when ehese were momentarily disordered by retreating elephants. Numidians were well suited for pursuit, "mose efficiene and formidable when in pursuit of a flying foe'', as Polybios puts it; bue perhaps because ehey expeceed such a vigorous pursuie, when beatett themselves 55

they wouJd rarely attempt to rally, but "once they give way continue their night for two or three days, trying to get as far away as possible". Sallust adds "no Numidian, after a rout, returns to bis post in che king's army".

Numidian and Moorish infantry skirmished with javelins in a similar manner to the cavalry. Syphax's Romanc:rained infantry were intended as dose combat troops, but despite an initial success against a Carthaginian force, the only accounts oflater battles in which these troops may bave been involved are heavy Numidian defeats. After ali, there was only so much that one Roman centurion could do in thc way of inculcating effectiveness. Later infantry appear again almost exclusively as skirmishers. Ligh1 infantry wcrc occasionally mixed in with 1he cavalry, giving the latter more willingness to charge; bue when Sallust mentions jugurtha doing this he specifically notes il was unusual, and Syphax' s cavalry were very disc:ressed when the Romans pulled the same c:rick on them.

THE SPANIARDS Spaniards wcre regardcd as good soldiers, fierce and de1ermined if ill-disciplined; they were in demand as mercenari es, and it took Rome 200 years to complete the conquest of che pcninsula. Spanish troops might get out ofhand ifthey were winning, becoming overconfident and disorganised; but in sharp contrasc to che usual GraccoRoman picture of"barbarians" they did not despair if things went badly, but fought doggedJy on. lberian garrisons would commi e mass suicide or sally out in desperate charges "blind wich fury" rather chan surrender; Stra bo says every Ibcrian carried a via! of poison for suicide. This stubborn inflcxibility did noc extend co keeping their word, at least to foreigners, and Spanish chiefs and mercenaries changed sides between Rome and Carthage with alacrity - Livy blames "the restless Spanish 1emperament, always hungry for adventure and change". The Celtiberians successfully combined Celtic ferocity with Iberian s1ubborness; Cehiberian Numanria held out as desperately agains1 sicge as Iberian Saguntum, ending with a similar mass suicide. They fought Rome off and on for 60 years, a s1rugglc Polybios called the "fiery war" because ofi1s uninterrupted ferocity.

Spanish equipmem and 1acùcs were famous for lightness and mobility. Livy says they were "well accustomed to mountain warfare, and well adapted by training and equipment to rapid assauJts over craggy and brokcn ground", and Plutarch describes "the swift attacks and retreats ofa set offleet mountaineers". Even the horses, says Strabo, were trained to dimb mountains. Nor were they stopped by rivers, Hannibal's Spaniards swimming thc Rhone supported on their shields. The Celtiberians may not have been so adept in difficuJt terrain, Livy recording in a battle against Rome "the Celtiberians usually fight by a series of rapid skirmishing attacks, but their speed of movement was rcndered useless by the rough and broken ground".

Naturally wìth Lighlly armed troops in "the narrow passes between wooded hills which are typical of Spaio" ambush was a popular tacùc, especially with the Lusitani. So were raids and skirmishes, in short ali the tactics one would expect from a country which invented the wordguerilla. In pitched battle the main force was the scurari1~ who would advance performing wardances and charge with a shower ofthrown spears, then closing with swords. This initial charge was often powerfuJ enough to brcak through even a Roman line; ifit was held, the Spaniards were stili formidable with swords, but Roman discipline and armour would usually beat them. A variati on on the usual single mass of infantry was 10 form three bodics with gaps betwccn them, to allow cavalry to charge through. The Cchibcrians used a wedge formaùon "in which they are so powerfuJ that the troops in any piace at which they hurl their attack are unable to withstand the shock".

Cae1ra1i would skirmish, or were prepared to charge home to drive off enemy skirmishers. Moorish infantry and Numidian cavalry were no match for cae1ra1i and Spanish cavalry, who could catch them, unlike many heavier troops, and then beat them at dose quarters - "troops who were their equals in speed and thcir superiors in suengch and daring". Otherwise caerrari might forma reserve in barde, deploy on che flanks, or support che cavalry.

In this case they would ride pillioa behind the horsemen, dismounting to fight.

The cavalry, too, were prepared to fight on fooc when necessary, someùmes led by a generai. Hanaibal used his Spanish cavaJry as shock troops; they and the Gauls are called "steady" in contrast with the Numidians. They would charge enemy cavalry or the flanks and rear of infantry. They might also skirmish; an Imperia! Roman cavalry tactic, the Cantabrian circle, betrays its Spanish origin by its name. Here che cavalry would gallop round in a cirde showering the enemy with javelins, presenting a fasc-moving cargec co retum fire and protecting themselves with thcir shields on the oucside of 1he cirde. Anocher form of circle was a defensive tactic, mec once when surrouaded cavalry sat their horses in a ring, facing outwards. Cavalry would scarc a batcle on che wings or in reserve. They were good troops, but not usually ofgreat tacticaJ importance, partly because ofthe rough terrain.

56

THE CELTS

There was a strong heroic ethos in Celtic society, and their warriors favoured single combat. Banle would open with champions advancing from the ranks to challenge the enemy. Tbey would work themselves up into a banle frenzy, leaping about, proclaiming their own heroic deeds and their warrior ancestry, waving their weapons and clashing them against their shields "bowling and singing as their custom was, shaking their shields above their heads and brandishing their spears", ali to encourage themselves and daunt their enemy. A Celt wbo killed an enemy, especially in a duel, would sever his head as a trophy, tying it to his horse's harness or sùcking it on his spear (heads would be taken home, preserved in cedar-oil and sometirnes displayed on the gates). Often chariot-warriors would be driven into battle but fight afoot, and this may be connected witb thc cult ofthe duci - Homer's chariot-heroes, who also fight dismounted, are a dose para Ilei. So may the rrimarkisia, "three horses", Pausanias describcs. He says each Celtic horseman was accompanied by two grooms, who tbough armed and mounted would halt behind the lines and not fight; they would carry offtheir master ifbe was burt, bring him a spare borse ifnecessary, and ifhe was killed or wounded one would replace him in banle, but otherwise they would 1ake no acùve pan. This seems an appalling waste oftwo thirds ofthe army's horsemen, but makes sense ifthe baule is seen as a series ofheroic duels.

The single combats would probably continue unti! one side was sufficiently encouraged by their champions' successes, 1heir own capering and chanting and the blaring of their many warhorns and trumpets, to deliver a generai charge; at this the less enthusiastic si de would probably collapse in preuy shon order, the result ofthe baule thus being largely decided by the duels and other psychological preparations.

This procedure was aU very well when fighting another Celtic tribe who would observe the conventions, but though 1here are instances ofRoman champions accepting challenges and defeating Gallic heroes in single combat - a non-Celùc army might well not oblige by standing around while heroes duelled. In this case thc Celts would generally deliver a ferocious mass charge, generally accompanied by the usual leaping and screaming to work up their own courage and daunt the enemy; "They rusbed at their adversaries like wild beasts, full of rage and tcmperament, with no kind oftraining at ali ... tbc blind fury never left them while there was breath in their bodies ... even with arrows and javelins sticking through them they were carried on by sheer spirit while their life lasted".

In the case described, the anack on Thermopylai, tbe Galatians kept up this almost berserk enthusiasm, despite making no headway, till their leaders called tbem off; but usually they quickly weakened ifthey did not break through the enemy witb their first charge - "Gauls in generai are most formidable and spirited in their first assault, while stili fresh" - not simply because they lost beart, but because, though strong, they lacked stamina, and were panicularly vulnerable to heat and thirst, so they were obviously at a disadvantage wben sweeping south into tbc hotter lands ofltaly and Greece. In addition tbey often lacked tactical sense, and since they usually charged in one mass, iftbis was stopped or outflanked they had no reserves to fati back on, and would often collapse at once.

The most ferocious cbarges were associated with infantry who fough1 naked, like the gaisatai at Telamon (thougb there they were wom out by vain charges against skirmishers) and as the custom declined so did Gallic ferocity, perhaps "softened" by contact with civilisation. In tbe cast, although the Galaùans stili fought naked in 189 against Rome, they did not show tbe savagery and spirit oftheir ancestors in 279, but took up defensive positions and waited passively for the enemy to come to them, instead of attacking. There may be some truth in the Roman idea that they had become enervated by the heat and the lwcury ofGreek culture.

Infantry usually fought in very closc formations, often overlapping shields, so that Roman pila are recorded piercing two sbields and pinning the wielders together. Livy calls their formation tescudo, the Roman " tortoise".

Such very close formaùons were probably cbiefly used to receive attacks, as tbey would narurally loosen up in a wild charge. Livy for instance only empbasises the Gallic close order at Sentinum in che last stages ofthe baule, when the Gauls were thrown on the defensive. They stili seem to have charged in fairly close order, however. The reluctance to abandon their dense forma1ions may help account for their generai ineffectiveness in rough terrain. Galaùans especially proved belpless in mountain country in Greece or at home in 189. Some tribes living in the bills naturally fought better in such terrain; Hannibal was anacked by Alpine tribes wbo "carne swarming down the rocky and precipitous slopes, sure-footed as they were from long familiarity with 1heir wild and trackless terrain". Similarly the tribes ofnorth ltaly were enough at home in forested country to stage a successful ambush by felling trees on a Roman army in the Litana forest in 216. But in generai Celtic tacùcs were best suited to open country. Tbey were vulnerable to enemy skirmishers, because oftheir dense formations, and because while most men had a spear or two notali were suited for long range shooting, and the supply was very limited. Galatians in 279 are recorded tearing enemy spears out oftheir own wounds and hu rling them back, and in 189 they resorted to throwing stones. Celùc infantry were not fast enough to chase offand catch enemy skirmishers. They had very few light infantry oftheir own; those there were are usually found supporting the cavalry.

57

The pic1urc ofthc Cclts as wild, ferocious, bul erratic, disorganised and lacking in caccical sense, while popular witb Greek and Rom:m writers, is nor rhe whole truth. As early as 386 Gauls decisively oucgeneralled che Romans at the Allia, seizing a key hill :md turning rhe flank. At Telamon the allied Gallic army, caughc berween rwo fues, formed up calmly and in good order facing both directions in a position which greatly impressed observers, and che Romans "were terrified by lhe fine ordcr ofthc Celtic host". Nor did rhey weaken when rhe banle went againsr chem, bue "rhough being almost cut to pieccs, they held their ground, equa! to their foes in courage" and most died where they stood. They had showcd similar determination at Sentinum, standing cheir ground even wbea cheir Samnice allies fled past them. lt is clear that, sometimes at least, Celts could produce more chan jusc a wild charge.

Cavalry would deploy on che flanks, opening the baule by charging the enemy horse with spear and sword, perhaps lhrowing javelins on che way in. They are found raiding, scouting, ambushing enemy scoucs, blockading enemy camps and seizing key hills. They would somctimcs dismount if thc terrain dictated it, as at mount Magaba.

Chariots usually supported the cavalry. At Sentinum they were in reserve behind the horse, at the "elephaat" battle of273 they staned behind the horse, but advanced through them to open the battle, while at Telamon, with the cavalry engaged separately, the chariots covered the flank of the infantry. They would usually skirmish, chiefly against enemy cavalry, shooting at them with javelins and trying to break up their formation. The warrior would sometimes dismount to fight against an enemy force which scood firm. At Sentinum chariots successfully charged home against Roman infantry, but only when thcse had been thrown imo disorder by the rout ofcbeir own cavalry. il is unlikely that chariots would charge infantry who stood firm. It is significant tbat at che "elepbant" battle, che conventional light chariots werc placcd on the wings, and only the scyched vehicles were in the cenere intended to break up che formation ofthe enemy infantry.

Celtic armies werc ofccn accompanied by large numbers of wagons. In the case of a tribal migraĂšoa, or a cribe having enlisted wholesale as mercenaries, these could conta in the warriors' families and belongings. Otherwise they migbt bold boory. Theycould be formed in a l:iagcr behind the battleline. At Telamon they are found on the flanks, probably as a barrier against outflanking rather 1h:10 intended 10 advance along with the chariots.

ROM E Roman tactics subordinated the other arms very much to the heavy infantry. A standardised formation and tactics had been developed which emphasised defence in dep1h and reserves. A fonified camp, with a standardised layout possibly influenced by that of Pyrrhos' camps, formed part of this systcm as a last line of defencc or a refuge fora beaten army. The camp-guard could be used as the last battlefield reserve. The nccd for this systematic tactical routine had much to do with the style of Roman gcneralship. Command was in the hands of elected magimates; they would ali have had some military experience, but often no cxpcricncc of command, and they were nor always cbosea for tbeir generalship; cbey often displayed a conspicuous Jack ofit. Roman armies ofcen had to win despite their generals. A common failing ofRoman commanders was overconfidence, which contributed for instance to che defeacs ofche Trebia and Cannae. Excessive zeal was also a problem with the men, who would sometimes demaad battle againsc cheir generals' better judgment, or attack without orders - nor always unsuccessfully. Another weakness was reconnaissance; 1ime after time 4th century armies blundered cheerfully imo Samni1e ambushes, most notably at tbe Caudine Forks. Thc lesson was temporarily !carne, as we hear of Roman scouts acrually detecting an ambush in 298, bue in 2 17 they were back to 1heir old habits, marching without scoucs into the ambush ai Lake T rasimene. The second Punic War was in fact che great tria! ofthe Roman army. le cmerged with the basics of ics sound caccical system unchanged, but transformed into a far more versatile and professional force.

In a normai rwo-legion army, Roman citizen infantry would form the centre with the two allied alae on their flanks - they were known as "the ala o[ the left" and "of the right" - while in larger armics legions and alae would usually alternate. An unusual deeper formation was twice used against Gauls, with one ala (and in one case the extraordi11ariz) in advance, and both legions behind in supporc, lest che allies wcre swcpc away by thc wild Gallic charge. There was a ceadeacy 10 use allies in such exposed positions, for instance on che outflanking march which turned che Macedonian posicion in hill country before Pydna. The extraordi11an"i also had the dangerous privilege ofleading the army on rhe march. Non-ltalian allies might be massed on one wing, like the Numidians at Zama or the Pergamenes at Magnesia, or puc in reserve if chey were not trusted, which was done with Ccnomani allied against other Gauls.

Each legion (and probably each ala) deployed in rhree lines - four including the vtlites skirmishing in front. The lrastati formed the first heavy infantry line, their maniples deployed with intervals betwccn thcm. The maniples of 1he pri11cipes in the second line were positioned co cover chcse gaps, in thc famous "chequerboard" formation .

Similarly the units oftria rii in che third li ne covered the intervals in the second. Livy says the 4th century rorarii and

58

acce11si deployed behind the m·an·i. Scipio at Zama unusually pur the maniples of pri11cipes and rriarii directly bebind tbose ofthe hascari, so tbat tbe intervals berween the maniples formed lanes right through che formation, down whicb tbe enemy elephants were co be lured. The chief purpose ofthe intervals was to allow one line to be relieved in combat by the uoops behind; this line-reliefsystem was one ofthe dìstinctive fearures ofRoman tactics and meant that ifthe first line were d'efeated fresb troops were always available to meet the threat. This contrasts sharply with the scant attenòon paid to reserves even by most ofthe sophisticated Hellenistic gene rais. Ifthe lzastari were defeated or exhausted they would dìsengage and reòre through the gaps between the maniples ofthe pri11cipes.

These would be facing an enemy òred and dìsordered by their fight with the lzasrarz~ but ifthe principes in 1urn were defeated they could similarly retire tbrough the third line, and the rn·arii, who had been kneeling quictly in rescrvc, would be able to continue tbe banle or cover a retrcat. (Ahernatively it was occasionally necessary to bring the pn"11cipes and rriarii up on to the flanks ofthe hasrari to face a long enemy line, as in the last sragcs ofZama.) lt became proverbial that the rrian·i were the last resort. (lo fact the last resort was often the camp guard; 1hese may someòmes bave been rriariz~ but could be foreign auxiliaries as at Magnesia.) It is highly unlikely, however, 1hat the maniples normally fougbt with these intervals left between them, as the danger of small units being outflanked and swamped must bave been great. Livy indeed says that when the rriarii entered battle they dosed up and fough1 as a solid li ne, and no doubt the others dìd too. The two cenruries into which a maniple was divided were cali ed prior and posrerior, front and rear; they probably advanced and manoeuvred with the posrerior century behind the prior, as the name suggests, and with gaps between the maniples, but when they advanced to fight 1he posrerior centuries would move up into line with the prior, filling the gaps and prcsenting a solid front. When a line withdrcw, the posrerior cenruries would break off first and resume their rear position, allowing the maniples to reti re 1hrough thc gaps in the line behind. Typical fighting depth of each line was probably six ranks.

Within the mani pie Polybios says the Roman soldier formed up in much looser formation than other heavy infantry.

Not only did each man take up a frontagc of3 feet190cm, but there was a three-foot gap be1ween each rank and each file; in short, there was one legionary to each six feet offrontagc, equivalent to the Hcllenistic open order. Polybios says tbis was necessary to use tbe sword and for the soldier 10 defend himsclfall round with his shield. This implies the Roman was essentially an individua! fighter, and indced he certainly was capable of fighting this way. For instance Livy records them taking on Celliberians in scrub which broke up 1he forma1ions, and bea1ing 1hem in man-to-man sword-figbting, while the Roman 's heavy equipmcnt gave him a decisive advantage agains1 a Macedonian phalangite when they were reduced to individuai combat. But it is dear from othcr accounts thai Romans would usually fighe in dose order; probably they would dose up aftcr throwing pila in most circumstances.

A dose formation could literally push the enemy back, shoving with the unusually largc and heavy Roman shicld.

Against slashing weapons such as the Celtic sword the legionary would sheher under his shield and stab upwards with his own short sword. Atone point it was thought 1hat 1he answer 10 1he long sword was the 1hrusting spear, which could outreach the Gallic sword and parry i1s blows; thus in one battle in 223 the spears of1hc rriarii were given to the front line rroops, but the experiment was not repeated. In the mid-4th century 1he spear was stili the majority weapon, only hasrari carrying pila. These were ini1ially a selec1 force, intended no doub1 10 dìsrupl an enemy charge with showers oftheir new weapon, whereupon the principes, on whom wc are 1old the Romans chicfly relied, would charge and take advantagc of 1hc confusion. (The problem that the princi'pes, whose name means "first", fought as the second li ne disappears ifit meant first in importancc rather than posi1ion.) The reliance on the pri11cz'pes, and the fac1 that the lzasraribcgan as a select - and hence small?- forcc wi1h a new weapon suggests there may have originally been fewer hasrati 1han pri11cipes, rhough later the numbers were equal.

Against the Macedonian phalanx volleys ofpila usually failed 10 disrupt the formation. Livy says pila were tbrown "to no effect against a formati on resembling a reswdo of dose-packed shields" in one dash in 198. This took piace in the breach of a ciry wall, so 1he Macedonian flanks were covered, and the Romans were unable 10 make any headway. This was the rypical resuh iflegion met phalanx head on with intact formations. The Romans would try to cut their way through thc hcdgc of pikcs with swords, or force the poims aside with their shields or their bare hands; but they would generally not succeed because of the sheer number of pikes facing them - five points projecring in front of each file of Macedonians - and many pikes would become firmly embedded in Roman shields, keeping the legionaries out of sword reach. A shoving match would resuh, to the advamage of the deep phalanx formation. Rome's victories over phalanx armies were won bccause the phalanx became disordered, and then the greater flexibiliry ofthe small Roman units and the heavier armamene ofthe Roman swordsmen gained the victory. At Kynoskephalai and Corinth thc phalanx was disordered by attacks on its flank and rear when it was already engaged. At Magnesia the legionaries stood off and did not charge till the pbalanx was 1hrown imo chaos by the elephants in its intervals. Pydna offers the best example of the legions' superior versatility; the Roman centurions led their men into gaps which appcared in thc phalanx as it advanced over rough ground, attacking the pikemen's exposed flanks. The Romans had both the cquipment to win this sort of individuai fighting and the leadership at maniple leve! to exploit it.

59

This flexibility and leadership was thc great advantage ofche legion in the latter part ofour period. lt is someùmes said that the introduction ofthe manipular system led to such flexible tactics, and that the example ofthe mobile Samnices inspircd improvement. This does not seem to be the case; there is little sign of mobile tactics or the individuai use ofmaniplcs throughout che 4th century or most ofthe 3rd. The potential mobility ofthe manipular syscem is used for lini e more than the line-reliefsystem; the bactle is a straightforward slogging match between sol id infantry lincs. Occasionally a whole li ne ofa legion will be used separacely; against Samnites in 297 the hastati ofone legion, relieved by thc pri11cipes in the usual way, chen marched around the Samrute flank under cover of high ground, to attack from the rear. But maniples are not used independently and it is panicularly noticeable that the sort of small-scale tactics and iniciative by centurions or cribunes which won Kynoskephalai and Pydna is not seen against Pyrrhos' phalanx.

The sccond Punic War was the greac divide. Hannibal handled his unics with an originality, flexibility and flair which made a deep imprcssion on the Romans, while the length ofche war and the size ofthe armies involved built up a pool of expericnced officers and men with much more service cxperience than the cssentially pan-ùmc soldiers of earlier wars. Scipio Africanus' battles represent major departures from the straightforward frontal attack by three-line legions. His tactics wcre inspircd by Hannibal's and needed good officcrs as well as generalship to implement. He favoured pinning the enemy cenere with light or auxiliary troops while enveloping the flanks with heavy infantry. The flanking movement at llipa in p:micular was a complicated manoeuvre, carried out quickly and efficiently with the troops whecling by maniplcs from line imo column, then into line again in position to outflank the enemy. After this war, a greater degrce of tactical flexibility and officers' initiative could almost be taken for granted.

One new tactic first mentioned in Macedon in 200 was the teswdo or tortoise. A group oflegionaries would overlap their shields for protection from missiles, thosc on thc flanks holding their shields outwards, those in the centre holding them above their heads to forma roof; this created a virtually invulnerable protection. lt was common in sieges or streetfighting to fend off missilcs from walls or rooftops, but in 200 was uscd in the field. One testudo was formed by 60 men (ie a century) and was fivc ranks deep.

Despitc the victories of the mani pie, it was replaccd by the larger cohort, which apparently was more useful against barbarians, cspecially in Spain whcre it seems to havc first bcen uscd. The open chequerboard formation ofthe manipular lcgion may bave becn admirably flcxiblc, but there was a risk it could be swcpt away by a wild charge, especially if caught before the maniple.s had dosed up for batti e. The cohort allowed greater concentration on the battlefield; it also, convcrsely, allowed greater strategie dispersal, making it easier to distribute forces ofless than legion strength, valuablc against guerillas in Spain. Fora century manipular and cohorc tactics co-existed, and by the late 2nd cenrury there was a regular drill available for forming cohorts from maniples on the battlefield.

Light troops and cavalry Roman light infantry were essentially skirmishers, who would open thc battle screening the hcavy infantry, withdtaw through their ranks, regroup on the triarii and either stay in rcservc or threaten the enemy flanks. By the 2nd cenrury however they were well enough armed and trained to attack enemy skirmishers hand to hand and chase them off, bcing used this way in hill country before Pydna. Their javelins were pa rticularly useful for harrying Celts, and they were well enough armed to stand up to individuai Celtic warriors in dose combat, though a concerted rush would have swept them away - ifit caught them. They could also support cavalry; outside Capua in 211, to compensate for the Campanian rebels' cavalry superiority, light infantry rode pillion bchind Roman cavalry, dismounted when dose to the enemy horse and hurled their javelins. The surprised Campanians were then charged by the cavalry and broken. Against Numidians in 203 light infantry forced their way into a cavalry fighe and turned the tables. T his contrasts markedly with the dismal performance oflight infantry supporting horse in the Ticinus sk.irmish in 218, and suggests a marked improvement in their capabilitics.

Cavalry wcre used much likc light infantry, harrying the enemy while the legions deployed, then falling back to the flanks or reserve. Sometimes citizen cavalr:r would form one wing, ltalian allies the other; altemativcly ali the ltalian borse could be on one wing, Numidians or other foreigners on the other. The cavalry would attempt 10 defeat the enemy borse and then fall on the flanks or rear oftheir foot. Roman cavalry were shock troops, charging homc with spear and sword; at Asculum they tried to dose at once while Pyrrhos' Greeks manocuvred for advanrage, while Macedonian cavalry in 200 were highly disconcerted to be charged at once by Roman horse. Ifa stationary melee resulted Roman horsemen were likely to dismount; to some extent this was an occupational hazard for any ancient cavalry, with no sùrrups and poor saddles ifany, but the Romans had a distinct preferencc for it. Polybios records them and their Gallic and Spanish opponents dismounting at Cannae, to fight in a confined space; hc says this was unusual, but other examples dearly show it was not, unless he means it was unusual for the Carthaginians.

60

Cavalry could also ctismount e11 masse to artack enemy infantry iftbeir own foot were in difficulties; while this was more typical of tbe Stb century, Livy records it against Hernici in 362 and Etruscans in 31O. Dismounted attacks had more chance tban a mounted charge of success against formed infantry. Even Samnites, not especially well equipped for fending olf cavalry, repulsed Roman cavalry charges on severa! occasions. Most successful charges were delivcred against the flank or rear of engaged infantry (or on one occasion against the flank of Samnite foot trying a rash outflanking move oftheir own) - especially in 4th-century wars in ltaly, where their Campanian cirizens usually gave tbe Romans a superiority in cavalry, but also at Zama. One peculiarly Roman technique was to delivcr a panicularly imperuous charge by ctiscarding their horses' bridles, to remove ali restraint, and "spurring them on so hotly that notbing could withstand tbe shock". Again this was characteristic ofthe Sth cenrury but in 325 such a bridleless charge broke a Samnite infantry line whicb severa! convcntional charges had failed to shake.

It succeeded again in Spain in 180, tbough by tbcn regarded as an unusual and old-fashioned tactic.

Cavalrydeployed in reserve migbt charge tbrough tbe infantry line. Against Etruscans in 310 and Samnites in 293 Livy records cavalry charging tbrough lanes between the ordines; this could mean between the files, but as he calls the units of the third line ordines it probably means between units. In one case the lanes were opened just before the cavalry charged, in tbc other specially left open in advance; so the cxistence of gaps between units in this case actually confirms tbat tbey were not usuaUy present, but that the units usually fought closed up.

Roman cavalry has been regarded as inefficient, chiefly because oftheir poor performance against Hannibal, but bere they were generally outnumbered. They ctid show couragc and determination, for instancc at the Ticinus, fighting on tbough heavily outnumbercd unti! tbey were surrounded. They also fought well against the Macedonian cavalry, defeating tbem severa! times where equa! numbers were involved. They ncver quite worked out how to cope witb skirmĂŹshing NumidĂŹans, but otherwise ctid as well as could be cxpected, often against heavy odds.

Elephants Rome first met elephants when fighting Pyrrhos in Lucania, and these "Lucanian oxen" made a great impressioa at Heraclea. T rying to work out how best to deal with thcm, Rome first cxpcrimcntcd with gimmicks, likc the wagons used at Asculum (see figure 180), squealing pigs, or incenctiary javelins. These werc no great successes, the most effective weapon in the Pyrrhic wars being the javelin fire ofthe camp guards at Bcncvcntum. Regulus at Bagradas tried to meet the elephants' charge with a deep formation in the hope of standing up to it; this failed miserably. Thereafter Rome reverted to relying on javelinmen, who woundcd and panicked the beasts; at the Trebia javelinmen drove off elephants partly by stabbing them in thc soft skin under the tail. Hcavy infantry also learnt to dea! with elephants by avoiding the charge, throwing pila from thc flanks or hamstringing them witb swords.

In Roman service elephants migbt be used in front ofthe line to brcak up the encmy formation; at Kynoskcphalai and Pydna their charge, supported by Latin allies, swept away one Macedonian wing. At Magnesia, where Rome's African elepbants were outnumbered and outclassed by Antiochos' Indians, and where moreover tbc enemy had abundant missile troops, they were wisely kept in reserve. Elephants were occasionally used in Spain and Gaul, perhaps chiefly to overawe the tribesmen.

THEOSCANS Tbe Samnites had a very high military reputation, and seem to have been the only ltalian nation whose warlike qualities the Romans feared; Livy calls them warlike, brave and srubborn, fighting "with more courage than hope" even in dire adversity. In the Sth century the Campani, too, were brave and ferocious, and in demand as mercenaries. The Mamertini, Campanian mercenaries, stili displayed tbese qualities into the 3rd century. In Campania itself, however, wealth and civilisation seems to ha ve sapped the Oscan marciai ardour. About 345, when they allied with the Sidicini against the Samnites, Livy suggests the Campanians stili had a good reputation, but it was no longer justified; they "brought a reputation rather than acrual strength to protect their allies" and had been "softened by luxurious living". He repeats the harsh judgment at the time ofH annibal's war.

What we know oftactics suggests they were common to ali the Oscan peoples, though the Campani must bave been affected by their greater proportion ofcavalry and hoplites. Most ofthe army was mobile, lightly-equipped infantry at home in hills and woods. Where possible they would surround an enemy force, avoicting contact but pelting it with javelins. Lucani and Brunii beat severa! Greek forces this way. Pyrrhos used sucb infantry to increase tbe flexibility ofhis phalan.x and to winkle Roman infantry out ofwoods. Most ofthe Oscans lived in country ideai for this sort ofwarfarc. Thus whcrcvcr possiblc the Samnites would ambush Roman armies rather than risk a baule in

61

che open, especially since chey were usually outnumbered in cavalry. They would charge downhill at the surprised enemy or blockade them in a narrow vallcy, as ac cheir greacest victory at tbe Caudine Forks. Tbe Lucani similarly blockaded Alexander ofEpeiros in 330 at Pandosia. Once the Samnites, having cbased a Roman army into its camp, tried to pen tbem up by building a palisade right round the camp. The Sarnnites, too, would build a pa!isaded camp for an army on carnpaign.

Wben an ambusb was irnpracticablc or discovcrcd too soon, the Samnites were always prepared fora straight fight.

Tbe Rornans believed the first Samnite auack was the mosc dangerous, and after a while tbey would run out of missiles and their spirits would flag. This strongly suggcsts they had no systern for relieving troops comparable to the Roman triple-line formation. Indecd, rcar lincs and rcserves are not mentioned, and their absence belps explain the Samnite vulnerability to cavalry attacks in che flanks or rear. Their infantry would usually charge fiercely and fighe ac dose quarrers ratber than skirmish with their javelins; che Romans seem to have had a slight edge in such a contese, but Samnite croops worsted rbem more than once. The cohors seems co have been the basic tactical unir. If the Samnices did lack a multi路line system, then since chey rarely seem ro have had a longer line than che Romans, their units may bave been drawn up in greater depth.

Samnite cavalry, though armed with javelins to their Roman opponents' spears, fougbt as shock troops rather than skirmishers, meeting Roman horse hand路to路hand. On one occasion tbey fougbt dismounted for possession of a fa!len general's body. More ofcen than not they lost sucb encounters; this is probably because the Romans and their Campanian a!lies would bave numerica I superiority in cavalry, rather than from any deficiency in quality. The non路 appearance ofSamnite cavalry in some battle accounts may similarly be because they were quickly brushed aside by superior numbers. They usually deployed to procect tbe flanks of their own infantry (though these were often capable ofkeeping ofTRoman cavalry unaided) and tbreaten tbc Rornan flank. However rasb attempts to outflank the Roman infantry could ex pose the Samnite cavalry's own flank to their Roman counterparts. Otherwise they are mentioned as esconing generals, harassing Roman carnps and marching columns, raiding baggage, and occasionally skirmishing as preliminary to a battle. The more numerous Campanian cavalry fought similarly in Roman service, for instance taking the Gauls in che rear at Sentinum.

THE ETRUSCANS According to Livy, in the 4th century Rome feared tbe Etruscans more than anyone else except the Gauls; "there was no other race ... whose arms were more dreaded, not only because cheir territory lay so near, but also because oftheir numbers". Notably absent is any reference to tbeir skill or courage. The Romans bada low opinion ofthcse qualitics, and wcre very surprised when an Etruscan army gave them a hard fighe in 310; "lt seemed as though the Romans were contending, not with che so often defeated Etruscans, bue with some new race. There was no sign of flight in any quamr". Descriptions ofthe 4th cenrury battles are ratber stereoryped, with che two heavy infantry lines rushing co dose quarters, and che Romans invariably triurnpbing sooner or later, chanks to superior courage or discipline, orto their reserves.

It is difficult to extract real Etruscan tactics frorn this, bue tbey dearly relied almost exdusively on their heavy infantry. Cavalry and skirmishers are rarely mentioned, though the cavalry had in che 5th century been singled out as their best troops. Despite other similarities to Roman organisation and armament, the Etruscans may have kept closer to Greek hoplite tactics, using one solid phalanx rather than successive lines, since one defeat in 311 is attributed to their lack of any reserves; there was a small reservc force, but barely enough to garrison che camp which may have been ics role in battle.

Atone battle in 357 the troops ofFalerii and Tarquinii were led by a contingent of priests "who advanced like furies brandishing snakes and burning torcbes, and dernoralised tbe Rornans by this unexpected spectacle". The priests relied solely on their spiritual powers rather than carrying weapons, as Livy calls them "harmless", and when che Romans rallied they "knocked aside the enemy's useless device".

62

MAJOR BATTLES OF THE PERIOD Because of space limitations only briefaccounts ofsome ofthe more significant or tacĂšcally interesring battles ofthe era can be included. Some important battles have to be omitted because details are obscure. Siege and naval actions are not covered. Where strengths are given for the armies they are generally those given by the ancient sources; not ali ofthese are reliable, especially since Greck :md Roman writcrs tended to exaggerate tbe numbers of" barbarian" armies. After each battle I have noted thc main primary sources and in some cases useful modem accoums; the works cited can be identified in the bibliography at the back.

ONOMARCHOS IN THESSAL Y 353 BC Philip ofMacedon fought the Phokians under Onomarchos somewhere in Thessaly. Numbers are unknown but the Phokian force was larger. Onomarchos took up a posicion at the foot of a steep, rocky range ofhills, and concealed forces on the slopes, including many stone-throwing catapults. He cben advanced into the plain, but afcer some skirmishing feigned a retreat in front ofche Macedonians, withdrawing half-way up the slope. The Macedonians took tbe baie and followed, but were surprised when che stone-throwers opened fire, shooting "huge rocks" imo the phalanx. Onomarchos rurned about and counter-anacked downhill, supported by the anillery fire, and thc Macedonians were soundly beaten.

Philip claimed he did not flee, but withdrew like a ram, 10 buu barder tbe next time, and crue to his word rerurned next year to Thessaly with 20,000 foot and 3,000 horse, many oftbem Thessalian. Onomarchos had 20,000 infantry but only 500 cavalry, and because ofthe shortgage he was beaten and killed at the Crocus Plain, his army trapped against the sea and destroyed.

(Polyainos II.x.xxviii.2; Diodoros XVI.35.)

KRIMISOS 341 BC The Corinthian Timoleon, with a Syracusan army of 10,000 foot and 1,000 borse, advanced against a larger Carthaginian army, sa id to be 70,000 foot and 10,000 borses (tbe total ofcavalry, chariot horses and rcmounts). Thc G reeks occupied a hill overlooking a crossing oftbe river Krimisos in western Sicily. The Carthaginians were at first unable to see them, since it was a misty morning, and crossed the river with no precautions against attack, chariots leading, then I 0,000 heavy infantry including the Sacred Band of2,500 Canhaginian citizens, followed by a disorderly straggle of other t roops. Thc Greeks could hear the noisy Punic advance before they could sce it, whilc they remaincd silent on the mist-shrouded hill.

As the mist lifted Timoleon decided co attack wbile the enemy were stili divided by the river. He sent his cavalry to prevent the Carthaginians deploying into line ofbattle, while he followed with the infantry, the best troops in the cemre. T he chariots howcver preventcd the Greek cavalry gerting at the infantry, and threatened to break up the horsemen's formation. Timoleon thus ordered tbe cavalry to switch to a flank attack while he led the infantry charge against the Canhaginian foot. The chariots must either have withdrawn, or else moved out to the Oanks to counter the Grcck cavalry again. As the Greeks charged, a thunderstorm suddenly broke, driving rain and hail imo the Carthaginians' faces, while the resulting boggy ground underfoot hindered them more than the lighter armed Greeks. After a ficrce fight the Punic infantry broke, and many were cut down by Greek light troops in the pursuit, while many more drowned in the swollen river. Tbey lost I 0,000, including the enti re Sacrcd Band, plus more than 5,000 prisoners.

(Plutarch, Timoleo1126-29; Diodoros XVI.79-80.)

SUESSA 340 BC After the first Roman-Samnite war had ended in an agreemem on "spheres ofinfluence" , a coalition ofthe smaller pcoples of the area - Latins, Campani, Aurunci and Sidicini - emerged to defy the two greater powers. The Roman consuls Publius Decius Mus and Titus Manlius Torquarus enrolled four legions, witbout the usual allies as there were none left loyaJ, but apparently comaining accensi as an emergcncy mcasure, and marcbed soutb to join the Samn ites. T he Latins an d aUies assembled at Capua. Their army did not include thc Campanian noble cavalry, who stayed loyal to Rome, nor the Latins ofLavinium, who did not arrive in time. Livy claims tbc Latin army was the larger ofthe two sides.

63

After some brushes between cavalry scouts the two sides met by a mountain near Suessa in tbe Auruncan country.

Manlius commanded the Roman rigbt, Decius the left; the Samnites were on one wing "at tbe foot of the mountain", probably in rougb ground to whicb tbeir style offighting was better suited than the Roman. Since the Laùns wcre armed, organised and deployed in the same way as the Romans, a prolonged infantry slogging match resulted. The Roman hasrati were pusbed back, and Decius, leading the pri11cipes to their support, was killed. The rorarii too were commiued, but the Romans were stili forced back. Instead of committing his triari~ the orthodox course, Manlius kept these kneeling out of sigbt in the rear and sent in the unreliable acce11Si. T he Latins, unable to see the irian·i, bclieved the accensi were tbe last Roman reserve and sent in their own tria rii. Tbese broke through the Roman line after a fierce fight, but were surprised by the sudden appearance ofthe Roman triarii, who charged tbe tired, disordered and astounded Larins and routed them at once.

The Campani retired to the camp, but suffered heavily when it was stormed and taken. Tbe Latins, who seem to have berne the brunt ofthe fight ing, did not rally on the camp but fled back to their own territory.

The Samnites are said to bave played a respectable pan in tbe victory, but Roman historians, prejudiced by long Roman-Samnite enmity, bave not preserved the details, and some denied tbe Samnites any credit at ali.

(L ivy VIII. vi-xi; Scullard ( 1935) pp. 94-5 and Salmon p. 208 for the site, etc.)

CHAIRONEIA 338 BC Pbilip of Macedon invaded centrai Greece wilb 30,000 foot and 2,000 cavalry. He was opposed by Atbens, Tbebes and severa! !esser states wbo fielded abouc 35,000 men, perhaps ali infantry as the famous Boiotian cavalry seem to bave played no part. Tbc allies deployed in a good defensive position with their lefc covered by hills and the walls ofCbaironcia itsclf, tbeir rigbt by a marshy stream. The Theban Sacred Band was on the right, then otbcr Thebans, allies in the cenere, and Athenians on the left. Philip commanded the Macedonian right, entrustfog the lefi to his son Alexander. Pbilip's wing withdrew, slowly and in good order, but the rasb and inexperienced Athenians assumed this was a genuine retreat and followed up exuberantly. This may bave exposed their right flank; it certainly secms to have opened a gap somewherc in tbc li ne, through which Alexander charged, nodoubt at the head of the cavalry; he was said to be tbc first man to break into the ranks of the Sacred Band, which was apparently simultaneously engaged witb the Macedonian phalanx. Tbc Sacred Band died wbere they stood, the rest ofthe allied right took fligbt, and Pbilip renewed bis auack on tbe Athenians, wbo broke and fled in rurn. This baule secured Macedonian hegemony over Greece.

(Polyainos IV.ii.2 & 7; Diodoros XVl.86; Plutarch Pelopidas 18; Alexander 9; Hammond (1973) eh. 16)

GRANlKOS 334 BC Alexander the Great invaded Asia with 43,000 foot and 5·6,000 cavalry, and was met at the river Granikos by 20,000 cavalry and nearly as many Greek mercenary infantry, under a commiuee of generals. The baule is unusually hard to reconstruct as there are two incompatible ancient accoums; I follow Arrian and Plutarch in preference to Diodoros. lt was early summer and the river would be fairly narrow, running in tbe cenere ofa broader flood channel; this bad steep banks, and beyond it on the Persian side the ground was at first flat, rising to low hills. The Persian cavalry deployed on the bank, massed strongest on the left oppositc Alexander himself. Thc infantry wcre wcll 10 the rear.

Alexandcr had cavalry on both his wings, supported on the right by arcbers and Agrianes, with tbc phalanx and hypaspists in tbe centre. H is Greek and Thracian infantry were not cngaged. Thc reason for this may be the same as for thc Persian attempt to hold a riverbank with cavalry, leaving the more appropriate infantry to tbc rear, namcly that this was an encounter battle, with both sides racing for the river. The Persians sent their cavalry ahead oftheir foot and reached the bank first, while Alexander in turn attacked before ali his infantry had arrived.

Hc sent prodromoi, Paionians and one ile of Companions across first, on the right; they werc pushed back but meanwhilc Alexander and the rest ofthe right wing had crossed, edging to their right towards the Persian flank, and charging into the melee on the bank. Afier heavy fighting the Macedonians forced their way up on to the leve! ground and broke througb the Persian li ne. SeveraI Pcrsian generals soughc Alexander out and for some time he was in considerable danger; but thrcc Persian leaders were killed and their men wcre suffering heavily from light infantry forcing their way into tbc cavalry mclee, pcrhaps from the flank. The Persian cenere, engaged by the phalanx and with Alexander pushing through the linc on its flank, was the first to break, and the wings followed.

64

The mercenary infantry, so far unable to take pan in the battlc, halted on the hills and asked Alexander for terms; but he refused, attacking them simultaneously with infantry and cavalry. Though surrounded the mercenaries put up a dcsperate fight, killing Alexander's borse under him and inflicting heavier losses than tbc cavalry bad done.

Evenrually ali but 2,000 Greeks were butchered; the Persian cavalry had lost 1,000-2,500, while Macedonian losses were said to be as low as 85 cavalry and 30 infantry.

(Diodoros' account has Alexander refusing to attack the Persian position, camping overnight, crossing unopposed at dawn, and fighting a battle on the Persian side ofthe river.)

(Arrian 1.13-16; Plucarch Alexa11der 16; Diodoros XVll.19-21; Polyainos IV.3. 16; Hammond (1980) follows Arrian, Lane Fox (1973) Diodoros.)

ISSOS333BC Alcxander marched along the coast road into Syria with 25-42,000 foot and 5,000 horse, only to discover that Darius had taken another route and was now in bis rear with 250-600,000 men, levied from the whole Persian empire except the eastern provinces. Alexander retraced bis steps and forced a baule in a narrow plain between mountains and the sea, where Darius' superior numbers would be restricted. The plain was biscctcd by tbc river Pinaros behind which Darius deployed, throwing bis 30,000 cavalry and 20,000 Ligbt infant ry across tbc river as a screcn.

30,000 Grceks wcrc in his centre, 60,000 kardakes on thcir flanks. Darius placcd himselfbehind the centre witb bis guards, with a vast mass oflevy infantry to the rear. He then pullcd tbc cavalry back, most to his right flank by the sea, a fcw Medes and H yrkanians on the left, and sent 20,000 light infantry (perhaps the screening force withdrawn) into the hills to work round tbc Macedonian right.

Alexander advanced with Macedonian and Thessalian cavalry under bis own command on the right, Greek cavalry on thc left under Parmenio, both wings supponed by Light infantry, and phalanx and hypaspists mixed with mercenary infant ry in the centre. The infantry, marching 32 ranks deep in tbc narrowcst part ofthe plain, expanded till they werc cight deep as thc pia in broadened near the river. Seeing Darius transfer even more cavalry to his right, Alexandcr sent thc Thessalians over to support Parmenio, and ordered tbc archers and Agrianes, with cavalry support, to face the light troops in the hills; these easily chased the Persians further up the slope, and they showed no sign ofwanting to come down.

With this flank secure, Alexander anacked at once ali along che li ne to minimise the effect of Persian archery. At tbc head ofthe Companions he broke through che Persian left and swung in cowards the centre; seeing this, Darius fled.

Meanwhile the phalanx was trying co cross che river and was being roughly handled by the Greek mercenaries. But Alexander's victorious cavalry crashed into the mercenaries' flanks and cut them to picces, rescuing the hardpressed Macdonian foot. The Persian right-wing cavalry charged across the river, but were counterchargcd and held by che Greeks and Thessalians; heavy but indecisi ve fighting followed unti! tbc Persians saw thcir mcrcenarics beacen, at which thcy turncd and fled themselves, the Thessalians eagerly pursuing them.

By now the whole Persian army was in flight; their losses were said to be 100,000, Macedonian 450 dead and 4,500 wounded. About 4,000 Persians and mercenaries escaped with Darius, though other fugitivcs no doubt joincd him later, and another 8,000 mercenaries escaped to the sea and cook ship.

(Arrian 11.9- 13; Curtius III. 8-12; Polybios Xll.17-22; Lane Fox (1973) pp. 168-174.)

GAUGAMELA 331 BC Darius raised another army, allegedly 40-200,000 cavalry and 200,000-t,OOO,OOO infantry, 200 scythcd chariots and 15 elephants. Tbc lower cavalry figure may not be far out; the infantry tota! is a mystery, but only 8,000 or so were any good; the elephants may not have fought. Darius cleared a plain to give a completely flac arena for che chariots. His line was mostly cavalry mixed wich che few good infantry, chariots and heavily armoured cavalry thrown forward, useless lcvy infantry to the rear. Bessos of Bactria commanded tbc left, Mazaios of Babylon che rigbt, and the king was in thc ccntrc. He proposed co surround Alexander with bis superior cavalry. To meet the threat of encirclement Alexander, with 7,000 cavalry and 43,000 infantry, took elaborate precautions. The Macedonian infantry were in the ccntre as usual, but with a second line, mainly Greeks, to face the rear if necessary.

Companion cavalry were on the right and Thessalians on the left, with strong flank guards of cavalry and light infantry held back beyond them. Alexander as usual commanded tbc right, Parmenio the left.

65

Alexander advanced obliquely to the right, to edge out ofthe cleared ground. Darius at first followed suit, but tben ordered Bessos' advanced cavalry to ride round the Macedonian right and stop them. They did so, cbarged and pushed back tbe lìrst unit ofthe flank guard, only to be counter-attacked and tbrown back by light cavalry and mercenary infantry, initially concealed behind the horsemen. Bessos was obliged to commit progressively more units ro this fight; soon ali his flank cavalry were engaged and the line between these and the Persian centre was getùng dangerously thin. The Persian right was similarly held up by Parmepio's force.

Darius had meanwhile launched the chariots to break up the Macedonian line, but most were shot down or surrounded by light infantry, and those which reached the phalanx refused to charge the closely-packed infantry blocks, veering instead down lanes which had been left between the units, to be dealt with by the rearguard wheo their momentum rao out.

Alexander and the Companions now charged at the weakened Persian left centre, and the hypaspists and right-wing phalanx uueis anacked too. The Companions broke through the Persian line, heading for Darius himself. When they broke the Persian horse guards Darius realised the baule was lost and fled. However a gap had opened io the Macedonian phalanx bctween the right-wing raxeis which advanced with Alexander and thc left-wing units which stayed with Parmenio's hard-pressed wing; through this gap chargcd some Persian and Indian borse. They attacked the Macedonian baggage but were counter-attacked and driven offby the rear line ofiofamry. Alexander now roll ed up tbc Persian lioe, heading towards Parmenio's flank. He had some hard fighting against Parthyaean, Persian and other cavalry ofthe beleaguered right centre, some ofwhom succeeded in cutting a way through tbe Companions and making their escape. Soon however the Persian flight was generai, with tbe Macedonians in hot pursuit.

Persian losses were reckoned at 40-300,000 killed, with even more captured, Macedonian at 100-500 dead.

(Arrian IIl.11-16; Diodoros XVI I.53-60; Marsden ( 1964).)

JAXARTES 329 BC While Alexander was subduing Sogdia, the north-eastem outpost ofthe Achaemenid empire, a large force ofSaka raiders massed on the far bank of the river jaxartes (the Syr Darya, also called Tanais by the Grecks) hoping to exploit the confusion. Realising they must be dispersed, Alexander prepared inflated skin rafts and deployed his artillery on the bank. When the catapults opened fue the Saka were surprised by their range and penetrative power - the catapult bolts cleared the river, which the Saka arrows could 001 - and fell back to a safer distance. Taking advamage ofthis, Alexander quickly ferried his troops across the river, archers and slingers first to keep the Saka at a respectable distance.

When his army was across, Alexander, realising that be would need more than a simple attack to catch the mobile horse-archers, sent forward a small cavalry force of four i/ai of prodromoi and a hipparchy of mercenary cavalry as bait. The Saka gleefully surrounded this small force and pelted it with arrows. Alexander then charged them with a mixed force of cavalry and light infantry, sending forward more cavalry in columns on the flanks ofthis force 10 complete the encirclement ofthe Saka. Now trapped between these new attackers and the force tbey thought they themselves had trapped, the Saka were slaughtered - 1,000 were killed and 150 captured. The pursuit which followed was hindered by heat, thirst and dysentery caught from drinking tainted water.

(Arrian IV .4)

HYDASPES 326 BC Alcxander, campaigning in India with perhaps 7,000 cavalry, 40,000 infantry and 5,000 lndian allies, was met at the river Hydaspes (the jhelum) by the loca! ruler, Poros, wbose army was estimated at 20-50,000 good, and an unspecilìed number of inferior, infantry, 2,500-4,500 cavalry, 85-200 elephants and 400-1,000 chariots. The river was swollen and Alexander realised he could not force a crossing in the face ofthe elephants; he proclaimed his intention to wait till the waters subsided and collected supplies fora long scay, but was looking for opportunities for an uncxpected crossing. He kept Poros on his toes by regular noisy nightùme feints; at first these were met by calling out troops and elephants but eventuaJly the Indians' vigilance relaxed and they concluded no real attack was imminent.

Alexander however had found bis crossing piace, a wooded island some 17 miles upstream which would help to conceal the crossing. He split his force into three; Krateros stayed in camp with the lndians, two phalanx raxeis and

66

some cavalry, with orders to atcack across the river if it was left unguarded; Meleager, witb tbe mercenaries and three phalanx taxeis was posted a few miles upstream with orders to anack once Alexaoder bad diverted tbe main lndian force. Alexander himselftook most ofthe cavalry, rwo phalaox taxeis, hypaspists and light troops to cross via his wooded island. He crosscd at dawn bue was sponed by Iodian sentries. Poros sent bis son with 120 chariots and I,000-4,000 cavalry to intercept thcm, but this force did not arri ve till Alexander had completed the crossing, and thc Indians wcrc bcaten by horse-archers and Companions; the chariots, useless in the muddy ground, were ali captured.

Poros now took most ofhis army against Alexander, leaving only the poor quality infantry and a few clephants to face Krateros, :md advanced till he reached firm ground, suitable this t ime for the chariots. H e strung bis elephants out in front ofthe infantry, with cavalry and chariots on che wings. Alexander advanced witb bis infantry beld back, most of the cavalry on che righe, and Koinos with a few Compaoions on che left, ordered to sbadow tbe lndiao mounted troops on that wing. He threw the horse archers forward co auack tbe lndiao left-wing mounted troops while they were stili advancing in column of route, and followed up witb tbe Compaoions. Seeing tbis wing beavily outnumbered, Poros sent che cavalry from his other wing over to suppon them. Koinos duly followed, so when tbe two lndian wings united they were attacked simultaneously by Alexander in front and Koinos in tbe rear. They fell back to rally on che elephants. The Macedonian foot now atcackcd while the lndian infantry and elephants were disordered by che retreating cavalry. Light troops shot at the elephants and attacked their flanks with swords aod axes, while the phalanx, though surfering from the elephants breaking up its ranks, did heavy damage to the lndian infantry. The lndian cavalry managed to attack the Macedonian horse again, only to be beaten aod forced back a second time. Successive Macedonian charges compressed ali che lndiao army imo a confused mass. Maoy oftbe elephant drivers were dead, and the beasts were trampling their own men as well as tbe enemy. Evenrually the exhausted elephants retreated, Alexander's cavalry charged again, and the l ndians, virtually surrounded, broke. lndian losses were said to be 12-23,000 killed, ali che chariots destroyed, and ali the elephants kiJled or taken.

Macedooian losses were 230-280 cavalry and either 80 or 700 infantry.

(Arrian V. 9-18; Diodoros XVll.87-88; Plutarch Alexander 60; Curtius Vlll.14.)

THE CAUDINE FORKS 321 BC The Roman consuls Titus Veturius Calvinus and Spurius Posrumius led their combined armies into the territory oftbe Caudini, tbe western Samnite tribe. Livy says they were on their way to Apulia, but alternatively this may ha ve been an attempt to knock tbc Caudini out ofthe war. Near the capitai, Caudium, was a pass called the Caudine Forks, two narrow wooded defiles witb a grassy plain, wider but stili surrounded by steep mountains, between thcm. The Samnite generai Gavius Pontius hid his troops in the area and blocked che further defile with a defended barricade of rocks and felled trees.

When tbe advancing Romaos reached tbe barricade they made no anempt to storm it but retraced their steps in haste, only to discover tbat tbc defile by which they had entered had been barricaded and garrisoned behind them.

Completely cut orr, tbe Romaos panicked and at first had no idea what to do; eventually they fortified a camp in the plain near a spring. Over the next few days they made severa! attempts to break out, but with no success, and as supplies were running out tbey were obliged 10 surrender. The Samnites inflicted a harsh peace, obliging the whole Roman army co pass "under tbc yoke" as a sign ofhumiliation.

(Livy IX. 1-6; Salmon pp. 224-228.)

THE HELLESPONT 321 BC In the M acedonian civil war berween Perdikkas and Antipater, Antipater's generai Krateros crossed to Asia and was met near the Hellespont by Eumenes of Kardia, the governor of Kappadokia. Krateros had 20,000 foot, mostly Macedonian, and 2,000 horse. Eumenes had the same number ofinfantry but tbey were of a mixture of nationalities and unlikely 10 stand up to the Macedonians, so he relied on his 5,000 or more cavalry, many ofthem levied from his own satrapy. Both sides deployed cavalry on the flanks oftheir infantry. Eumenes was on the right with his own guard. On the left be had lranian, Greek and Thracian horse, opposite Krateros himself, since he was afraid his Macedonians might desert. Though outnumbered, Krateros' horse on both wings charged ahead ofthe infantry.

Krateros himself fell from bis horse, either because it stumbled or because he was wounded, and died. His men, discouraged and probably outflanked, were routed. Krateros' generai on the other wing was Neoptolemos, a persona! enemy of Eumenes. The two commanders sought each other out, fighting first with swords and then 67

grappliog. They pulled each other offtheir horses and fought on the ground; Eumenes was wounded but killed Neoptolemos, whose cavalry thereupon fled to the phalanx, on whom they rallied. Krateros' phalanx had not contacted Eumenes' infaorry, wbo were beld back. They were now outflanked by Eumenes' victorious cavalry on both sides, and surrendered when summoned.

(Diodoros XVIll. 30-32; Plutarch, E11menes.)

PISIDlA 320 BC Antigonos One-eye force-marched with 40,000 foot, 7,000 horse and up to 70 elephanrs, 40 miles a day for seven days, to surprise Alketas, one ofthe now-dead Perdikkas' last supponers, who was in Pisidia with 16,000 infantry and 900 cavalry, including 6,000 Pisidians. Antigonos' men seized a commanding position on some rugged hills.

Alketas rashly attacked uphill with his cavalry trying to turn the enemy off and give his troops time to deploy. Not surprisingly he was beaten and fled back to his infantry. Antigonos now attacked these in turn, with infantry and elepbants charging downhill and cavalry anacking from che flanks. He completely routed Alketas' force, most of which surrendered and was recruited into Antigonos' army. Only Alketas' bodyguard and che Pisidians remained loyal; they retired co Termessos, wbere, though the youog soldiers were devoted to Alkeras, their elders bad biro murdered to save the town from Antigonos.

(Diodo ros XVIII. 44-46)

PARAITAKENE 317 BC Eumenes ofKardia was now allied with most ofthe eastern satraps of Alexander's empire against tbe rising power of Antigonos. After some months manoeuvring in western Iran, Eumenes risked a pitched battle in Media. He bad about 6,300 cavalry, a phalanx of 17,000 including the 3,000 elice Argyraspides, 18,000 light infantry including 10,000 Persian archers and slingers, and 125 elephants provided by Eudamos, governar of India. Antigonos' numbers are less clear because Diodoros' overall fìgures do not match che tocals ofhis unit strengths, but clearly Anùgonos was superior in cavalry and phalanx, and outnumbered in elephants (and pr obably light infantry, for whom there is no figure); he had 8,500-11,100 cavalry, 28,000 phalanx and 64 elephants. "Phalanx" figures probably include some peltast types.

Antigonos was initially deployed on a range ofhills, and Eumenes' left also rested on hill:s, but the main battle was fought on a plain. Eumenes had his best 2,900 cavalry, mostly heavy, on the righe undler his own command; he himselfwas ili, aod was uansported in a litter. To the left ofthe phalanx was Eudamos with 3,400 cavalry, mostly light. The elephants and light infanrry were spread evenly along che front. Clearly Eumenes intended to win with his strong righe wing and refuse his left. Antigonos was able to see aod react to this from bis hilltop positioo, and be deployed Median, Parthyaian and Tarantine light cavalry on bis left under Pithon of Media, ordered to avoid frontal combat with the strong enemy righe. He led the right-wiog cavalry himself; one oftheir unit commanders was his 17-year-old son Demetrios, later called " the Besieger", in his fìrst battle. Most of Antigonos' elephants were on this flank, and be advanced down imo the plain obliquely, with the right leading.

Nonetheless Pithoo's lighc cavalry made the firsc comacc, riding round Eumenes' right flank and shoocing ac che elephants. Unable at fìrsc to catch chem, Eumenes summoned some lighc horse from che apposite wing, and with these and che light infantry chased offthe skirmisbers, roucing Pichon's whole wing as che elephants followed up, and pursuing ic back to che hills. M eanwhile che phalanxes mec, and despite cheir inferior numbers Eumenes' men were victorious, thanks chiefly to the vetera n Argyraspides.

Things looked black for Antigonos, but he snatched a chance from apparent defeat; as Eumenes' cenere pursued a gap opened becween them and Eudamos' left wing. Antigonos charged imo this, hitùng Eudamos' units on their ioside flank and routing them. At chis Eumenes recalled his own pursuers, wbose flank was oow threatened, and this enabled Aotigonos to rally many of his men; both sides reformed their lines but it was too dark for further fighting.

Antigonos claimed victory because he camped on the battlefield, but he had lost more men - 3,700 fooc and 54 horse killed, over 4,000 wounded, to Eumenes' 540 foot and "very few" horse killed and 900 wounded. In effect neither side had won, and both retired in search of su pplies aod wimer quaners.

(Diodoros XIX. 27-32; Scullard ( 1974) pp. 85-90) 68

GABIENE 316 BC Antigonos emerged early from tbc winter quaners wbere be had retired after Paraicakene, and force-marcbed against Eumenes' quarters, in tbc district of Gabiene near Susa, hoping to surp rise him. He caugbt Eumenes' elephants isolated and anacked tbem witb light cavalry and psiloi, but Eumenes relieved them before they had sufTered severe losses. With his forces reunited Eumenes prepared for another banle, which was fought on a broad sandy pia in in January 3 I 6.

Antigonos had 22,000 infantry (t his probably excludes psi/01), 9,000 cavalry and 65 clcphants. He put his best cavalry on thc right, nominally under Demetrios though in fact he commanded rhis wing himself, while Pithon again had 1he left wing borse. The phalanx was in the centrc and thc elephants and psiloi distributed across tbc whole front. Eumenes this time had 36,700 foot, 6,000 borse and 114 elephants. Learning Antigonos was relying on his righe wing, Eumenes massed bis best cavalry apposite; these included his own guards, on the extreme left, and thc bodyguards of most of the allied sauaps, witb 60 elepbants and supporting psi/oi in front of them. Pbilip, in charge ofthe right wing, had the weaker ofbotb cavalry and elepbants and was instructed 10 avoid battle. Unlike Paraitakene, bere both sides' strong wings tbus faced eacb otber.

Tbe elephants and cavalry joined battle first, and under the cover ofthe dust clouds they ra ised Antigonos sent some Median and Tarantine light horse unnoticed round Eumenes' Oank, to capture his baggage. Antigonos then charged, routing Peukestes, satrap of Persis, with over 1,500 borse. This isolated Eumenes himselfon the left; he tried 10 cut his way towards Antigonos, but tben bis lead elephant was killed, discouraging the other beasts.

Scenting defeat at this, Eumenes abandoned tbc losing wing and rode to join Philip on the right. He tried to rally the fugitives, sending messcngcrs to Peukestes, but tbat satrap only witbdrew funher, though apparently in reasonably good order. Meanwhile Eumenes' infanrry had again, thanks to the Argyraspides, beaten their opponents and pursued thcm. But Antigonos sent Pithon's cavalry to harass these, while he himself confronted Eumenes.

Eumencs' phalanx, without cavalry support, formed square and retreated 10 Peu kestcs' position. Rather than renew the battle, the Argyraspides negotiated with Antigonos, and in exchange for the return ofthei r baggage they seized Eu menes and handed him over. (One of tbeir commanders, Antigencs, had been involved in the similar betrayal of Perdikkas to Ptolemy five years before.)

Antigonos, now dominating the eastern satrapies, executed Eumenes; understandably mistrusting the

Argyraspides, he had them posted 10 the centrai Asian frontier, whence they never returned.

(Diodoros XIX.39-43; Plutarch, Eumenes; Scullard (I 974) pp. 90-94)

GAZA312 BC Ptolemy I, assisted by the fugitive Seleucus, invaded Antigonos' province ofSyria with an army of 18,000 infantry and 4,000 cavalry, mixed Macedonians, mercenaries and Egyptians. He was opposed by Antigonos' son Demetrios with 4,400 borse, a phalanx of 11,000, about 2,000 light infantry and 43 elephants. Demetrios planned to attack with a strong left wing under his own command, witb 1,300 Macedonian heavy cavalry, 100 Tarantine light cavalry and 1,500 other mixed horsemen, preceded by 30 elepbants and 1,500 light infantry. The other 13 elephants with supporting light infantry covered the phalanx, wbile the right wing of 1,500 cavalry was held back at an angle ordered to refuse battle. Seeing this, Ptolemy massed bis 3,000 best cavalry apposite Demetrios. In front ofthese were arcbers, javelinmen and men carrying some arrangement of iron spikes linked by chains, which had been prepared against the elephants.

T bc advance guard ofDemetrios' cavalry was inicially successful, but Ptolemy's cavalry outflanked them and threw them back. Meanwhile the elephants attacked, but were panicked by the spiked obstacles and the missiles; most of the drive.rs were shot and their beasts captured. At this reverse, Demetrios' cavalry fled, and the rest ofthe army followed theirexample. Deme1rios only lost 500 killed, mostly cavalry, but 8,000 captured; the high figure suggests tbc infantry surrendered wholesale when the cavalry left. The prisoners were taken into Ptolemy's service.

(Diodoros XIX.80-85; Scullard ( 1974) pp. 95-97)

TUN1S310BC Agatbokles of Syracuse had invaded Africa 10 draw ofT Carthaginian pressure from Sicily. Somewhcrc between Tunis and Canbage his 13,500 men were faced by a superior Carthaginian army, aUegedly 40,000 foot, 1,000 borse

69

and 2,000 chariots. Tbc Cartbaginians dcployed with the Sacred Band undcr Hanno on tbc rigbt; tbc lcft, under Bomilcar, was drawn up in greath depth because the tcrrain rcstricted the superior Punic numbers. Cavalry and chariots wcre in front. Agathokles faced this with a main line of 12,000 infantry, plus 500 arcbers and slingers divided betwcen the two wings, and 1,000 pickcd hoplites under his own command in front ofthe lcft, facing the Sacred Band. Sailors or camp-followers, unarmed but carrying shield-covers stretchcd over sticks 10 fool the enemy, wcrc used to boost the numbcrs.

The Punic cavalry and chariots charged; some wcre shot down, othcrs allowed to pass througb gaps in tbc Grcek line, bui most werc turned back against their own infantry. The Carthaginian Sacred Band now attacked, and fought on dcspite tbe immediate death ofHanno; but the rest ofthe Carthaginian force fled, probably bccause the Greek infantry attacked while tbey wcrc disordered by the rctreating mounted troops. The Sacred Band finally brokc when the rese oftheir force bad taken flight.

(Diodoros XX. IO)

RIVER THATES 310 BC Satyros, claiming the Bosporan throne, fielded 2,000 G reek and 2,000 Thracian mcrcenaries, and an allied Scythian force of over 20,000 foot and 10,000 horse. He deployed the merccnaries on the righe, and pickcd cavalry under his own command in thc cenere. H is opponent Eumelos was allied with King Aripharnes ofthe Sarmatian Siracae, wi th 20,000 borse and 22,000 foot. Satyros' picked cavalry cbarged and broke Aripbarnes and tbe Siracian cenere. The Siracian left defeated Satyros' righe, routing the mcrccnaries, but Satyros rcturned from pursuit in time to secure the victory.

He pursued the defeated Siracae up thc river Thates to the "city'', probably a hill-fort, whicb was tbcir capitai, suffering on the way from Siracian archers, shooting from the cover of woods. Attempting to storm tbc town, Satyros was wounded with a javelin and bis mercenaries repulsed by the defenders' missiles. Satyros died from his wound, and Eumelos became king.

(Diodorus XX.22-23)

IPSOS 301 BC The other Macedonian rulers combincd against the dangerously powerful Antigonos, and an allied army under Seleucus and Lysimachos met him at lpsos in Phrygia. Amigonos had 70,000 infantry, 10,000 cavalry and 75 elephants, to tbc allies 64,000 infantry, 15,000 cavalry, 100 scythed chariots and 400 elephams. Demetrios led a strong right wing ofcavalry which opened tbc battle by routing Seleucus' son Antiochos on the aUied left (unless thc fligbt ofAntiocbos' cavalry was a deliberate ruse), and pursuing rum dangerously far offthe field. T his exposed the flank of Antigonos' pbalanx, whicb Seleucus barassed witb borse-archers. Thc fight between Antigonos' left and tbc allicd right, perbaps under Lysimachos, sccms to bave becn indecisive.

When Demctrios tried to return he found his way blocked by Sclcucus' large elephant corps, and be was unable to force a way past these to help his father's beleaguered phalanx. Surrounded and under constant lire, some of Antigonos' infant ry surrendered; the rest were at last cbarged and rou1ed, Antigonos dying amongst them, hoping to the last that Demetrios would return to save him.

(Plutarcb, Demetrios 28-29; Diodoros XX. 113-XXI.2; Bar-Kochva (1976) ch.6)

SENTINUM 295 BC Tbc coalition of Etruscans, Umbri, Samnites and Senonian Gauls assembled a large army near Sentinum in Umbria, under tbc Samnite Gcllius Egnatius. Tbc Roman consul Fabius Rullianus marcbed against them, bui decided thc combined encmy were too strong for him, an impression reinforced when the Gauls ambushed one of his legiones. Tbe other consul Publius Decius Mus (son of thc Decius killed at Suessa) was sent 10 help him.

Together tbey mustered four legiones, 1,000 Campanian cavalry additional 10 the usual cavalry complement, and a larger number of allies - perhaps 45,000 men. The Et ruscans and Umbri were drawn offby Roman rescrvc forces raiding imo Etruria, leaving thc Samnites and Gauls to face tbe Romans. These two alone proved a close match for 1he Roman army, so were probably similar in numbers. Livy was sure that tbe presence ofthe Etruscans and Umbri would bave turncd the battle.

70

The Gauls were on the allied righe, with cavalry on their flank and chariots in reserve, the Samnites on the left.

Fabius faced the Samnites with legiones I and III, Decius the Gauls with V and VI. The imperuous Decius attacked the Gauls at once, leading the cavalry charge himself. He broke the Gallic cavalry, but as his men pursued, crashing into some ofthe Gallic infantry, they were counterattacked and routed by the chariots. These now pursued in rurn, following the routed Roman cavalry imo the ranks ofthe Roman infanrry. Seeing the legionaries thus disordered, the Gallic infantry charged and pushed the Romans back. Decius cried to rally his fleeing uoops, but met with no success and was killed himself in a desperate charge which Roman tradition claimed was a deliberate dewcio, a religious sacrifice whcreby Dccius dedicated tbc enemy to the gods below.

On the other wing, Livy says that Fabius deliberately held back from a decisive attack unti! the Samnites tired and ran out ofmissiles, but the truth may be that a long, gruelling infantry struggle was necessary before any impression could be made on them. Fabius' cavalry presumably chased olf what Samnite cavalry there was and retumed to contro! while the infantry fought; certainly he eventually sene them in against the Samnite infantry flank.

Combined with the infantry pressure, this broke the Samnites and they fled past the Gallic Line to the allied camp.

Once he knew Decius was in difficuhies, Fabius had sent some ofhis reserves to his colleague's assistance; now learning ofhis death, Fabius despatched 500 Campanian cavalry to take the Gauls in the rear, with the principes of the third legion to support them. Many ofthe Gauls had probably scattered in pursuit; those that rema ined were engaged with survivors of Decius' wing and Fabius' first batch of reinforcements. This new force broke up the Gallic infantry formation, but some resistance continued. Meanwhile Fabius had stormed the Samnite camp, where Gellius Egnatius was killed trying to organise the defence. Fabius then took che Gauls in the rear, completing the rout.

Livy says allied losses were 25,000 killed and 8,000 taken; the Romans lost 7,000 from Decius' army and 1,700 of Fabius'. The Samnites lost even more as they retreated through the territory ofRome' s Paelignian allies.

(Livy X.25-29; Salmon pp 265-267)

HERACLEA 280 BC Pyrrhos of Epeiros, in ltaly at the request of Taras, faced an invasion by the Roman consul Publius Valerius Laevinus. Pyrrhos had his expeditionary force of28,500 men and 20 elephants, less storm losses and garrisons, plus the reorganised Tarantine militia, perhaps 35,000 men in total. The Roman army was larger, perhaps four legions plus allies, 40,000. The rwo armies camped on opposite sides ofthe river Siris near Heraclea. Laevinus decided to attack at once before Pyrrhos could be joined by his Samnite and Lucanian allies. He sene his infantry across a ford, but his cavalry splashed through the river at severa! points, threatening to surround the small guard force Pyrrhos had posted on the bank. The guards thus withdrew.

Pyrrhos charged the Romans with 3,000 cavalry, hoping to catch them before they had completed the crossing, and to give his own infamry time to form up. However his horse was killed under him by an Italian cavalry officer and the Epeirot cavaJry repulsed. Pyrrhos thus brought up his infantry, and a prolonged struggle between legions and phalanx ensued. Pyrrhos himselfhad exchanged his distinctive cloak and armour for those of a companion, Iest any other ambitious Roman should seek him out; but the ruse backfired when the substitute was killed, and Pyrrhos had to ride bareheaded along hjs line to encourage his men and persuade them their king was stili with them! The advantage in the battle was said to ha ve changed hands seven times before a decisive success was scored by Pyrrhos' elephants, who broke through the Roman cavalry. As the Roman line wavered Pyrrhos charged in his own Thessalian cavalry to complete the rout. The Romans lost 7-15,000 men, Pyrrhos 4-13,000, the first "Pyrrhic victory".

(Plutarch, Pyrrhos 16-17; Garoufalias pp 69-76) THERMOPYLAI AND DELPm 279 BC The Galatians had invaded Greece, plundering the northern states, under Brennos and Akichorios. In tbe harsh winter of280-279 the centrai Greek states blocked the pass ofThermopylai with an army of24,000 foot and 1,500 horse, chiefly Boiotians, Aitolians and Phokians, under the Athenian Kallippos. Athenian ships, perhaps financed by tbe Macedonian governor in Athens as they do not seem to bave constituted an officiai Atbenian fleet, stood offshore.

The Greeks tried to hold the river Spercheios beyond the pass, sending cavalry and light infantry to break down the bridges and hold the bank. Brennos however sent 10,000 men downstream to wade across the river where it 71

broadened into a shallow marsh. Outflanked, the Greek advance force fell back to the main army, and Brennos conscripted locai labour to build a bridge for bis main body. He then tried a fronral anack on the Greek position in the pass. The Galatians fought with furious courage, but suffered heavily from missiles, both from Greek light infantry on the inland slopes of the pass and from Greek ships offshore. Unable to make any impression, the Galatians withdrew.

Brennos now planned to weaken the defenders by drawing off the Aitolians. He sent 40,000 infantry and 800 cavalry to pillage Aitolian territory; they brutally sacked the !inie town ofKallion, and this succeeded in gening the Aitolians at Thermopylai to march to the defence oftheir country. They were joined by the home levy, even women and old men coming out to fight, and by a few Achaian hoplites. These were badly mauled by the Galatians, but the Aitolians avoided dose combat and harassed the Celts with javelins. In tbc Aitolian mountains these tactics were very effective, and less than halfofthe Galatians rejoined the main force.

Meanwhile at Thermopylai Brennos outflanked the Greek position with 40,000 men. lt was said he used the same path as the Persians two centuries earlier, and similarly surprised the Pbokian guards, charging out ofthe mist. This time, however, messengers warned the main Greek army before it was encircled, and the Greeks withdrew and dispersed to their homelands.

Brennos' flanking column now headed for tbc treasure-filJed sanctuary of Apollo at Delphi in Phokis, Akichorios folJowing with thc main force and the baggage. Brennos was harassed on the way by Phokian and Aitolian light infantry, later joined by a few Lokrian hoplites, while the bulk ofthe Aitolian army similarly harassed Akichorios.

Again unable to reply 10 tbc skirmishers the Celts suffered heavily, and were further discouraged by the cold weather and by earth tremors and thundersiorms, attributed to Apollo's displeasure. As Brennos' men camped at night, rockfalls killcd many more. Tbc ncxt day tbey were attacked frontally by the Lokrians and Aitolians, and outflankcd by Phokians sbooùng from the slopes. Brennos was wounded, bis men retreated, and during the next night this developcd into panie fligbt. The survivors rejoined Akichorios and the main force. Akichorios now leamed that the full Boiotian and Athenian armies were in the field again, so decided to withdraw. The Aitolians harassed bis retreat as far as tbc Spercheios, and then the Thessalians and Malians turned on the invaders. The survivors retreated out ofGreece completely.

(Pausanias X.19-23; Tarn (1913) eh. VI) ASC ULUM 279 BC Pyrrhos fought the Roman consul Publius Dentius Mus near Asculum in Apulia. Both armies were about the same size, either 40,000 men or (less probably) 70,000 foot and 8,000 borse. Pyrrhos, now joined by bis Oscan allies, had slightly more cavalry and 19 elephants. Dentius Mus had four legions, 20,000 Roman citizens plus allies, with a slight infantry advantage. He also had 300 special anù-elephant wagons, hung about with spiked beams and inflammable weapons.

The first day's fighting was among woods and hills which hindered Pyrrhos' cavalry and elephants; it rnay bave been forthis baule that he mixed ltalian units between the phalanx speiraito givc greatcr flexibility. Ifthe incidents described by Dionysios belong bere, Pyrrbos' Macedonians broke the Roman first legion and Latin allies, on their left wing, but the Roman third and fourth legions beat the Tarantines, Oscans and Epeirots of Pyrrhos' centre.

Meanwhilc a force ofDauni attacked bis camp. He sent reserve cavalry to deal wi th the breakthrough, more cavalry and elephants to drive offthe Dauni. When they withdrew to an inaccessible stcep bili he switched the elephants against tbc third and founh legions; tbese 100 took refuge on wooded heights, only 10 be shot at by tbc arcbers and slingcrs cscorting the elephants, to whom they could 001 reply. Pyrrhos then seni some Athamanian, Akarnanian and Samnite infantry to winkle the Romans out oftbe woods, but thcy were intcrceptcd by Roman cavalry. Botb sidcs withdrcw at dusk, ncithcr baving gained much advantage.

Ncxt dawn Pyrrhos sent light infantry to occupy the difficult ground which had given him trouble the previous day, and thus obliged tbc Romans to fight a set batùe in the open. As at Heraclea, an even figbt between legions and phalanx followcd, unti! tbc elepbants, supported by light infantry, broke througb tbc Roman line. lt was probably now that the anti-elephant wagons were driven against them; after a bricfsucccss thcy were overwhelmed by the light infantry. The Romans then retreated to their camp. Their losses were 6,000 men, Pyrrhos' 3,500, a narrow victory.

(Plutarch, Pyrrlzos 21; Dionysios XX. I-3, describing only one day's battle, mostly relevant to the first day; Garoufalias pp 88-93.)

72

BENEVENTUM 275 BC Pyrrbos anacked 1he Roman consul Manius Curius, who was encamped in a s1rong hill position near Beneventum in Samnium, operating against Pyrrhos' Samni1e allies. Reports tbat Pyrrbos had 80,000 infantry and 6,000 cavalry, and that be outnumbered the Romans tbree 10 one, are probably grossly exaggerated. Pyrrhos tried 10 outflank tbe Roman position by a night marcb witb picked troops and the best elephants. But the route led along narrow forest tracks and the marchers lost tbeir way wben tbeir torches went out. They evenrually arrived uphill of the Roman position at dawn, but were exhausted and too late to surprise the Romans. Manius Curius led his fresh troops out to anack this tired farce and defeated them, capturing severa! elepbants. h may have been now that che Romans released squealing pigs to frighten the elepbants.

Manius now followed up into the plain below against Pyrrbos' main arrny. The Romans were successful on one wing, bue were beaten on the other by the elepbants. Manius threw in his camp-guard who charged downhill at tbe pursuing elephants, possibly using incendiary javelins. One young elephant was wounded, and its panie proved infectious. The elephants were forced back into their own lines, trampling Pyrrhos' men and completing tbe rout.

Pyrrbos was sa id to have lost 23-33,000 men and two elepbams killed, 1,300 men and eight elephants captured; che figures may be exaggerated, but the defeat marked tbe end of Pyrrbos' halian campaign.

(Plutarch, Pyrrhos 25; Dionysios XX.10-12; Scullard (1974) pp 111-115; Garoufalias pp 118-121.)

THE "ELEPHANT VICTORY" 273 BC The Seleucid king Antiochos I marched against the Galatians, wbo had been devastating Asia Minor. He is said to have been heavily outnumbered, and over balfhis army was ligbt infantry. However he had 16 elephants, sent via Babylon by his governor ofBactria. Tbe Galatian centre was a deep forrnation ofinfantry, tbe front ranks allegedly armoured, with 20,000 cavalry on tbc flanks. 80 scythed chariots and 160 two-horse chariots were behind the line.

Since only 20,000 Galatian fighting men had crossed to Asia, however, the figure of20,000 for cavalry alone is clearly exaggerated; Antiocbos may even bave outnumbered tbe Celts. But his farce was clearly not ofvery higb quality, and the Galatians were stili reckoned virtually invincible, so his reponed lack of confidence was understandable.

On the advice of the tactician Theodotos of Rhodes, Antiocbos somehow concealed his elepbants. Wben the Galatian cavalry charged and the infantry opened ranks to lei 1he chariots through, he sent four elephants against che cavalry on each wing, and the remaining eight against the chariots. Unused 10 elephants, the Galatian borses bolted through their own infantry, many of wbom were trampled or cut down by the chariot scythes. The elepbants followed up into the Galatian infantry, completing the rout.

(Lucian, Zeuxis or Antiocl1os, 8-11; Bar-Kochva (1973) and (1976) p240 n85 for revised dating.)

BAGRADAS 255 BC The Roman consul Marcus Atilius Regulus bad invaded Africa with J 5,000 foot and 500 cavalry. Tbe Canhaginian citizen levy was supplemented by newly recruited Greek mercenaries, including an experienced Spanan officer, Xanthippos. H e was loudly criticai ofthe gene rais ' performance to date, and at the insistence ofthe mercenaries the Cartbaginian leaders agreed to obey bim. When Regulus advanced across tbe flat Bagradas plain towards Cartbage, Xanthippos moved to meet him, with 4,000 cavalry and 100 elepbants, tbough only 12,000 infantry.

He deployed the Canhaginian citizens in tbe centre, with mercenaries on thc right. Tbe elepbants were in front of the citizen infantry, with cavalry and ligbt infantry in front ofthe wings. Regulus split his few cavalry between bis wings and, in the hope of stopping the elephants' charge, drew up his infantry in unusual depth, with skirmisbers in front. The Carthaginian cavalry soon beat tbeir Roman counterparts, bue thc left wing Roman infantry, avoiding the elephants, charged and broke tbe mercenaries on tbc Canhaginian right, chasing tbem as far as tbc camp. (Silius ltalicus claims this was a feigned fligbt, nota rout; he is unlikely to bave invented a story whicb discredits tbe only Roman success in the battle, so the tale may be true, but it is difficult to reconcile with tbe casualry figures.) Tbe elephants crashed into the front ofthe remaining Roman infantry, who soon found their their deep formation only allowed them to be trampled more easily, and were soon also anacked in flank and rear by cavalry. Those who managed 10 cut their way past che elepbants were slaughtered by che Canhaginian infantry.

2,000 infantry from tbe Roman left, who had been pursuing the mercenaries, got away; 500, including Regulus, were captured; tbe rest were killed where they stood or hunted down by cavalry as tbey !led across tbe plain. The

73

Carthaginians lost 800 mercenaries, but few other casualties. Xanthippos left Carthaginian service, possibly for that of Egypt.

(Polybios 1.32-34; Silius V.326-328.)

PANORMOS 250 BC T he Cartbaginian generai Hasdrubal son of Hanno, in Sicily with a large army and 140 elephants, advanced to destroy the com around Panormos before the Romans could harvest it. T he city was held by the Roman generai Lucius Caecilius Metellus, wbo at first kept bis men inside the town, giving tbe impression that he was afraid to face the elephants, as ali Roman generals had been since Bagradas. When Hasdrubal advanced dose to the walls, however, Caecilius sent some of his skirmishers out to harass him, steadily reinforcing them until the whole Carthaginian army was deployed to fend the Roman skirmishers off. Meanwhile Caecilius ordered the civilian population ofthe town to bring supplies ofmissiles to tbe walls, and drew up his heavy infantry by agate opposite the Carthaginian Jeft flank.

Hasdrubal's elephants charged, apparently without being ordered to do so, the Roman skirmishers, but this advance brought them within the range ofarchers on the city wa!ls and more javelinmen drawn up in front ofthe city ditch. These men were ordered to take refuge in the ditch itselfif charged, but this does not seem to bave been necessary; the hail of missiles wounded many ofthe elephants and they fled in panie, breaking through tbe lines of their own troops behind them and trampling many. Caecilius now sallied out with his fresh heavy infantry, hitting the disordered Carthaginians in the left flank and routing them. The Romans captured ten elephants in the battle, and rounded up most ofthe rese, driverless, afterwards. Caecilius led 120 in his triumpha! parade.

(Polybios I. 40; Scullard (1974) p 151.)

TELAMON 225 BC Alarmed by Roman expansion into Cisalpine Gaul, the Gallic lnsubres, Boii and T aurisci allied and called in the mercenary Gaesatai from the Rhòne, to invade Italy. The combined army numbered 50, 000 foot and 20,000 horse and chariots; some troops had to be left behind because the Gallic Cenomani, with the Veneti, Umbri and Sarsinates, were allied with Rome and threatened to invade Boian territory. In addition to this force the Romans raised three armies every one ofwhich was dose to the Gallic host in numbers. The consul Lucius Aemilius Papus was at Ariminum on the Adriatic, where the Gauls were expected to attack, with four legiones plus allies - 50,800 infantry and 3,200 cavalry. A praetor defended Etruria with 50,000 Etruscan and Sabine foot and 4,000 borse (of whom the Sabines at least were Roman citizens, though this force does not seem to have been counted among the normai legio11es), and there was a reserve force near Rome itself of equivalent strengtlii to a consular army. The second consul, Gaius Atilius Regulus, was campaigning in Sardinia with anotber four legiones and allies.

Instead of attacking Ariminum the Gauls bypassed the praetor and devastated Etruria. When the praetor's army carne up behind thero they succeeded in attacking him by surprise, routing the army and killing 6,000 men. The survivors rallied on a hill which the Gauls blockaded overnight with their cavalry.

At this point Aemilius, who had heard ofthe invasion ofEtruria, appeared in the Gallic rear. Seeing his campftres they resolved DOt to fight and slipped away DOrthwards before dawn. Aemilius collected the praetor's survivors and followed. Wben they reached TelamoD, however, the Gauls were surprised to encounter yet another Roroan arroy - tbe consul Gaius Atilius, who had rerurned froro Sardinia and landed at Pisa. Atilius advanced with bis infantry in line ofbattJe, rushing ahead in command ofhis cavalry to seize a key hill overlooking the Gallic line ofmarch. The Gauls sent their own cavalry and light troops to dispute it, and when Aemilius carne up he threw in his own borse to support his colleague. The Gauls, finding themselves trapped between two Roman armies, drew up their infantry in two lines, with the Gaesatai and Insubres facing Aemilius, and behind them tbe Boii and Taurisci facing Atilius. The chariots were OD the wings, their booty, with a force to guard it, on a nearby hilJ. lf Polybios is right in saying that each consuJ had four legiones instead oftbe usual two, the combined armics numbered over l 00,000, plus the survivors ofthe praetor's force now with Aemilius, so the Gauls were outnumbered nearly two to one.

After a hard cavalry fight in which Atilius was killed, the Roman cavalry succeeded in taJdng the bili. The Romans sent forward their skirmishers to harass the Gallic infantry with javelins, to which the Gauls couJd not reply. Some ofthe Gaesatai charged out in an attempt to catch the skirmishers, but without success. When the enemy were worn down the Romans withdrew their skirmishers and sene in the heavy infantry; surrounded, the Gauls held out fighting fiercely unti! the Roman cavalry charged down and completed their defeat. Most ofthe Gallic infantry were cut down where tbey stood, but many ofthe surviving cavalry got away.

74

The Gauls lost 40,000 killed and 10,000 captured, including one oftheir kings, while another committed suicide.

(Polybios II.22-31.)

SELLASIA 222 BC Antigonos Doson of Macedon invaded the Peloponnese to support tbe Acbaian Leaguc against Kleomenes III of Sparta. Amigonos had 13,000 Macedonian infantry, 15,000 mostly light-armed allied, mercenary and Acbaian infantry, and 1,200 cavalry. Kleomenes had only 20,000, including 6,000 Spartans armed as Macedonian phalangites, so he took up a defensive position, near Sellasia where the road to Spana ran along the river Oinous between two hills. He fortified both hills with ditches and palisades, and positioned himself on Olympos, tbe righthand bili, with rhe Spartans and some ofthe mercenaries; cavalry and mercenary light infantry in tbc centra! valley; and allies and perioikoi undcr bis brotber Eukleidas on tbe otber bili, Euas. Antigonos tricd for severa! days to Iure him out ofrhis strong position, wirb no success.

Finally Antigonos decided to attack Euas with the Macedonian Bronze Shiclds, with Illyrian infantry in between the speirai ofthe phalanx, to give it the flexibility to cope with thc slope. Behind these were Cretan and Akarnanian light infantry, the Cretan archcrs no doubt shooting up at the bili over the heads ofthe assault force. Cavalry and Achaian infantry were in the valley, while Anrigonos himself was to attack Olympos with Macedonians and mercenarics.

As the Illyrians advanced, their exposed left flank was charged by Kleomenes' light mercenaries from che valley, but these were coumer-attacked and driven offby Achaian cavalry undcr the young Philopoimen. A generai cavalry melée developed, with thc Maccdonians gaining tbe upper hand. Relieved of this pressure, the Illyrians and Macedonians attacked Euas frontally while tbe Akarnanians worked round thc flank. Eukleidas rcceived tbem at the halt, instead of countcr-charging, so tbat once his men began to be pushed back they lost tbe advantage ofbeight to the enemy and were soon brokcn.

On Olympos the two kings at first skirmisbed witb mercenaries, but when Kleomcncs saw his left in rout and his cavalry losing, he tried tosa ve tbe day by a downbill phalanx chargc. This at first pushed tbe Macedonians back but they scem to bave countcr-attacked in closer formation while troops from thcir successful rigbt and centre rook tbc Spartans in flank and broke tbem. Kleomenes' losses were very heavy, including ali bue 200 oftbe 6,000 Spartans engaged. He himself fled with a few horsemen and sailed ro Egypt, Ptolemaic subsidies having financed his campaigns, where he died trying to stir up another revolution.

(Polybios 11.65·69; Pluiarch, Kleomenes 28, Philopoimen 6.)

AP OLLONIA 220 BC Molon, Selcucid satrap of Media, rebelled and overran most of the eastern provinces. The young Antiochos III, later called "the Great", marched cast threatening to cut Molon offfrom bis Median base. Molon fell back from Babylonia into the hill country east ofthe Tigris, and near Apollonia both sides' light infanrry scours met while crossing a range ofhills. After some skirmisbing, tbey separarcd and both armies camped for tbc night. Molon planned a night attack, and scnt a picked force to high ground above the royal camp. But reo men deserred ro rhe king, and Molon, fearing his plans were berrayed, hurriedly fell back to his own camp. His unexpected return surpriscd and panicked tbc troops lcft in tbc camp, wbo feared at first tbat tbc enemy were attacking. So tbe rebel army was already tired and dispirired wben be led it out to battlc.

Antiochos dcployed heavy cavalry on his right, then Cretans, Galatian rhigosages and Greek mercenaries, with the phalanx in the centre and Companions on tbc left. His ten elcphants wcre in front oftbc Line, witb infantry and cavalry reserves behind botb wings, inrcnded ro advance and outflank tbc cnemy. Molon put rhureophoroi, Galatians and othcr infantry in tbe cenere - be sccms to havc had few if any acrual phalangires - with cavalry on the wings and lighr infantry, notably numerous Kynii slingers, beyond thesc, probably on bills. He bad scythed chariots spaced out in front ofthe line. Numbers are unknown.

Molon's righe wing, under his own command, attacked vigorously, but Antiochos' reserves seem ro bave succeeded in enveloping both rebel wings. Tbc rebel lefr, surrounded and faccd by che rightful king, surrendered. Sceing bis army collapsing and himselfencircled, Molon committed suicide rathcr than be caprured.

(Polybios V.51 ·54; Bar-Kochva (1976) eh. 8.)

75

THE TREBIA 218 BC Hannibal had invaded north ltaly, won a cavalry battle on che Ticinus, and in cold, snowy December 218 was camped near Placentia. On che other bank of che river Trebia were che Roman consuls Tiberius Sempronius Longus and Publius Cornelius Sci pio, with four legiones, ltalian al Li es and Cenomanic Gauls. Sci pio was recovering from a wound and out ofaction; Sempronius was eager co fight, and Hannibal was aware ofthis.

He concealed l,000 Numidian cavalry and 1,000 infantry under bis brother Mago in an overgrown stream bed as an ambush, and at dawo sent the rest ofthe Numidians over the river to harass the Roman camp. Meanwhile the rese of the army had a hot breakfast and rubbed themselves down with oil to keep out the cold. Sempronius reacted as expected; not giving bis men time to eat be led out first che cavalry and then the whole army to chase the Numidians across the freezing river. On the plain in front ofhis camp Hannibal deployed bis 8,000 Light infantry screening his 20,000 heavier infantry (with the Gauls, one-tbird or more ofche tota!, in the centre), and the 10,000 cavalry and 37 elephams split between the wings. Sempronius drew up bis 16,000 Roman and 20,000 ltalian infantry in the usual three lines, with che Ceoomani on che righe and che 4,000 cavalry on che flanks.

When che Light infantry clashed che Carthaginians had the better of it, as the Romans had used up most of their javelins on the Numidians; when the skirmishers withdrew through the heavy infantry, Hannibal's light troops switched to shooting at the Roman cavalry. Heavily outnumbered, these were soon chased offby che Carthaginian cavalry. When the heavy infantry clashed, the Roman flanks were attacked by che elephants and soon shot at by che light infantry and Numidians. The Roman light iofantry succeeded in chasing off the elephants, but Hannibal regrouped these and they charged and broke the Cenomani. Mago's ambush now bit the cold, hungry, outflaoked Romans in che rear. Most ofthe Roman wings broke and fled, many drowning as they tried to cross the Trebia.

10,000 infantry from che Roman cenere however succeeded in breaking through Hannibal's Gauls; seeing the rest oftheir army in flight they made their own way to Placemia.

Roman losses bave been estimated at 15,000. Hannibal's were slight, and were mostly expendable Gauls, though most ofthe elephants and many men and horses died of cold soon after.

(Polybios 111.70-74; Livy XXI.53-54; Lazenby pp 55-58; Caven pp 110-1 13.)

LAKE TRASIMENE 217 BC Hannibal was marching through Etruria towards Rome, devastating the country to provoke the Romans into battle.

He was closely pursued by the consul Caius Flaminius with 25,000 men, including two legions. Hannibal led the Romans along the shore ofLake Trasimene, aod during the night concealed his army in ambush, in a narrow defile between steep hills and the water. His Africa o and Spanish infantry were placed on a hill crest across the Roman lioe of advance, the light infantry in the hills on the flaok, and tbe cavalry and Gauls in position to cut offtbe Roman retreat.

At dawn Flaminius set out after che enemy, in thick mise, with no attempt at reconnaissance. Seeing the Africans and Spaniards, the Romans began to form line to attack them, only to be completely surprised by che rest of che Carchaginian army charging downhill out ofthe mist imo their flanks and rear. Unable to orgaoise any effective resistance, 15,000 Romans were killed, includiog the coosul, some drowoing in the lake as they tried co flee. A body of 6,000 Romaos cut cheir way out, only to be surrounded and captured the next day. Hannibal lost I ,500-2,500 men, mostly Gauls.

(Polybios 111.82-84; Livy XXII.4-7; Lazenby pp 61 -66.)

RAPHIA 217 BC Antiochos 111 overran Ptolemaic Palestine; Ptolemy IV re-organised the army and counter-anacked, meeting Antiochos on a flat plain near Raphia. Ptolemy had 70,000 foot ofwhom at least 56,000 were armed as phalangites (25,000 klero11cho1~ 20,000 Egyptians, plus Libyans and mercenaries), 5,000 borse and 73 African elephants.

Antiochos had 62,000 infamry of whom 32,000 were armed as phalangites (includiog I 0,000 argyraspides), 6,000 cavalry and 102 Indian elephants.

Both deployed with phalanx in che cenere, lighter infantry on che flanks, and cavalry on the wings, with elephants in front ofthese and ofthe flank infantry. The 60 elephants in Antiochos' righe easily beat the 40 smaller elephants on Ptolemy's left, forcing them back on the infantry behiod, and the Seleucids' Greek mercenaries completed che rout. Antiochos on the sa me wing then led 4,000 borse against Ptolemy's 3,000, broke them and chased them offthe

76

field, bue Peolemy himself escaped. On the other wing, 2,000 Ptolemaic mercenary cavalry outflanked and beat 2,000 Seleucid borse, whiJe Ptolcmy's Greek mercenaries beat the Arab and M edian levies. Each side's righe had thus beaten the other's left, but while Antiochos was offin pursuit, Ptolemy joined his phalanx to encourage it. The two cencres charged, and the largcr Ptolcmaic force was victorious almost immediately. Antiochos and his cavalry returned to find the rese of che army in flight.

Ptolemy lost 1,500 foot and 700 borse killed, Antiochos 10,000 foot and 300 borse, plus 4,000 prisoners. There are highly contradictory accounts ofelephant losses.

(Polybios V.63-65 and 79-86; Scullard (1974) pp 137-145; Bar-Kochva (1976) eh. IO; Galili (1976/7). For the view that Ptolemy's phalanx was only 25,000, including the Egyptians, see Mahaffy ( 1898) and Griffith ( 1933) pp 122-123.)

C ANNAE 216 BC Hannibal was in Apulia with 40,000 foot and 10,000 borse, opposed by the Romao coosuls Caius Terentius Varro and Lucius Aemilius Paullus, camped on the other bank ofthe river Aufidus with 80,000 iofaotry and 6,000 cavalry - eight legiones, understreogth io cavalry, with allies including a Syracusan force of archers and slingers. The consuls comroaoded on alternate days; when it was Varro's rum he led most ofthe army out, leaving 10,000 men in the larger ofthe two camps, with orders to attack Hannibal's camp, and probably 5,000 men in the smaller Roman camp. Varro drew up on a plain with the river on his right. On that flank were the Roman cavalry, in the centre the infantry, screened by skirmishers and with the heavy infantry drawn up io greater depth than usual to batter through Hannibal's centre, while the ltalian allicd cavalry were on the left. Hannibal crossed the river covered by his light troops, and deployed the Spanish and Gallic cavalry on his left under H asdrubal, the Numidian cavalry on his righe under Maharbal. In the centre were the Gallic and Spanish infantry in alternate units, arranged in a crescent formatioo with the cenere chrown forward. The centre units were dr awn up in greater depth chan che less exposed flanks, aod each flank ofthe crescent was supported by a body of Africao infantry, held back.

The battle opened with indecisive skirmishing by che light troops, while on ooe wiog the ltaUan cavalry tried unsuccessfully to come to grips with the skirmishing Numidians, and on che other tbc Gallic aod Spanish cavalry charged the Roman borse. Here, after a desperate fighe in che confined space between the river and the infantry lines, the outnumbered Romans finally broke and fled. The survivors were pursued along che river bank, but Hasdrubal soon regrouped aJI or most ofhis cavalry and swung round che rear ofthe Roman infantry to help the Numidians.

Meanwhile che centre of the Roman heavy infantry li ne had charged the projecting cenere of Hannibal's crescent formation, and gradually forced the Gauls and Spaniards back. Tbc convex Une wavered, maightened, and became concave as the Romans forced the centre back and thrcatcncd to break through. But this advaoce placed the so far unengagcd African foot on the Roman flanks, and, perhaps supported by che light iofantry, they charged in and threw the dense Roman formation into disorder.

Hasdrubal's cavalry had appeared in the rear ofthe ltalian allied borse; these broke and fled at once, and Hasdrubal left the swifter Numidians to pursue them, while he regrouped his heavier horsemen yet agaio, and charged into the rear ofthe beleaguered Romans. The dense Roman formation meant that many in the middle ofit could not reach the enemy, nullifying the effect ofthe Roman numerica! superiority; surrounded, tbc Romans were gradually cut to pieces. Paullus was among che dead, Varro one ofthe few survivors. The victorious Hannibal now also broke the Roman camp guard, who were io rhe course ofan attack on his camp. 47,500 Roman infantry and 2,700 cavalry were killed, 19,300 captured. Hannibal lost 4,000 Gauls, 1,500 Spaniards and Africans, and 200 cavalry. This was Hannibal's greacest victory, bue failed co break Roman resiscance.

(Polybios Ill.11 3¡ 117; Livy XXIl.41-50; Lazenby pp 75-85.)

IBERA 215 BC The brothers Gnaeus and Publius Cornelius Sci pio, with a Roman army of25,000 men in Spaio, were attacked near Ibera on che river Ebro by Hannibal's brother Hasdrubal wieh an army ofsimilar size. The Scipios deployed in the usual ebree lincs. Hasdrubal apparently cried to repeat his broeher's cactics, as he placed his Spanish infantry, expendable unenthusiastic tribal levies, io the cenere, flanked by better infantry, Africaos and mercenaries on the left, "Poeni" (probably citizen troops from the Punic towns ofSpain) on the right. Numidiao cavalry with a few elephaots were on the righe wing, ocher cavalry on the left.

77

When tbe Romans anacked, Hasdrubal's Spanish cenere broke in short order, and though the flanking infantry closed in the Romans successfully faced out and fought them ofT. The Carthaginian cavalry fled when they saw the cenere break, and swept the elephants away with them. Hasdrubal's casualties were heavy, the African and Poeni foot being almost wiped out.

(Livy XXIII.29; Lazenby p. 128.)

CASTULO AND ILORCI 212 BC The Sci pio brothers divided their forces to dea! witb dispersed Cartbaginian armies. Publius took cwo-thirds ofthe Romans and Italians against Hasdrubal son ofGisgo and Mago, Gnaeus took the remaining Romans and 20,000 Celtiberians against Hasdrubal son ofHamilcar, Hannibal's brother.

Near Castulo in the valley of the Baetis (che Guadalquivir) Publius was harassed by Numidian cavalry under Masinissa, who by cuning up foragers and outposts effectively confined the Romans to their camp. Hearing further reinforcemencs were on their way to tbe enemy, the Spanisb prince lndibilis with 7,500 Suessetani, Publius gambled on beating the allies separately. He slipped out ofthe camp by night, leaving a small garrison, and met lndibilis the next day. The resulting battle was confused and disorderly as neither marching column had time to deploy into line ofbattle, but the Romans were winning till Masinissa's Numidians aririved to harry both flanks, closely followed by the Carthaginian generals who charged the Roman rear. The Rom~ns broke when Publius was killed, and many fugitives were cut down by pursuing cavalry and light infantry tĂŹll nightfaU allowed some to escape.

Meanwhile the other Hasdrubal had bribed the Celtiberians to desert. Gnaeus tried to retreat, but Hasdrubal was joined by bis victorious colleagues and their Numidian cavalry harassed the Roman withdrawal till Gnaeus improvised a defensive position on a slight hill near Il orci, sening up a barricade of pack-saddles and baggage, since there was no timber on the hill for palisades and che soil was too stony to dig ditches. When the Canhaginian infanery carne up they were at first held up by che barricade, bue were able to prise it apan with poles, and then soon overran the outnumbered, dispirited Romans. Many Romans managed to escape through nearby woods. The survivors ofthe two Roman armies regrouped north ofthe Ebro, but their hold on Spaio was nearly broken.

(Livy XXV. 32-36; Lazenby pp. 130-131; Cavea pp. 181- 183.)

BAECULA 208 BC The Roman forces in Spain were now under the younger Publius Cornelius Sci pio (later known as "Africanus", son ofthe Publius killed at Castulo), who had taken New Carthage and won over severa! Spanish tribes. He marched against Hasdrubal son ofHamilcar, who took up a strong position near Baecula wich 25-30,000 men. His camp was on the leve! summit of a steep bili, surrounded by a steep bank; below this was anotber open, slightly sloping cerrace, surrounded by a similar bank. A river covered his rear. Hasdrubal occupied this lower terrace with Numidian cavalry, Balearic slingers and African light infanery, and refused to come out and fight in the open.

Scipio, wbo had perbaps 35-40,000 men, had to attack in case he was caught by another Carthaginian army operating nearby. He sene some of the velires and a picked force of heavy infantry, supported by camp servants throwing stones, up the steep slope to the lower terrace. These suffered heavily from missiles, but eventually gota footing on the open ground and quickly broke the Carthaginian light troops. Hasdrubal, wbo seems only now to bave realised that bis position was not impregnable, began to Jead bis main force out ofthe camp and deploy on the level billtop. Scipio sene his remaining light troops to support the force which had cleaired the terrace in a frontal attack on this plateau, wbile be and Caius Laelius each took halfthe remaining army to :skirt Hasdrubal's position and fall on his flanks while the Carthaginians were pinned by the froneal anack. These flank attacks were successful and the Carthaginians collapsed. Hasdrubal lost 8,000 killed, 10,000 foot and 2,000 horse caprured, but managed to extricate part ofbis army, including the elephants, wbich had not been engaged, and the treasury. He withdrew northwards, ineending to join his brother in ltaly.

(Polybios X.38.7-40.1; Livy XXVII.18-19; Lazenby pp. 141-142; Caven pp. 205-207.)

THEMETAURUS207 BC After his defeat at Baecula, Hasdrubal marched across the Alps into nonhern Italy, recruiting on the way to bring 78

his army up to perhaps 20-30,000 men. Near tbc rivcr Mctaurus in Umbria hc was faced by the Roman consul Marcus Livius Salinator witb a two-legioo army possibly including Gallic and Spanish auxiliaries sent from Spaio

by Scipio, perhaps 20-25,000 men, plus the praetor Lucius Porcius Licinus, with a weak two-legion army, perhaps 15,000. Tue other consul, Caius Claudius Nero, was watching Hannibal in the south, but with a picked force of 6,000 infanuy and 1,000 cavalry be slipped away and made a forced march north to join his colleague. Learoing of thcsc rcinforcemems, Hasdrubal wisbed to avoid baule, and tried to slip away by night. But his guides ran off, and he could not fmd a ford to cross tbc Mctaurus. Tbe Romans caught him up and deployed for baule, Nero on tbe right, Livius on the left, Licinus in the centre.

Hasdrubal put his Gauls on tbe left, wbere they were protected by hills. These were the least reliable pan of his army; many had desened during the nigbt marcb, while others slept through the baule, drunk. The cenere ofthe line was made up ofLigurians, the rigbt wing ofveteran Spaniards. Hasdrubal's ten elephants were posted in front of either tbc Ligurians or the Spanish. At first there was hard but fairly even fighting, the elephants getting out of band and doing damage to their own side as well as the enemy. But Nero, unable to come to grips with the Gauls apposite him, marcbed round bebind the Roman lines and fell on the flank and rear of H asdrubal's Spaniards.

These were almost wiped out, and the rest ofthe army also suffered heavily, many Gauls being butcbered in their beds, though some Gauls and Ligurians got away. Hasdrubal deliberately sought death once he realised be bad lost.

10,000 ofhis men were killed, 5,000 or more captured. Six oftbe elephants died, some killed when they ran amok by their own drivers, and the other four were captured.

(Polybios XI.1-3; Livy XXVII.47-49; Lazenby pp. 186-190.)

MANTINEIA 207 BC Philopoimen, now strategos ofthe Achaian League, led the reformed Achaian army to fighe Machanidas ofSparta near Mantineia. Numbers engaged are not specified but a full levy might produce about 20,000 Achaians and ratbcr fewer Spartans. Macbanidas had more and better trained mercenarics. Philopoimen deployed light mercenary infantry on a hill on his left flank, then thorakirai and Illyrians, with Tarantine cavalry on the flat probably in front ofthc left wing. Tbc phalanx was in tbe centre in severa! bodies, ics front covered by a broad, deep, but dry and gentlc-sidcd ditch. Citizen cavalry were on the right. Machanidas bimself commanded mercenary infantry and Tarantincs on bis righe, with tbe citizen phalanx on his lcft and catapuhs at intervals in front ofthe line.

Rather than stand around and be shot at by tbe catapuhs Philopoimen resolved on an immediate attack, opening with a charge ofthe Tarantincs, probably at the catapults. Machanidas' Tarantines countercharged, and both sides' mcrccnary infantry became embroiled in a generai fight. Eveotually the more numerous Spartan mercenaries won, and chased thcir opponents, including tbe thora.kitai and lllyrians, off tbe field. Machanidas rashly joined the pursuit himself, but Philopoimen, like Ptolemy at Raphia, joined bis phalanx, and promptly wheeled the leftmost unit on to tbc posiĂšon formerly held by Machanidas, outflanking the Spartan cit izcns. These, carricd away by che succcss of their mcrcenaries, charged the Achaian citizcns, but lost their formation crossing the ditch and were routcd by the Achaian countercharge. Tbe ditcb then of course promptly hindered their flight .

When Machanidas returned from pursuit to see tbe phalaox in rout, some ofbis men seem 10 bave joined in the flight at once. With the others be tried to force tbe di tcb, but his men lost heart seeing that the Achaians held the only bridge, and they too fled. As Machanidas tried to fmd a safe crossiog point, be was spotted by Philopoimen himself, and killcd as bis borse struggled up out ofthe ditcb. 4,000 Spartan ciĂšzcns were killed, plus an unknown number of mercenaries, and even more were captured.

(Polybios Xl.11-18; Plutarch, Philopoimen I O)

ILIPA 206 BC Hasdrubal son ofGisgo, recruiting in the west ofSpain, had assembled ao army of50-70,000 infantry and 4,500 cavalry, plus 32 eleph:mts. He marched east and was met near flipa by Scipio with 45,000 infantry and 3,000 cavalry, including allied Spanish horse and foot. The two armies encamped near each other and there followed some days ofindecisivc skirmishing. Each day, at a late hour, Hasdrubal would lead bis army out ofhis camp and deploy with African and Carthaginian infantry in the centre, the less reliable Spaniards on the flanks with the elephants in front ofthem. Sci pio would follow suit, with bis own Spaniards on his wings; neither side would advance their ma in lines, and soon both would reti re to camp. Once this had become a babit, Sci pio deliberately cbanged his practice to disconcen the enemy.

79

One evening he ordered his mento be armed and breakfasted before the next dawn, and at first light sent the cavalry aod light infanery to harry the Carthaginiao camp, and under cover ofthis he deployed the rese ofhis men, surprisiog even them by putting the Spaoiards in the cenere, and the Romans on the two wings. Hasdrubal mobilised ali his men wichout waiting for them to breakfast, aod deployed in his usual formation, while the light croops and cavalry skirmished. The skirmishiog wenc on till almost noon, Sci pio deliberately prolonging it ciii che enemy began co feel che efTects ofhunger. Then he pulled back his skirmishers and advanced, ordering che Spanish cenere to advance slowly while he manoeuvred the Roman wiogs 10 outflank che enemy. On each wiog the man iples and 111rmae wbeeled so the li ne became a column faciog outwards, and then marched till the beads ofthe columns were opposi te the ends ofthe Carthaginiao line {wbicb was longer than the Roman, because ofsuperior numbers). The columns theo wheeled and advanced quickly towards the enemy, then redeployed in line so tbat the beavy infancry faced the enemy while the cavalry and velites were io position to outflank them. Hasdrubal was unable to respond to this; the Spanish levies on his wings were probably incapable of much manoeuvre, wbile aoy move by his African centre against the Romao columns would risk exposing their own flanks to Scipio's cautiously advancing Spaniards.

H asdrubal's Spaniards put up a good fight, but they were hungry, outflanked, aod oucclassed by the Romans. In addition they were soon trampled by their own elephanes, who were forced back by the javelin lire ofthe velites. Tbe Spaniards retreated, first io good order bue soon in rout. Hasdrubal's cenere would oo doubt oow ha ve beeo attacked by the victorious Romans, but a sudden viole ne rainstorm put an end to the day's bartle. Hasdrubal planned to make a stand in his camp, but the next day his Spanish allies began to desert, and he slipped away overnight.

Scipio's outflanking move is probably the most complex manoeuvre carried out by a Roman army during this whole period, and shows the manoeuvrability and flexibility ofthe manipular system at its best.

(Polybios Xl. 20-24; Livy XXVUI.12-16; Lazenby pp. 145-151.)

ZAMA202BC Sci pio had invaded Africa, Hannibal had been recallcd to countcr bim, aod they met at Zama. Hannibal had about 40,000 men and 80 elephanes. Scipio had 23,000 Roman and ltalian foot with 1,500 cavalry, plus 6,000 Numidian infantry and 4,600 cavalry under Masinissa. Scipio deployed the Roman iofantry in three lines, but instead ofthe usual chcquerboard formation tbe maniples of pn¡ncipes and triarii were directly behind the hastati; the lanes between the units, inecnded 10 cbaooel che elephane charge, were filled with the velites. The Roman cavalry under Caius Laelius were on che left wiog, Masinissa's Numidians on che right. Haonibal put bis elephants in front oftbe infantry, who were io ebree lines reminisccnt of che Roman formatioo. Tbe first li oc was of Gauls, Liguriaos, Moorish archcrs aod Balearic sliogers, the sccond of Carcbaginian aod Libyan in fancry, the third of Hanoibal's veterans, some ofthem Brunii. Carchaginian cavalry wcre on the right, Numidians on the left.

The baule stancd with skirmishing betwcen the Numidian cavalry, tben an elephant charge. The elcphants were disconcerced by tbe Romao trumpets and the cheers ofthe men; some refused to charge, turning back against their own side's Numidian cavalry. Masinissa charged these men wbile tbey were disordered by the reti ring elephants, pursuing tbem ofTthe field. Those elephams which did charge were met by the advancing Roman velites. Losses were heavy on both sides. Some elephanes were thrown back on the Carthaginian cavalry and Laelius, like Masinissa, exploited the confusioo and chased the enemy cavalry from the field. Other elephanes chased che velites down the lanes between che heavy infanery; these were killed or driven ofT, but che velites were so badly mauled tbat they were out ofthe battle.

Thc Roman lzastati, supponed by the principes, now charged and routed tbe first Carchaginian iofaotry line. Tbe Punic foot oftbe secood lioe refused to let the mercenary fugitives througb their ranks, so they fled out to tbe wings.

These Puoic and Libyan foot badly mauled the pursuiog lzastati, but the interveneion oftbe principes broke them, and they too fled to the flanks as Hannibal's third line refused to let them through. Sci pio paused and re-organised his infantry, bringing the principes and tria rii up on to the flanks ofthe lzastati. The need to lengthen the line like tbis argues that some ofH annibal's fugitives had rallied on the flanks ofbis veterans. The rwo infanery lines closed, aod were eveoly matcbed unti! Laelius and Masinissa returned from pursuit, charged Hannibal's veterans io tbc rear and broke tbem. Few escaped, as the Roman and Numidiao cavalry hunted down most ofthe fugitives. Hannibal lost 20,000 killed and almost as many caprured; Scipio lost 1,500-2,500, and Masinissa rather more.

(Polybios XV.9-14; Livy XXX. 32-35; Lazenby pp. 220-226.)

80

KVNOSKEPHALAI 197 BC Philip V of Macedon, wich a phalanx of 16,000, 7,500 otbcr infamry and 2,000 cavalry, was campaigning in Thessaly againsc 1hc Roman consul Tirus Quinctius Flamininus, who had about 18,000 Roman and ltalian troops, 8,000 Greck allies, moscly Aitolians, and a few elephants. Afier some skirmishes in the gardens and orchards round Pherai, where there was no suitable ground fora ba1tle, boch sides marched for Skocoussa seeking supplies, and losc track ofeach other when they were separated by the range ofhills cali ed Kynoskephalai. On thc third morning ofche march, in thick mist, advance panies ofthe two sides met acop che Kynoskephalai hills and skirmishing began. At first tbe Rom ans were thrown back by the Macedonian forcc; Roman and Aicolian reinforcemems turned the tables; and then the Macedonians were reinforced by ali Philip's cavalry and mercenaries, who pushed the Romans back down the slopcs, only thc Aicolians saving them from rout. lt was now clear that a full-scale battle was developing, :md bo1h sides led out cheir ma in armies. Flamininus got co the fìght firs1, pushing back the Macedonian merccnari cs wit h his lcft wing. Philip hurricd up with the Macedonian peltasts and halfthc phalanx, which he deployed whilc the rest ofche phalanx, under Nikanor, trai led along behind.

Both gene rais regrouped their advance forces. Philip doubled the depch ofche half ofche phalanx he had with him, closing up to the right to allow Nikanor room 10 deploy, and charged down bili, pushing back the Roman lcft.

Seeing his left was losing but that the Macedonian left wing was stili hurrying up in column of route, Flamininus led bis right wing, wich tbc elephams in front, to che auack. Thc elephants broke the leading Macedonian units before they had time co deploy, and mosc of Nikanor's forcc joined in tbc rouc. Most ofthe Roman right pursued tbem, but a tribune took 20 manipuli from the victorious right wing and charged downhill imo the rcar of Philip's embattled wing ofche phalanx. Unable to turn to face this, the Macedonians suffered heavy casuahies and broke.

Philip, seeing no chance ofretrieving che situation, fled wich what croops he could collect. The Maccdonians lost 8,000 killed and 5,000 prisoners, tbe Romans only 700.

(Polybios XVIII.18·27; Lìvy XXXIII. 3· 10; Plutarch Flami11i1111s 8; Walbank ( 1940) pp. 167-172.)

THERMOPVLAI 191 BC Antiochos Ili crossed to Greece ac the instigaùon of tbc Aitolians, but met with linle locai support, while the Romans seni a much larger force against him. Too heavily outnumbcrcd for an open battle, Antiochos hcld the pass ofThermopylai with his 10,000 foot, 500 horse and six elcphants. The pass was blocked by a rampart and ditch, with a stone wall cxtending from this up tbe landward cliffs. Antiochos had psiloi and "peltasts" (perhaps the argyraspides) in front, the phalanx behind tbem round che rampari, cavalry and elephants by the sea, and light infamry on the inland cliffs. Thcre were catapults mounted on the wall. 2,000 Aitolians occupied the three pcaks of Kallidromon, Teichious and Rhoduntia, which commanded the paths by which oucflanking moves might be made.

Another 2,000 Aitolians in nearby Herakleia played no pan in 1he ba1tlc.

Tbe Romans bad 20,000 fooc, 2,000 horse and 15 elephants, under thc consul Marcus Acilius Glabrio. Dcciding to outflank tbe pass as 1he Persians (and indeed the Gala1ians) had done, he scnt Marcus Porcius Caco and Lucius Valerius Flaccus with 2,000 men each against tbc Aitolians on the peaks. Meanwhile he attacked thc pass fron1ally, beat 1he fìrs1 Seleucid line and afier a hard fìgbt forced tbc phalanx back. Thc phalanx however 1ook up a posi1ion on the rampare, whence the Romans were unable 10 dislodge 1hem.

At this point Cato's force, having su rprised and routed 1he Ailolians on Kallidromon, appeared in the Seleucid rear near cbeir camp, pursuing che Ai1olians. Tbc Seleucids, worried for cheir baggage and uncertain ofcbe size ofCa1o's force, fell back to defend che camp; but tbe Romans followed up and forced 1heir way imo che camp along wi1h the retreating Seleucids, turning che recreat imo a rout. For a t ime che elephants held up the Roman pursuit in the narrow pass, but eventually 1hey were ali killed; aJJ An1iochos' infantry scem 10 have been killed or taken, but the cavaJry escaped. The Romans lost 200 men.

Flaccus had been rcpulsed by the garrisons ofTeichious and Rhoduntia, and i1 was probably these Aitolians who raided tbe Roman camp; but they withdrew without a fìght when che main army rcturned from pursuit.

(Livy XXXVl.1 5-19; Appian Syr. 17-20; Bar-Kochva ( 1976) eh. 13.)

MAGNESIA 190 BC A Roman army ofcwo legions plus Pergamene and otber auxiliaries, about 30,000 men and 16 African elephants, 81

defeated Antiocbos Ill's Seleucid army of 45-58,000 foot, I 2,000 cavalry, 54 Indian elephants, camels and scythed chariots.

The Seleucids opened the battle with a charge by the scythed chariots posted in front oftheir lef1 wing. Eumenes II of Pergamon, in command ofthe Roman right, sent forward light infantry and cavalry to shoot at them from front and flanks. The chariots tumed back under the lire and in their flight disordered the camels and cataphracts supporting them, the latter particularly suffering from the scythes. Eumenes, following up with his Roman and Pergamene cavalry, routed not only the disordered enemy cavalry but also the left wing infantry.

Meanwhile Antiochos, commanding his right wing, outflanked the Roman left, because the Roman advance had opened a gap berween the small left-wing cavalry force and the river on whicb this flank had originally rested. He also attacked frontally with his cataphracts, breaking the whole Roman left wing, and pursuing it a considerable distance towards the Roman camp. Here he was held up by the camp commandant Marcus Aemilius with bis camp guard of Macedonians and Thracians; Aemilius also managed to rally some ofthe fugitives.

The Seleucid pbalanx was not outflanked by Eumenes' cavalry and light infantry, and the light troops skirmishing in front of it al so fell back. The outflankcd phalanx formed a hollow square, with tbe light troops hiding in the cenere rather than replying to the lire ofthe Roman light troops and cavalry. At fĂŹrst the phalanx withdrew in good order, the Romans reluctant to dose with it. Eventually though the elephants posted berween the ten bodies ofthe phalanx began to panie under the constane rain of missiles, breaking up the infantry's formation; the legionaries charged the disordered phalanx which collapsed in rout. The Romans pursued and sacked the Seleucid camp.

Antiochos, engaged with the camp guard and the rallying fugitives, was now also anacked by Eumenes' brother Analos with 200 cavalry from the victorious Roman right. Hc routed this force, but seeing the rest of his army broken, retired from the lield with his own right wing. The Romans lost 350 killed, Antiochos 50,000 or more killed or captured, and 15 elephants taken.

(Livy XXXVII.39-44; Appian Syr. 30-36; Bar-Kochva ( 1976) ch.14.)

MOUNT OLYMPOS AND MOUNT MAGASA 189 BC The new Roman consul Gnaeus ManliusVulso took over the army which had won Magnesia and marched against the Galatians. Their three tribes occupied well-supplied fortilied camps on two mountains, the Tolistobogii and Trocmi on mount Olympos, the Tectosages on mount Magaba, three days march away near Ankara. They believed they could hold the steep slopes against any Roman assault and wait out a siege. Manlius Vulso however paid special attention to the supply of missiles, realising the Galatians' vulnerability to skirmishers. He marched lirst against Olympos.

He led the main army along the road up the south face ofthe mountain, detaching cwo columns up more difficult approachs on the flanks; neither ofthese routes proved practicable, and the rwo flanking forces later rejoined the main road in Manlius' rear. Ali three forces were oflight infantry supported by legionaries. The Galatians held a summit overlooking the road with 4,000 men, so Manlius sent velires and Pergamene skirmishers to shoot at them, following up cautiously with the legions. The Galatians soon ran out of missiles, resoning to thrown rocks, and suffered heavily from the Roman lire; a few charged but were cut down by the velires. Tbc survivors retired to the camp, where the ma in army d rew up in front ofthe ramparts. Again they were forced by missiles to reti re, inside che camp. While Roman light troops shot over the rampare imo the milling mass of warriors and non-combatants inside, a lcgionary attack broke the forces left to guard the gates, and as tbe Romans charged into the camp the Galatians fled out of it. The Romans pursued, and Galatian losses were estimated at I 0-40,000, including noncombatants captured.

Manlius thcn marched 10 Magaba. During negotiations the Galatians treacherously ambushed him with 1,000 picked cavalry, but these were driveo offby Roman foragers. The Galatians deployed on the mountain with 50,000 Tcctosages and Trocmi in the centre, 10,000 dismounted cavalry on the rigbt, and 4,000 Kappadokian and Paphlagonian ali ics on the left. After careful reconnaissance Manlius sent light infantry followed by heavies against the cenere, and two more columns against the flanks. The battle was almost a repeat ofOlympos, with the centre so dcmoralised by missiles that it broke as soon as the Roman heavy infantry charged; thc wings waired unti! the Roman flanking columns arrived, then they too broke at the fĂŹrst charge. Galarian losses were 8,000, mostly in the pursuit.

(Livy XXXVIIl.19-29.)

82

PYDNA 168 BC The Roman consul Lucius Aemilius Paullus, witb rwo legions plus Italian, Pergamene and Numidian allies, about 37,000 men and 34 African elephaots, had invaded Macedoo, outflaoking King Perseus' army of 21,000 phalangites, 17,000 otber infaotry and 4,000 cavalry, which had been occupyiog a fortifìed position in tbe Elpeus valley. Perseus retired and offered battle in a flat plain suitable for the pbalanx, near Pydna; but the Romans refused and pitched a camp on rough ground. The next day fìghting broke out at the stream where both sides were drawing their water, between Perseus' Thracians and Paullus' Italian allies. Perseus advanccd his whole army to support them. Paullus formed up tbe Roman army in response, but the Macedonians advanced very quickly and as a result the battle took piace dose to tbe Roman camp.

The Romans opened thc attack with ao elephant charge supported by tbe Laùn allies, which broke the Macedonian left wing. The Macedonian phalanx at fìrst pushed back the legions, while the pehasts broke a cohon of Paeligni and perhaps anotber of Marrucini. But as the phalanx followed up over the rough ground gaps began to appear in its formation. Paullus rode along the line ordering the maniples 10 act iodependently, and they thrust into the gaps, breaking up the solid front oftbe pbalanx. As the Macedonians lost impetus and the battle broke up into single combats the bener-equipped Romaas cut tbe cnemy 10 picces. 3,000 elitc Macedonians, pcltasts or agema, fought 10 the death, but the rest ofthe Macedoniaos broke and ran after a hard fìgbt. Pcrseus and thc cavalry had already fled. The battle had only lasted an hour. Macedonian losses were 20-25,000 killed and 11,000 prisoners; the Romans only lost 100 dead.

(Livy XLIV. 40-43; Plutarcb, Aemilius Pau//us 18-23.)

ELASA 160BC The Seleucid generai Bacchides marched against Judas Maccabeus, wbo controlled ali Judaea after defeating previous Seleucid expeditions. Bacchides had 20,000 infantry and 2,000 cavalry; I Maccabees claimsjudas met him at Elasa with a mere 3,000 picked men, ofwbom ali but 800 fled on seeing theenemy. Tbe courseofthe battle makes these numbers scarcely credible, and this seems to be an example ofthe sources' deliberate understating ofjewish numbers. Judas controlled roughly the same area from which Jonathan could expect 30,000 men a mere eight years later. Thcre may stili be truth in the story ofheavy desenion when faced with a pitched banle on a flat plateau suitable for the Seleucid phalanx aad cavalry.

The Seleucids deployed with an advaaced scrcen of skirmishers and with cavalry on both flanks; Bacchides commanded the right, his colleague Alkimos may ha ve led the left. The main forces engaged and fought a long and at fìrst indecisi ve baule (which hardly sounds like 800 men against 20,000) but tbc Seleucid cavalry wings seem at first to have been held back. Judas with a picked force charged and broke Bacchides' wing, pursuing them a considerable distance from the bartlefield (probably far and fast enough 10 indicate that tbe pursuit was at cavalry speed). Seeing this, the Seleucid left wing cavalry left their position, chased after Judas' men and took them in the rear, killing Judas and routing his force after a fierce fìght. Their freedom of action indicates either tbat they were unopposed by enemy cavalry or had already beaten them; either case means the Seleucids had a comfortable cavalry superiority. Judas' infantry were tbus isolated and beaten, though whether before or after tbc Seleucid cavalry rerurned is not stated. As a resuh of this banle Seleucid authority was restored in Judaea, aod Judas' brothers Jonathan and Simon were reduced to guerilla leaders.

(/ Maccabees 9. 1-21; Bar-Kochva (1976) ch.16.)

AZOTOS 148 BC Jonathan had restored an independent Judaea by guerilla warfare and playing offriva! claimants for the Selcucid throne. The Palestinian coastal towns werc beld for Demctrios li by the govcmor ApolloniosTaos, so to gaio them for his own kingdom, Jonathan anacked Apollonios io the name of the rivai Seleucid Alexander Balas, and took Joppa. Apollonios had an army of 3,000 borse and 8,000 foot (or perhaps more) referred to as "the powcr of the ciùcs", so possibly composed largely ofthe mi licia of the coastal cities. Jonathan started tbc campaign with I 0,000 men, but had been joined by his brother Simon, perhaps with another contingent; Jonathan's roral resources were 30-40,000 men. Apollonios had a comfortable cavalry superiority, so lured Jonathan into tbc plains near Azotos, and set an ambush of 1,000 horse-archers in bis rear.

Jonathan took up a stationary defensive formation while tbc horse-archers surrounded and harassed his army fora considerable time. Evenrually the horse-archers "tired"; perhaps their ammunition ran out, or tbey may even have been driven off by Judaean counterfìre. M eanwhile Apollonios' main forcc apparently did nothing, which is

83

surprising; however firm the Judean defensive formation, this wou ld seem to have been his best moment 10 attack it, as if he couJd not beat Jonathan with the support of the horse archers, 1hen he never would. Perhaps the inexperienced militia refused to advance.

In any evem, when the horse·archers were neutralised Simon "brought forth his host" and routed the Seleucid phalanx, while tbc remaining cavalry were al so driven off, possibly by Jonathan at 1he head ofhis own cavalry. The prominence given to Simon suggests he may have "brough1 forth" his men from ambush; although I Maccabees emphasises the open plain, some sort of cover existed 10 conceal 1he Seleucid ambush, perhaps olive·groves or orchards, so Simon may have used the same sort ofterrain. Following up 1he victory, Jonathan took Azotos and other towns. Total Seleucid casualties in the battle and the mopping·up were 8,000.

(/ Maccabees I 0.69-85; Bar-Kochva ( 1975).)

CORINTH 146 BC The Roman consuJ Mummius, with 23,000 infantry and 3,500 cavalry (probably two legions plus ltalian allies), with Cretans and Pergamenes, advanced into the Peloponnese against the revoluùonary government ofthe Achaian League. The Achaian generai Diaios camped at Corinth with 14,000 infantry and 600 cavalry (plus probably some survivors ofanother army beaten earlier). The Achaians made a successful night attack on the camp ofthe Roman advance guard, innicting heavy casualties.

Encouraged by this success they offered battle thc next day, bui their cavalry, hcavily outnumbered, did not wait to receive the Roman cavalry charge but flcd a1 once. Thc Achaian infantry however beld tbc legions umi) a pickcd force of 1,000 Roman infantry chargcd their flank and broke 1hcm. Some Achaians took refuge in Corinth, but no defence was organised because Diaios flcd 10 Arkadia, and Corinth was sacked with unusual tboroughness by che Romans.

(Pausanias VII.15-16.)

84

DRESS AND EQUIPMENT 1.

DARIUS 111

The last Achaemenid king of Persia, "a tali and handsome man" is shown as he appears in thc lssos mosaic, from Pompeii, standing in his chariot (for which see the note to figure 176). He wears the usual Persian (originally Median) costume; the headgcar is thc tiara, a clotb cap with two lappets ticd under the chin and a third at the neck. lt is here dark yellow, with thc pcak stiffened to stand erect. This was a distinguishing mark ofthe king, commoners wearing the peak llopping to one side as in the following figures; to wear one's tiara erect was equivalent to laying claim to the throne. Around the cap is tied the diadem, a narrow fillet ofblue cloth spotted white. Satraps are also shown in diadems, and they were adopted by the later Hellenistic kings. His tunic is grey with a broad white stripe down tbe front; Xenophon mentions royal costume ofpurple and white, the Book ofEsther blue and white. Oarius' legs are hidden by the chariot, but Xenophon mentions red trousers as royal wear; Aischylos mentions satTron shoes, while an earlier king's shoes carved at Persepolis ha ve traces of red and blue paint. What appears to be a purple cloak is probably the sleeved ka11dys ofthe next figure; Xenophon conflrms a purple ka11dys as royal wear, while Curtius mentions cloth-of-gold ornamented with golden hawks. The king wears a gold torc round his neck, like most Persian nobles.

Weapons carried in his cbariot seem to have been bow, shield, dagger, and perhaps javelins. Cunius says the royal dagger sheath was cut from a single precious stone, while the belt was knotted, like a woman's, Persian men's belts usually having a clasp.

2.

PERSIAN OFFICER ANO ST ANDARDS

This man's costume is from a Persian silver statuette, showing baggier trousers than usual and a sleeved overcoat, the kandys, worn as a cloak. lt is also shown on tbe Persepolis sculprures and the Alexander sarcophagus from Sidon. Xenopbon says cavalry put their arms througb the sleeves only for royal inspections. This example shows a turned down collar or hood (see figure 2a for the rear view), wbat appear to be buttons or buttonboles down thc right front, and braid on the sleeves.

Many officers would wear scale armour. Many iron scales, and a few bronze and gilded iron ones, bave bccn found at Persepolis, but it is rare to see scale armour in contemporary representations of Persians as it was usually worn

2

85

under a tunic. This figure's cuirass, from a Sth-century Greek vase, may be the only ancient picture of Persian armour surviving. lt is similar to Grcek types, with thc scales scwn on a linen foundation, and shoulderpicccs brought ovcr from tbc back and laccd in front. Tbe ccnual pane! oflozcngc-pattcrncd clotb may conceal a mernl plate (see figure 28). Bclow tbe waist is a curious skin perbaps of padded clotb, and tben a layer of pteruges, protective lcather or stiffened linen strips.

The standard he carries is from che Issos mosaic, and unforrunacely much ofthc originai is missing. The surviving fragment seems to be grey-brown with an unclear red design and perhaps a fringe. Possible emblems for Persian flags would include gryphons, winged bulls, lions, the winged suo disc or the similar emblem of2c. The crossbar ofthc standard does not seem to be exactly horizontal on the shaft, pcrhaps being ticd with some flexibility. I have thus intcrpreted 2b, a rather curious standard from an earlier vase painting, assimilar but witb two crossbars, each bearing a flag, before and behind the shaft. Xenopbon in bis Kyropaidia says each senior officer had a distinctive standard. He describes tbe Persia o royal standard as a son of gold eagle on a shield. The eagle does not featurc as a royal emblem in Pcrsian an, and Xenophon is probably describing the eagle-winged emblcm of Ahura-Mazda; 2c shows chis as a standard.

3 and 4.

PERSIAN CA VALRY

Thesc two examples ofthe typical Persian cavalryman are based on thc Issos mosaic, and may represent the royal guard as chey are shown fighting in defence ofthe royal chariot. Similar warriors are depicted in numerous other sources.

They wear typical Mcdian costume. Ali thc caps on tbe mosaic are dark yellow, including those ofDarius and bis charioteer. Yellow caps are known from othcr sources too, but cannot be universal as at least one infantryman's is whitc (see figure 9) wbile some (though not of soldiers) in Greek art are striped or otherwise patterned. Thc mosaic does suggest cap colour may ha ve been uniform within a unit. The cavalry show no other sign of uni formity, though chere are indications of it elsewhere (see note to figure 8). Persian dress generally was brightly coloured and decorated, tbougb no doubt thc poorer men wore fewer and duller hues. Purple was wom by royal relatives and thc higher nobles as a sign of rank. Pale soft shades were qui te popular - white, mauve, pale ycllow, beige, light greco, light blue; so were dark brown, red-brown, sand, red, crimson and grcy. They could be dccorated by bcing covcrcd with squares, ci rei es or rosettcs in contrasting colours (usually reserved for formai long robes, but figure 4 's brown and whitc tunic is similar), vertical or horizontal scripes like figure 8's trouscrs, or rows ofembroidery at neck, bem, cuff, centre front and back oftunics, along the shoulder and round the upper arm, like figure 2's ka11dys or 3's tunic.

This is similar to Saka styles (scc figure 76). Shoes are white on the mosaic, yellow-brown elsewhere.

Figure 4 wears a bright red whi te-edgcd quilted linen corselet, with shoulder-pieces, rectangular neck-flap and probably preruges likc Grcek armour. Non-metallic armour seems to have been quite common, though metal scale was the ideai. 4a sbows a linen or leatber cuirass, probably with a metal plate on the chest like figure 29, from a Lykian relief of a Persian-style cavalryman; note that the check lappets ofhis tiara are worn loose. Figure 3 has no visible armour, but probably wears scale under his rather voluminous tunic, a practice memioned in severa! written sources.

Main weapons are two corncl-wood javelins called palra, about five or six feet (1.5 to 1.8 metres) long witb iron or bronze beads. One pa/1011 wou ld be thrown, tbe second could be thrown or used to thrust, Xenophon recording it was a better thrusting weapon than tbe flimsy Greek cavalry spears. The superiority ofthe Macedonian cavalry xys1011 lcd Darius to rearm some ofbis cavalry with similar long spears before Gaugamela; bow widespread was tbe reform is not known, bue he would presumably have started with the guard and may noc have got much furtber.

(Curtius adds that swords and shields were also issued to men who previously bad been armed only with javelins; this is probably an attempt to arm decently some ofthe worse equipped levy infantry, rather than another reform of cavalry armament.) Against Macedonian cavalry tbe Pcrsians seem generally to have thrown both javelins and closed witb band wcapons. Figure 3 has a sword similar to the G reek kopis (see figure 18). Straight swords are also sbown on the mosaic. Figure 4 has tbe traditional akinakes, a long dagger tbonged to bis right thigh. Hilt and scabbard could be covered with gold or ivory decorated in rclicf, while tbe uefoil chape might be gold, ivory or bronze. lts charactcristic decoration began as a lion fighting a ram, but degcnerated into sryliscd abstraction. He al so carries an axe, as rccorded at the Granikos. The sagaris of figure 7 might be substĂŹtuted.

5.

PERSIAN EXTRA¡HEAVY CAVALRYMAN

Some Persian cavalry were heavier than tbc above majority type, witb extra armour and riding part-armoured

86

horses. The borse-armour (see figure 167) derived from 1he Massagetai, a Saka people. Xcnophon dcscribes 1he 600 guards of 1be re bel sa1rap of Lydia, Cyrus, in 40 I wearing belme1s, cuirasses and thighguards, riding armoured horses and figbting witb javelins and Greek curved swords. He mentions Persian noblcs, no doubl from 1his guard, wearing crimson kandys and embroidercd trouscrs on thc march, some wi1h torcs and bracclets. In thc Kyropaidia, a dcscription based on these men adds that ali the armour was bronze, thc helmct white-plumcd and the cuirass covered by a red tunic (which Plutarch confirms). A rclief from near Daskylion in Phrygia shows cavalry with thighguards and conica! crestless helme1s, probably the locai sa1rap's guard. Xenophon's refcrencc to crcsts may indicate Cyrus' men wore Greek bclmcts; some Persian satraps wear crested Attic helmets on their coins, and one seal depicts a cavalryman in thighpieces and Greek-style crest.

This figure is based on tbc Daskylion reliefsupplememed by Xenophon. His belmet is thc typc shown on 1he relief, identica I to a bronze one captured from the Sth-ccntury Persian invaders of Greece. The overtunic is pulled up 1hrough his beh as on figures 7 and 8, the skirt ofhis undenunic sbowing at tbc rear. The thighguards are sbown hcre as scales, the most likely medium ifthey wc re bronze, as Xenopbon daims. Tbey are al so shown on the Lykian sarcophagus of Payava, those being narrower than here. They would be suspended by a strap over the base ofthe horse's neck, and fastened to the rider's beh a1 thc rear. Xenophon calls this defence both parameridia, thigh-pieces, and paraple11ridia, side-pieces, as tbey protect both the horse's sides and 1he rider's legs. 5a, from a Lykian tombpainting of the early 5th century, sbows what may be a prototype version of thc parameridia anachcd to 1hc saddlcclo1h of a rider in Persian dress. This suggests 1ha1, unlike thc horse-armour, they may ha ve bccn of Anatolian ori gin.

Onc defencc noi shown but possibly used is a tubular guard for the bridle-arm, like thosc offigurc 10 but probably of leathcr. Xcnophon recommends this for Greek cavalry, and though he does not mention Persian horsemen wcaring it he does ascribe it to Persian chariotccrs. lt may 1herefore have been of Persian origin, like the other innova1ions hc recommends.

87

Tbese cavalry were probably ncver numerous, as the lightcr varicty is far commoncr in art, even from Anatolia wbence com es much of our evidcncc for the heavicr type. The Massagctai and Bactrians in front of the left wing at Gaugamela rode armoured horses, and so probably did the Kappadokians and Armenians in a similar position on the right - certainly the Armenians later adopted tbem on a large scale. Thcsc two nations may have provided the armoured·horse cavalry recorded on the Persian right at lssos, since the casterners did not arrive in time for that battle. Cenainly not many ofthe Persians themselves seem to have adopted the extra armour, despite the example oftbe guards ofCyrus and the satrap ofDaskylion; the lssos mosaic suggcsts that cvcn thc royal guard had not taken it up.

6.

P ERSIA N "APPLE-BEARER "

Thc "Applc·bcarcrs" were thc royal foot guard, so callcd from thc goldcn apples decorating their spearbuns, an elaboration of thc traditional spherical counterweight ofthc short Pcrsian infantry thrusting·spear. This figure is based on two fallcn warriors on the Issos mosaic. Neither has a horsc, so they are unlikely to be fallen cavalrymen.

Their proximity to thc King suggcsts thcy are the foot guards. One has a shield ofthc typc carricd by Greek hoplitcs while the othcrcarries a bow, so wc may deduce that the unit was armcd with bow, spea r and shield (or possibly was a mixed force of spcarmen and archers). The shield has a bronzc rim, thc rest of the face polisbed 10 a mirror brightness, with no apparent blazon.

Both figurcs wcar thc sa mc dark yellow riara as che cavalry, and red-brown tunics. The shielded figure 's trousers are also visible, and are thc samc brown, with both tun ic and trouscrs dccoratcd in whitc. He has a gold torc and bracelct. The bowman has a mid-brown leather cuirass, thc torso divided imo horizontal strips, wbich is reinforccd with bronzc studs. 6a is wom by the only cuirassed Pcrsian on thc Alcx:mder sa rcophagus. lt is the same cuc as thc Macedonian cuirasses but decorated in panerns visible now only under ultra-violet light, including rwo beasts, pe rhaps lions, facing each other.

7.

PERSIA N INFANTRYMAN OF THE KARDAKES

There has bcen some debate over the armament ofthe kardakes when they were caUed out to fight at Issos, since of eyewitnesses, Ptolemy (quoted by Arrian) calls them hoplitcs, Kallistbenes (in Polybios) peltasts. Most modero scholars deduce from thi s that they were an atte mpt to produce a native Persian close-figbting infantry to support thc mercenary G reek hoplites against the Maccdonian phalanx. Xenophon (in tbc Kyropaidia) describes thc kardakcs' normai hunting weaponry as two javelins, bow and quivcr, a kopis or a sagaris (the bronze pick-like Saka battleaxe which had bcen adopted by Persians) and a wickcr shicld (gerrho11); Strabo confums they used bows and javelins. The Alexander sarcophagus shows Persian infantry, some or ali ofwhom may be kardakes, some ofwhom carry hoplite shields though more ha ve crescent·shapcd shiclds like the Grcek pelrai (see figure 20) which could be Xenophon'sgerrha, since the pelre was made ofwicker. Othcrs shoot bows. lfthe kardakes took the field with cheir normai javelins and perhaps their bows, some carrying wickcr shiclds while ochers had chosen to adopt the heavier hoplice cype, then this could explain the apparently contradictory dcscripcions of"hoplites" and "peltascs". While their function clearly was 10 suppon the Greeks, thcrc is no cvidcncc that they were re-equipped in a hcavicr stylc, and really no need to assume it. T bc absence of specific references to them carrying bows at Issos may ha ve bccn because they did not carry these imo bartle, or may havc been simply because che speed ofthe Macedonian advance gave little time for tbeir use. Arrian does mencion thc need to dose rapidly with Persian archers at Issos, and while these arcusually assumed to have been a scrcen in front ofthc mai o line, they may well bave been che kardakes as Lhe originai light infantry screen seems co havc been withdrawn.

This figure's shicld blazon is actually from the insidc of a pelte on the sarcophagus, again visible by ultra-violet. 7a is a rimless round or ovai shield from the Issos mosaic; from tbe outside in, dark green, red, white, dark green, red centre.

8.

PERSIA N ARCHER

The mainstay of 5th-cenrury Pcrsian armics had becn bodics of close·order infancry with spears, bows and wickcr shields; while the "apple·bearers" wcre srill apparently armed in similar stylc the bulk ofthe Persian infantry was now skirmishing archers and slingers. The changcover probably took piace about 400, as Xenophon faced the old stylc ofinfantry at Cunaxa bu t was harassed on his retrcat by archers and slingers, some ofwhom were definitely real Persians rather than subjecrs, since a Pcrsian armed only witb a bow and a ligh1 axe "like thosc which che Amazons carry" was caprured.

88

This figure (including bis pauerned trousers) is based on the Alexander sarcophagus. He wears two tunics, Lhe outer one pulled up through the beh to shonen it. This outer tunic may have been dyed in a uniform colour for at least some units, since at Cunaxa ali King Artaxerxes' men wore white tunics over their armour, while ali Cyrus' rebels were red ones. However this was a special case, with two Persian armies fighting each other and the obvious need to distinguish between the two sides, so it may not have been normai practice.

The Persian bow was a composite weapon quite large by ancient standards (perhaps 4ft./l.2m long). Xenophon notes that Persian archers outranged the Cretans, but that later the Cretans practised long-range shooting at a high trajectory with caprured Persian arrows. This suggests that the longer Persian range was the result oflighter arrows and different training, rather than any difTerence berween the bows. Persian arrowheads were usually bronze and three-edged, either 3-4cm long and socketed (Sa and b) or triangular, tanged, and about Scm long (&);a few Larger (8cm) tanged iron heads have been found. Socketed heads were fitted to a wooden foreshaft which was in tura inserted into the main shaft, light, hollow, and made from reed. The Median-style combined bowcase/quiver hung from the belt at the left hip (oddly, as arrows must have been drawn with the righe hand). Figure 7 shows the case closed, with the flexible upper portion fastened to cover the bow, while figure 8's is open. This man's secondary weapon is che usual aki11akes fascened to the right leg, but an axe like figure 4's could be substituted.

Other Iranian archers would be similar, such as the reliable Marcii and levied Uxii, " Red Sea men" and perhaps Sitakeni at Gaugamela, and the Medes, Karmanians and others who served che Seleucids. An archer dressed Like this figure on the Artemision relieffrom Magnesia (see note to figure 34) may be based on the Persians recorded as stationcd there under Seleucus II, while a relief ofSeleucid or Parthian date from Qizqapan (western Media/Iraq i Kurdestan) shows two Median archers dresscd as bere, one adding the kandys.

e

a

89

9.

PERSIAN SUNGER

Sliogers are usually coupled with Persiao archers in Achaemenid and Macedonian armies. Xenophon records the slings shot stones "as big as one's fìst" which had only halfthe range oflbe smaller Rbodian lead sling-bullets. This figure is based on tbe opponent offìgure 31; his armarne or is no looger visible and he may perhaps bave beeo some kind ofjavelinman. Cap is white, tunic pale green and trousers beige, all undecorated and probably typical ofthe relatively pia in dress ofthe infantry. Tbe shield is pale yellow-brown with a red-brown scar. I am slightly suspicious ofthls, sioce it is from a Macedooian paioting and bears a dose resemblance to Macedonian patterns (see figure 40); it may thus not be a real Persian blazon.

The Seleucids' Kyrtii slingers, and perbaps those escorting their elephants, would be similarly dressed.

10. SCYTHED CHARIOT DRIVER Xenopbon describes Persian cbarioteers in the Kyropaidia, protected below the waist by the body ofthe chariot, above it by cuirass, armpieces and belmet, with only the eyes showing. This last, probably a slight exaggeration, suggests a Greek helmet with cheekpieces rather than the open conica! type, and this man has an Attic helmet from satrapa! coinage. He is shown with a scaled composite cuirass and laminated armpieces oftbe type mentioned under fìgures 5 and 48. Later Seleucid and Pomic charioteers would probably be similar.

The driver ofDarius' chariot io the Issos rnosaic is dressed like figure 4, with yellow ca p, red quilted cuirass, and grey tunic braided down the sleeves.

11.

ARACHOSIAN CAVALRYMAN

Thls figure has tbe twisted head-fillet, tunic, baggy lrousers, high boots and akinakes worn by Arachosians, Arians and Draogianiaos in the Persepolis reliefs. Herodotos notes tbat Draogianians wore panicularly brightly coloured clothes, so perhaps their neighbours did too. Arachosian cavalry fought for Darius at Gaugamela; their satrap also commanded the "mountain Indiaos", probably the Paropamisadai ofthe Hindu Kush, who fought brigaded with Arachosian cavalry for Alexander and for Eumenes ofKardia. They later probably fought for the Greek kings of Bactria. They were fairly certainly light javelin cavalry; Alexander raised similar lzippako111is1ai, mounted javelinmen, in nearby Aria, while the references to lzippakomistai in his lndian campaign probably mean the Aracbosians and Paroparnisadai. Tbe impact of Amigonos' horse-archers at Paraitakene suggests Eumenes had no similar troops, and thus that ali bis ligbt horse bad javelios.

90

12.

PARTHYAIAN HORSE ARCHER

I bave used the Greek spelling "Parthyaian" to distinguish the inhabitants under the Achaemenids and Macedonians ofthe province usually called Parthia from the later nomad Parni invaders who founded the Parthian state. Antigonos had 1,000 Median and Parthyaian horse¡archers and javelinmen (longcl1ophoro1) at Paraitakene, later reinforced by another 1,000 Medes. A reference to 2,000 Median longcl1oplzoro1~ plus the generai dose relationship between Medes and Persians, suggests that the Medes provided the javelinrnen and the Parthyaians the archers, bui alternatively the two nationalities may have been mix.ed. T he Medes would bave resembled figure 3 (without the armour he is presumed to wear under his tunic, as this corps was clearly light skirmishers). This figure wcars the Parthyaian dress shown on the Persepolis reliefs, similar tO the last figure save for tighter trousers and shorter boots. Darius' Parthyaians at Gaugamela would no doubt bave been sirnilarly equipped.

13.

BACTRIAN OR SIMILAR LIGHT CAVALRYMAN

This man wears Persian drcss, with a floppier cap than usual, and carri es a bowcase (unusually, on his right) and 1wo spea rs. No hand weapon is shown on the originai source, a chalcedony intaglio dated by Rostovtzeffto the early Parthian era on stylistic grounds; it may alternatively be late Achaemenid Persian. Whatever che date, it indicates the survival through the late Achaemenid era of the earlier practice of arming cavalry with both bow and spear, which Herodotos records for Persian and similar heavy cavalry, and also for Bactrian lights. Since the Persian cavalry themselves had now discarded the bow, this man is almost certainly an eastern lranian, very likely a Bactrian. Whatever factors led thc Persians to drop the bow may not havc applied in thc eastcrn provinces, where nomad horse-archers were the usual enemy. Figu res dressed like this appear on gold plaques found in Bactria.

14.

BABYLONIAN LEVY ARCHER

The Babylonians who provided some ofthe rear li ne of poor quality infantry at Gaugamela will have been those who held "bow land" in exchange for militia service. Some were lranians or other non¡Babylonians by the 4th century, but this figure represents the dress ofa native Babylonian, who might well be mixed with men dressed like figure 8, and perhaps in a variety of other costumes too. The cap, long tunic, sandals and bracclets are from the Persepolis reliefs. Some show a longer ankle-length tunic, and ali show a robe draped over the tunic, but both these would probably be too restrictive for wear in battle.

12 13

91

15.

GREEK HOPLITE

This is the typical appearance ofthe Greek hoplite at tbc start ofour era; he comes from thc early 4th cenrury Nere id monument from Xanthos in Lykia, and is probably a mercenary. This sryle of equipment was worn through tbc 4th and 3rd cenruries, but during the 3rd was gradually replaced, at least among the major powers ofGreece, by new types ofinfantry. After 220 Lhere are few clear references to the traditional hoplite, though he may have remained in use by some ofthe minor powers.

Weapons ha ve not essentially changed since the 7th century. The main weapon is a thrusting spear eight 10 nine feet (2.4-2. 75m) long, with a cornei wood shaft, iron head, and iron or bronze buttspike, which could be used 10 plant the spcar in thc ground (hence the frequent use ofbronze, which would not rust) orto fight with ifthe spear broke. The centre ofthe shaft is bound with cord fora secure grip. The spear was usually used in the overarm thrust shown, but it could be thrust undcrarm or even, in desperate circumstances such as just before breaking in flight, thrown. The underarm thrust was awkward in dose formacion as it would be hindered by the hoplites' closely-packed shields, but may ha ve become more common in the more open hoplite warfare ofour era. The sword, slung on a baldric, has a leaf-shaped cut-and-thrust biade about 2 feet /60cm long; it is very much a sccondary weapon. The "Argive" shield (called aspis or occasionally hoplon) is 32-39 inches (80cm-l m) in diameter, with a distinctive broad, flat rim, and the rest ofthe shield somewhat convex. lt is ofwood covered with leather, with the rim always, and the whole front ofthe shield often, covered with thin bronze sheet. lt is carried by a distinctive arrangement oftwo handles, a bronze porpax in the centre through which the forearm passes and a cordami/a beat the rim. The extra cord looped round the inside can be slung over the shoulder to carry the shield. This man has a leather apron attached to his shield, as an extra defence against missiles. It is attached to a baton fixed usually to the outer face ofthe shield as in I 5a, but occasionally to the inside. At this time the apron seems only to be used in Asia, where arrows wcrc a commoner threat than in Greece.

This man has thin bronze greaves, clipped round his legs by their own elasticity, which are no longer uni versai for hoplites but are stili common. They are moulded to re present the muscles ofthe lcg, and sometimcs show a rcd cloth lining turned over the edges. His bronze helmet is oftbe Accic rypc, with bingcd cheekpieces and a horsehair cresi held in a decorated crest-box. Crests were most often white, red-brown or black, from natural horsehair, but could be dyed. Alternate sections could be oftwo or more colours. Since hoplitcs generally provided their own equipmcnt - though mercenary employers would sometimes outfit their men - equipment was rarcly very uniform, and

92

helmets seem to have varied more than most gear. Under their he lmets many men wore a headband Like l 5b, which restrained the hair and may have provided some support for the helmet. Alternatively a padded cap migbt be wom.

Costume is a woollen tunic, cl1iton; ehis man's is longer than usual, though not uniquely so. Some forces used uniform crimson tunics {see note to next figure) bue otberwise colours would rarely be uniform. In ci vii !ife, gentlemen prided themselves on the whiteness of their clothes, but this is unlikely to have been practical in wanime.

Colours used included red, terracotta, brown, grey, black, green {including shades called "frog-colour" and "the colour of an un ripe grape") and saffron yellow. Blue may have been uncommon, bue was cenainly used. T unics usually had bands ofembroidery at the edges, and were sometimes panemed ali over. Athenian tpheboi are variously said to have worn black cloaks, and to have wom yellow; perbaps tbeir tunics were yellow.

Over his tunic be wears a non-metallic corselet, the commonest style ofhoplite armour since the end of the 6th century. Modem writers have variously identified it as leatber or as stiffened linen. Ancient writers frequenrly refer to linothorakes, linen corselets, sometimes coupled with spolades, wbicb seem to be ofleather; the linothorax and tbe spolas are probably linen and leather ve rsions, respectively, of this cuirass. They could be reinforced with metal scales, but this is less common now than earlier. T he cuirass wraps round the body, tyi ng under the left arm (or, some vase-paintings suggest, occasionally down the fronr) and is split below che waist into strips called pteruges {feathers) for case of movement; a second layer ofpteniges is fixed inside the firsc, covering the gaps becween them.

A U-shaped yoke is fixcd across che back and che arms ofche U are broughc forward over the shoulders and laced down on the chesc; the shoulder pieces have decorated bronze finials where the laces are attached. l 5b shows tbe back view; note the rectangular flap which stands up to protect tbe back ofthe neck. T he cuirass is usually white, often decorated in red and black, with che shoulder-yoke and waistband sometimes in a darker colour.

I 5c shows a variant, also from che N ercid monumenc, with an undivided skirt instead of pteniges. I 5d is a jerkin of some flexible materiai, with scallopcd edges at the bottom and perhaps the sleeves (or sleeveless, with shon tunic sleeves showing?). h is from the early-4th century tempie of Apollo at Bassai in Arkadia. The helmet appears to be an Attic variant withouc a crest.

16.

SPARTAN HOPLITE

This is a reconstruction from several sources. Tbukydides and Xenophon say Spanans wore the spolas; so the appearance ofthe usual corselet, as shown bere, on Athenian reliefs sbowing trophies ofSpartan arms, confirms that this is tbe spolas. H eadgear is a pilos, characteristic panicularly ofSpanans and Arkadians, but also widely used in tbe rest ofGreece, soutb ltaly and Lykia, thougb not apparently by the Macedonians. T he name means felt, and it seems to bave originated in a felt cap, shown on Greek vases worn in dvii dress long before it emcrges, about 430, as military wear. Bronze caps were soon made to the same shape, and examples such as I 6a, from an Athenian vase, with sharply-outturned brim, are clearly metal, thougb felt versions continued to be worn in battle. The pilos was occasionally crested and/or fitted with cheekpieces, especially in ltaly and Lykia {see l 6b, a south ltalian statuette, or 21). What had originally been a lightweight substitute fora helmet had evolved into just another helmet type.

This man has a beard, probably uni versai in Sparta and usual over the rest ofGreece till Alexander set a fashion for sbaving, and he has long plaited hair, which seems to have once been tbe universal Greek custom but was now exclusive to Spana. H is tunic is charactcristically dyed dark red or crimson, and out of baule he would wear a crimson cloak. T his became quite a common colour for G reek soldiers, worn for instance by Xenophon's T en Thousand, by tbe Achaians under Pbilopoimen, on Boiotian grave monuments and Italiote Greek vases. Ptrioikoi as well as Spartiates wore red, and once at least (at Koroneia in 394) Spanan allies and mercenaries al so wore red, but this army bad been campaigni ng together for some years, and tbe usual sbort-term allied conringents raised for one campaign are unlikely to have been thus uniformed.

Most Spartans used a very short sword, vinually a dagger, shown on one reliefwith a biade !inie more than a foot (30cm) or so long (16c)- allegedly because Spartans fought at such dose quaners. 1t was occasionally used outside Spana, by mercenaries, and was popular with sword-swallowers. Another Spanan weapon was the xuele, apparently a sickle-shaped knife, but this was probably nor used in battle, and seems to bave had a ritual significance.

93

a

e

94

17.

UNARMOURED HOPLITE

This sryle oflightly equipped hoplite emerged during the Peloponnesian War; che equipment is widely depicced but this figure is based chiefly on che stele of Lisas, from Tegea in Arkadia, who died in che late 4th cenrury in Anika. He may have been in mercenary service, as there is no record of Arkadian state forces there at the time. h is unlikely that entire armies, or even units, were equipped in this way; in the Spartan army the younger hoplites of each mora were used for mobile operacions, and could for instance dash out ofthe ranks to chase ofT skirmishe rs. h was probably these who were this lightly armed. The Nere id monument shows unarmoured hoplites mixed in with those armed as figure 15, suggesting a similar arrangement.

His runic is an exomis, which leaves the right shoulder bare. Originally a lower-class working garment, it was adopced by those who led a si mple life, or wished co give that impression, and became qui te popular with soldiers.

Other unarmoured hoplites wore che c/1iron, sleeved or sleeveless; l 7a shows a panerned rype worn by hoplites on Athenian vases, apparently ofheavier materiai than usual.

The felt or bronze pilos is usually worn by these troops, but some Athenians are shown in Attic helmets, while one Theban relief shows che Boiotian style (see figure 25). These lightly anned men are common al so in Italiote Greek are, often wearing the pilos; if not, ltaliotes might wear Attic helmets, occasionally feathered in native ltalian sryle ( 17b is from T arantine sculprure) or a debased version ofthe old Corinthian sryle, worn on che back ofthe head, usually called Etrusco-Corinthian ( l 7c, from a 4th-cenrury vase).

18.

HOPLITE lN MET AL ARMOUR

Metal body-armour in ibis period was generally worn only by cavalry, officers, and occasionai elite units. Thus Dionysios ofSyracuse, equipping a large mercenary force in the 390's, provided corselets only for officers, cavalry, and a picked infantry guard. Agathokles later had a similar guard, and these, and other crack units, may well have been similarly equipped. In ci tizcn forccs thcre may have been individuals who felt that the protection ofbronze armou r justified the cost and weight; this figure is from the monument of Aristonauces, a late 4th cencury Athenian who may have been one such.

His armou r is a bronze "muscled" corselet, breasc and back plates shaped to represent che muscles ofche torso. Of necessicy ic was closely and exactly fittcd 10 che individuai who was to wear it, which added to the cost. lt fastened down the sides with hinges and buckles, and at the shoulders with straps similar to those ofthe linen or lcather cuirasses; this appears co be a 4th-cenrury innovation, as the earlier muscled cuirasses were fastened at tbc shoulders only with pins. A variant had small metal shoulder-pieces hinged directly to che backplace, decorated in reliefwith a thunderboh or similar. T he lower cdge ofche cuirass covers the abdomen, curves up 10 clear che hips for case of movement, and in some examples curves down again at che back. Preruges were usually worn at the waist (though often in ltaly, and sometimes elsewhere, they were dispensed with), in this case three layers. Pteruges ofthe short style shown here could also be dccoraced with embossed bronze plates. They were probably attached to a soft, perhaps padded, undergarment, as la1cr Imperia! Roman pteruges cercainly were.

Aristonautes' helmet is ofthe cype called Thracian, ofwhich severa! styles were common in the 4th century, and which was che ancestor of most later Helleniscic scyles. lt is distinguished by a peak 10 shade and protect che eyes, whilc che skull is often high and crested, che mosc distinctive rype curving up like a cap, as bere. The cheekpieces usually found wich Thracian helmets are longer than ocher types, dosely shaped to che face and poinced ac the chin, somecimes embossed with a mouscache or stylised curls to represenc a beard. I Sa shows a low-cresced variane from che tempie of Apollo at Bassai, with less distinccive cheekpieces; 18b is a type with horsehair crest and feathers, from a 5th cc ntury Athenian vasc. Feathers, chiefly ltalian decoration, are occasionally found in Greece; Philopoimen's Achaians in the 3rd century wore thcm. T he decoration on this helmet could be embossed or painted.

The cuirassed Aristonautes could perhaps have been an officer, but he lacks the triple-crested helmet whkh, we learn from remarks in Aristophanes' Peace and Acl1amians, was, with a crimson cloak, che emblem of Athenian taxiarc/1oi. These were probably Attic helmecs ofthe rype shown in J8c, worn by che city's patron goddess Athcne on coins and clsewhere. Emblems of rank do not, however, seem to bave been common in Grcek armies, chough we do hear of an Aitolian epilekrarchos in a triple-crestcd helmet.

The sword shown is the machaira or kopis, a single-edged curved slashing weapon, quite common as an alternative to the straight sword. h was characteristically uscd in an overhand slashing stroke brought down from above one shoulder, and this is thc pose shown by Aristonautes' monument.

95

19.

GREEK HOPLITE SHIELD BLAZONS

T he designs on hoplite shields were painted, either on a painted background or on tbe polished bronze facing ofthe shield. Some earlier blazons cut from thin bronze sheet have be.en found, but these were probably for ceremonial

96

pu rposes or dedications, not for battle. On Spartan shields the background was always polished bronze. Whi1e painted shields were tracli1ional a1 Argos, and common elsewhere. Some ofTimoleon's mercenaries had expensive purple pain1ed shields, inlaid wi1h gold, ivory and amber. From the late S1h cemury severa! states, though not apparently all, adopted uniform blazons for ease of recognition, in some cases 1he initial letter ofthe s1ate's name. (i) is Sparta's blazon, the letter lambda, initial ofthe city's alternative name Lakedaimon. Sparta may perhaps bave introduced the use oruniform blazons, and encouraged her Peloponnesian allies to adopt them. Cenainly two ofthe relatively small number of known state blazons belong to Peloponnesian cities allied 10 Spana in the early 4th century; (ii) and (iii), the sigma (lettcr S) or Sikyon and the 1riden1 or Mantineia in Arkadia. Both cities joined the Achaian League in the 3rd century, though whether League members kept their distinctive blazons is unknown. (iv) is the initial ofSparta's old enemy Messene. (v) is the Theban blazon, the club orHerakles, which at Mantincia in 362 was also carried by Thebes' Arkaclian allies. (vi) may be another state blazon. It is a monogram composed or thc letters lambda and alplia (A), the first rwo letters or the city La rissa Kremaste, on wbose coins it appears on a shield carried by a nymph. It may perhaps have been confined to coins, but could have been carried by Larissaian troops. (vii)-(xi) are from Athenian vase-paintings, blazons at Athens apparently remaining a maner of individuai choice.

The bull's head (vii) is the emblem orPhokis on coins, and Silius Italicus mentions a cavalryman from Akragas in Sicily in the second Punic wa r with a bull blazon. The gorgon (viii) was especially popular at Athens as it was the goddess Athene's blazon. (xii)-(xvi) are from south Italian Greek art. The star or(xi) and (xii) is the same style used

by Macedon (see figure 40); it was quite common in Greece and very popular in Italy. Geometrie patterns like (xiii) and (xiv) were also common in Italy. (xvii) is an oddity, from the i11side or a shield carried by a 4th-cenrury Theban hoplite, Rhynchon, on his gravestone. (xviii) is from an epic scene on 4th century Greek, probably Olbian, goldwork. (xix) is from a clay model shield from Thessalonike. (xx) is the only known shield to bave been decorated with words, rather than initial leners. It is the blazon carricd by the Athenian orator Demosthenes at Chaironeia, agatlie ryclie, "with good fortune", inlaid in golden letters.

20 and 21.

GREEK MERCENARY PELTASTS

Peltasts were lightly-equipped infantry, who were nonetheless better armed than the ordinary skirmishers, and they had emerged as the characteristic mercenary infantry or the 4th century Greek armies. Their style of fighting originated in Thrace and the first Greek peltasts were recruited from Greek cities orthe Thracian coast. By now they were or more diverse origins. The peltast took his name from the pelte, a light shield, and was customarily unarmoured and armed with javelins and sword. Xenophon 's peltasts at Cunaxa in 401, for instance, had no other weapons; they used swords, not spears, against charging Persian cavalry. The Athenian mercenary leader Iphikrates won the greatest peltast victory by destroying a Spartan mora outside Corinth in 390 with javelin fire, and again there is no hint that bis men had any other weapons. But thereafter the character ofthe peltast becomes less clear, as Diodoros and Nepos ha ve accounts oflphikrates re-arming his men with long spears (perhaps up to 12 feet/3.6m long), swords, pelta i and linen armour. This has been variously explained; as a lightening or hoplite equipment, as the sources imply; as a reform standardising and increasing the weight of peltast equipment; or as a complete fallacy, J. G. P . Best suggesting there was no such reform and thnt ali the gear or the "Iphikratean" infantryman was already used by the peltast. This seems unlikely as the linen armour was not characteristic ofthe peltast, but certainly there is no evidencc for any new or reformed type or infantryman acrually seeing service.

Perhaps Iphikrates' plans were never put imo effect.

Thrusting spears are certainly included in some illustrations of peltasts well before Iphikrates, and some peltasts may have carried them instead or, or as well as, javelins. There seems no evidence that they were commoner ar1er lphikrates than before, or that the very long type implied in accounts ofthc "reform" was ever used. No bartle accounts speak or peltasts using spears, so though their existence cannot be denied they may nevcr have been standard equipment, merely an addition to the peltast panoply carried by individuai choice. The tlmreoplioroi who later took over the pehasts' role (see figure 41) cenainly seem to ha ve carried both spear and javclins, so the pracĂšce may have been esiablished before the adoption ofthe tliureos.

The absence of 4th-cemury illustrations of peltasts prevents a definite statement about their appearance and armamen1. Figure 20 is based partly on carlier illustrations, whence comes his Attic helmet. Some orthesc earlier peltasts wcar greaves; it is unlikely that 4th-century peltasts would want to be bothered with them when even many hopli1es we re discarding them. The boots he wears may be the iplrikratĂŹdes, light and easy to untie, said to be introduced by lphikratcs.

97

The pelte is usually shown in art as crescent-shapcd, though some literary refe rences imply round ones. lt was made ofwicker, sometimes left uncovered to judge from vase-paintings, but usually covcred with goat or sheepskin, or occasionally with polished bronze. lt could be carried by a centrai armstrap and a hand le at the rim, as shown, in which case the armstrap is often shown as indistinguishable from thc hoplitc's porpax, or by a single centrai handle.

1t sometimes had a baldric, as Thracians slung their pelta i on their backs when evading an enemy charge. 20a-f are pelte blazons. The first rwo are typical ofthe simple faces shown on pelta i carried by Thracians. 20fis the only one which acrually dates to the 4th century, the others being earlier.

Figure 21 is raken from the Lykian sarcophagus of Payava, about 400 BC. Hc has exomis, pì/os with crest and cheekpieces, and a round pelte. He is thrusting overarm, probably with a long spear. With the addition of linen armour, be might look very much like the accou nts oflphikrates' supposed new troops.

22.

GREEK LIGHT JA VELINMAN

Javelinmen were rec ruited chiefly from the poorer, hilly areas ofGreece, such as Aitolia, Akarnania, and the srnall states round the Thessalian plain. T he poorer classes ofmore civilised states would sometimes fight as light troops, and most of these would throw javelins or stones, as they would lack the specialist training to use bow or sling.

Hoplites' servants can sometimes be found fighting in this role. However, the use ofsuch improvised troops seems to have declined wit h the introducùon of mercenary peltasts and other professional light troops. This figure is based on the 3rd-2nd cenrury coinage ofthe Ainianes, a small bili state. Apart from javelins he carries a light round shield with a metal boss, perhaps a round pelte (the cresccnt typc could be uscd too). The charactcristic shicld of Aitolian javelinmen was the sakkos, convex and probably round, made ofwicker covered with leather. This Ainianian also carries a sword, probably rare for light troops.

The ha bit offighùng with the right foot bare (to improve the footing on slippery ground)and the lcft shod - thc lcft, which led in baule, possibly being chosen as bcner for kicking the enemy - is an Aitolian custom recorded in a fragment of Euripides, and occasionally mentioned for othcr nationaliùes too. Poorer men such as thcse would gcncrally wcar tunics in the various greys and browns of unbleached wool, or instead wear a shccpskin tunic Pausanias rccords these still bcing wom by the poor in Euboia and Phokis in the Imperia! Roman era. Earlier sources also mention countryrnen in leather jerkins.

Some of Alexander's javelirunen were issued with axes at the Hydaspes, to hamstring lndian elephants. Axes like 22a were used for hunting in Macedon; this one is shown on a mosaic from Pella, swung two-handed against a deer, while 22b is used, also in two hands, by humsmen on the Alexander sarcophagus. h was presumably the established use oflarge axes for hunting which inspired their use 10 hunt elephants on the battlefìeld.

98

23.

GREEK SLINGER

This man is based on 4th-3rd cencury coins from che Greek ciry of Aspendos in Pamphylia. He is abou110 release 1he pouch ofhis sling and whirl ii abou1 his head 10 bui Id up speed. The sling seems surprisingly long, bue Ainianfan coins show them if anything longer. Perhaps these were only for long-range shooĂšng (see note to figure 119).

Ammunition was ovai lead bullets usually weighing 0.7-1.05 ounces (20-30 grammes) bue occasionally up 10 1.9 ounccs (55 grammcs). Xenophon reports that these easily outranged the large stones shot from Persian slings. Livy mcntions Achaian sli ngers prac1ising wilh seashorc pebbles, though they probably used lead sbot in battle. He says thcy were more accurate 1han 1he famous Balearic islanders. h is some1imes argued !hai slingers, unlike arcbers, could use shields to protect thcmselves while shooting; but while the use ofthe sling does not arguably restrict the use ofa shield as much as docs a bow, a shicld can have been little use toa slingerwhile loadingor,as bere, preparing to shoot; and thcrc seem to be no contcmporary illustrations of sbielded slingers, tbough one 5th-cemury painting shows a slingcr protccting himsclf with an animai skin draped over his left arm. One Ainianian coin sbows a slinger who al so has a pair of javclins, but while such unonhodox weapon combinaĂšons may occasionally bave been used, they were not common.

The slingers of Rhodes were traditionally the best in G reece; othcrs carne from hilly areas such as Achaia and Akarnania, or from the rural pans of other states - Elis levied slingers from her subject hill-towns.

21

22 a

a

b

e

99

b

24.

GREEKARCHER

Mosc Greek archers were mercenaries, che Cretans being by far the most famous. Tbey were also noted pirates, and specialised in ruses and ambushes. lt is of course possiblc that "Cretan", like "Tarantine" (see figure 44) carne to indicate a style offighting rather than a nationality, and that many Cretan units in H ellenistic armies were actually composed of mercenaries of mixed ori gin armed in the Cretan style. Ouring the 3rd and 2nd centuries increasing numbers of cities began to train their young citizcns in archery, sometimes hiring Cretan inscructors, and inscriptions recording archery contescs are numerous.

This figure is a typical Greek archer based on an Athenian vase painting ofthe early 4tb cenrury, unarmoured and carrying a recurved composite bow. 5th century Lykian paintings show Greek archers in white ruoics and red leather belts and quiver, barefoot and bareheaded. The quiver could be slung either ove r the back, or at the left bip.

Tbe pose ofthe figure, taken from an earlier vase-painting, shows how inconve nient it was to draw an arrow in this position. Arrows were of reed, usually with bronze heads, probably because bronze arrowheads were easy to cast in large numbers, but occasionally with iron. A large, tanged, barbed arrowhead was typically Cretan (22a, about 4 1/4" / l lcm long) but shoner, unbarbed, socketed types were also used (24b, three-edged, 11/4" /3. Scm, and 24c, rwoedged, 2" / 5cm, both from Olynthos).

Some archers seem to ha ve been better protected than this figure. On the Nereid monument one archer in a linen or leatber cuirass kneels beside an unarmoured colleague, and othe r sources show cuirassed or belmeted archers.

Tbere are also three references to Cretans with shields, in Xenophon's Ten Tbousand (bronze-faced pelta1), in Seleucid service in 210 (aspidiota1~ a diminutive of aspis, probably indicating small round shields) and with Perseus at Pydna (Cretan peltai appear in the Roman triumph). The shields suggest tbese men were prepared 10 fight hand to hand, in which case they probably carried swords too. Alexander cenainly sometimes used bis Cretans in circumstances where they would have found such equipment useful. lt has been suggested that these sbielded Cretans may not have been archers, and that they may be the rather mysterious Neocretans wbo occur in Aitolian, Seleucid and Ptolemaic service. But there is no positive evidence to associate Neoc.retans with shields, or indeed with any distinctive style of equipment. Shielded Cretans serve with Xenophon and Perseus when there seem to be no Neocretans in these armies, and when those Cretans wbo are present are specifically called archers. lt seems clear, then, that Cretan archers could carry shields, though not clear whether they ali did. Ifshield and sword were standard Crecan equipment, however, making Cretans much better able to look after themselves at dose quaners than ocher Greek archers and slingers, this might help explain tbeir popularity. It would also explain why we hcar of Achaians, Apollonians and orhers "armed in che Cretan style" ratber tban simply being called archers, if"the Cretan style" meant more than carrying a bow.

As for the Neocretans, those in Aitolian service in 218 at least seem co have been real Cretans, sene by Knossos to help her Aitolian allies, which disposes ofthe possibility that these "new Cretans" may bave been foreigners in Cretan equipment. The most likely explanation seems to be chac tbey are neo1~ members ofyoutb associations that existed in Cretan cities; 11eoi Kreces would easily become 11eokretes. Perhaps tbe members ofsucb associations would serve in a body in war. Their identifĂŹca tion simply with bodies ofCretan young men makes it seem less likely that they were armed in any way distinctive from other Cretans, and cenainly there is no evidence for it.

25.

GREEK HEA VY CAVALRYMAN

The commonest G reek cavalry type. He is based chiefly on a 4th or 3rd century relieffrom Thespiai in Boiotia, but similar equipment is shown in other sources. He wears tali boots ofan originally Thracian style now popular with Greek horsemen, and could wear a cloak. Body-armour is a bronze muscled piace corselet. The type worn by 18 is shown in use by cavalry on a coin of c.334 and an early 3rd century sculpture from Taras, and on a coin ofAlexander of Pherai of c.360, but it was not really suitable for cavalry si nce tbe extension over the abdomen made it awkward to bend at the waist and 10 sita borse. This figure's cuirass is one attempt to coumer tbis problem; it flares out widely at about hip level at both front and back, to allow freedom of movement. The few surviving examples ofthis rype are ali from south ltaly, but tbc Thespiai reliefshows it and confirms its use in Greece. While some cavalry may have worn linen or leather armour, it is clear tbat cavalry cuirasses were usually heavier than infant ry.

His helmet is the Boiotian type, resembling a hat witb the brim folded down, made in bronze. lt could be worn with a trailing crest as bere, or crestless as in tbe next figure. This was a very popular cavalry headpiece in the 4th century and at least the early 3rd, worn by tbe horsemen of Atbens, Boiotia, Thessaly, Macedon and Ptolemaic Egypt, at least. In Boiotia and possibly Tbessaly it was also wom by infantry, but was not popular for them elsewhere.

Weapons are two javelins, about 6 feet (I .8m); Xenophon recommended thc Pcrsian corncl-wood palea. Altbough cavalry often skirrnished there is no evidence for them carrying more javelins than these rwo. An alternative weapon 100

was a thrusting spear the same length as the hoplite's, or perhaps a little shorter, but apparently flimsicr judging from its tendency to break on impact. At this date cavalry did not carry shields. Xenophon, in his work On horsemanship written shortly beforc the start of our period, recommends much ofthis equipment - corselet with pteruges, boots, Boiotian helmet, javelins - and also the kopis rather than the straight sword. He also favoured extra armour for the borse and for the ridcr's arms and legs; this was not widely adopted, however, though a few riders adopted the parapleuridia shown in figure 5.

26.

GREEKUGHTCAVALRYM AN

Hc is similar to the last figure except that he wears no armour. Severa! Athenian monuments show 4th century cavalry in tunics only, sometimes with cloaks, and either bareheadcd or, more often, in Boiotian helmets.

Aristophanes' Lysisrrata mentions an Athenian cavalry officer in a bronze pilos. Earlier, Sth century vase paintings show similar riders in tunic, cloak and high boots, but mostly wearing the petasos, the hat shown in 26a. This is common headgear for huntsmen and travellers, and could perhaps have been worn in war by other helmetless troops too. There are unarmoured riders on the Nereid monument, one in a petasos, and in other sources, while another 4th century Athenian wears petasos with muscled corselet.

27.

THESSALlAN CAVALRYMAN

Earlicr Thessalian cavalry shown in 6th and Sth century art are unarmoured, with petasos and javelins, like the Athcnian light horsc mentioned abovc. But aftcr tbc mid-4th century representations ofthese cease andare replaced by hclmcted men, of whom those few who are depicted clearly enough for us to be sure are wearing armour indistinguishable from that of other Greek or Macedonian cavalry. Thessaly seems to have become markedly wealthier in the late Sth and 4th century, which is probably the explanation for the increased use of expensive armour. This man is based on a relieffrom the Thessalian city of Pelinna, ofthe mid 4th century. He has a bronze cuirass reaching only to the waist, with shoulder straps and preruges. This represents another attempt to make the muscled cuirass suitable for cavalry; it was apparently successful as this short-waisted type became standard for later Hellenistic armoured cavalry. His helmet has a Thracian skull with Attic style cheekpieces and forehead pla1e.

The Boiotian style was probably commonest for cavalry, and vague coin representations from Pelinna seem to show it worn by Thessalian hoplites too. 27a is an interesting variant from coins of Skotoussa, a Boiotian helmet wi1h cheekpieces.

Weapons seem to bave been the standard Greck choice ofjavelins and spear; Diodoros describes Thessalian cavalry using javclins in 368, while coins show thrus1ing spears used underarm.

101

28

102

28.

PHILIP Il OP MACEDON

This reconstruction is based on equipment from the rich tomb excavated in 1977 by Professor Andronicos, in the great tumulus at Vergina. Vergina is now accepted as Aigai, the originai capitai ofMaceon, and the tomb had been convincingly identified with that of Philip. lt contained some ofthe finest Greek arms and armour yet found.

The helmet is iron, photographs suggesùng uaces of gilding. lron was used increasingly for annour after the mid-4th century, but bronze remained commoner beyond the end of our period. This helmet combines Thracian and Attic features, like that ofthe last figure and severa! later Hellenistic types. There seems no sign offittings for a horsehair crest atop the high flat iron comb.

The remarkable cuirass consists ofthin iron pla tes, four for the body and two for each shoulder, hinged where they join. The backplate has a rectangular extension for the back ofthe neck, but the breast plate leaves the upper chest uncovered. The iron was covered inside with cloth, outside with cloth and leather. This would make its external appearance indistinguishable from that ofthe common linea or leather cuirasses (the covering would probably be white, like these) ofwhich it is presumably an imitation, and this raises the question whether any ofthese may have had internal metal reinforcements, not visible in art. In particular, some cuirasses are shown with scales down the sides but not the front, as in figure 2; and it now seems likely that tbis crucial area may have been protected by a concealed metal plate rather than left unreinforced. However, the cost and complexity of even a simpler version of Philip's iron corslet suggests it would never have been common among the rank and file. The cuirass had narrow gold band with relief decor:nion at all edges, and a broader band around the cenere ofthe body. On the front were six small gold lions' heads, four with rings througb the mouths to hold the laces which fastened down the shoulder pieces. The pceruges were decorated with 56 gold strips (the number indicates two layers) decorated in relief and ending in palmettes. The result, in white and gold, must have been very impressive.

Four pairs of greaves were found; one pair, the left greave an inch shorter than the right to accommodate Philip's limp, was gilded. The open-toed boots shown are taken from a hunting scene painted on the tomb's facade.

No reconstrucùon of the unique elaborate shicld has been publishcd at the time of writing. h was of traditional hoplite style, with a plain bronze cover to protect it. The shield's outer rim was decorated with a spi rai meander pattern in ivory and dark glass, covered with gold leaf. lnside this was another band of gold and silver, apparently pattemed. The centre ofthe shield was gilded, with two ivory figures attached, a youth seizing a maiden probably in a mythological scene. 28a shows a highly tentative reconstrucùoo. Inside the shield was a large rectangular gold plaque decorated with two pairs oflions, to which the porpax was probably attached. Arranged around it, perhaps in a cross, were four golden bands bearing victories in relief and ending in palmettes. Whether this ceremonial shield indicates the use ofhoplite·style shields by some Macedonian troops in action is unclear; perhaps this was the size ofsbield used by the cavalry when dismounted.

Weapons included several long spears, exact length uncertain. Another tomb in the tumulus contained a spear with iron head and buttspike and the shafl completely encased in gold. Javelins were also found. Philip's sword was ofthe straight type, with the guard covered in gold and two gold bands on the pommel. The scabbard is wood and ivory.

29.

ALEXANDER THE GREAT

This figure is based chiefly on the Issos mosaic, copied from a contemporary painting. Equipmcnt is basically that of the Companion cavalry. Alexander wears a white or pale grey long-sleeved tunic and light purple cloak. His composite cuirass is white, with fringcd pteruges at shoulder and waist, each decorated near the end with two narrow rcd stripes and one green between them. The cuirass is reinforced with iron scales about the waist and what see ms to be a rectangular iron plate on the chest, bea ring a gorgon's head. The shoulder-yoke may also be iron, but it is not too clear and may ali be red leather, as is its edging. The shoulder pieces bear a white thunderbolt moti f. The whole cuirass compares incerestingly with Philip's. Round the waist is knotted a narrow green sash edged and fringcd in yellow and black, hencefonh a common cmblem of officers' rank. The high sandals here are taken from a bronze statue at Naples, also believed to be a copy of a contemporary work. The statue shows Alexander in a short plate cuirass with pteruges.

Both these representaùons show Alexander bare-headed; he shaved, an innovation widely copied especially among the you nger generation. The helmet here is taken from a coin showing Alexander fighùng Poros, with on the reverse a figure of Alexander standing. Both versions on the coin show what is clearly revealed in enlarged photo· graphs as a high-crowned Thracian helmet with crest and vertical ornaments, apparently standing feathers. This

103

may be the sryle Alexander wore ar the Granikos, which had a whitc fcathcr on each sidc. At Gaugamela his helmet was polished iron, with a gorget attached; two types of gorger are known, that worn by figure 71 and a higher type recommended by Xenophon, which could cover thc face up to the nosc, but neirher ofthese seem co have been auached co the helmet. Alexander's Gaugamela gear also included a captured Persian linen cuirass, Sicilian tunic, and ornate Rhodian cloak. T he helmec he is shown wearing on the Alcxander sarcophagus, made a few years after bis death, is shaped like a lion's head; traces ofyellow paint suggest it was polished bronze or gilded. There is no trace of fìtùngs for crescs.

Weapons on the mosaic are a shon scraight sword on a red baldric and a long spear, the xys1011 (see next figure). The horse has a bridle of gilt or bronze plaques; what is visible of the saddlecloth is yellow wirh black markings perhaps representing leopardskin. h seems on the mosaic 10 be knoned at the front, but the rest ofthe outline shown bere is restored from tbe Naples starue, tbe mosaic being damaged.

30.

MACEDONIAN COMP ANION CAVALRYMAN

Tbis would be tbe appearance of a heavy cavalryman under Philip and Alexander and into the 3rd cenrury. He is based cbiefly on figures from che Alexander sarcophagus, with details from the Issos mosaic, the monument to Alketas (kiUed in Pisidia in 320) and a late 4th century grave painting from Egypt. Most ofthe last figure's equipment could be subsciruced. Long-sleeved tunic, cloak and shorc boots are from the sarcophagus, though other sources show short sleeves. The Egyptian painting shows a white tunic (as does another of an unarmed horseman, and compare Alexander's own) and red-brown cloak. A figure behind Alexander on the mosaic has a sandy-brown cloak, wbile Dionysios ofHalikarnassos mentions the dun cloak of one of Pyrrhos' cavalry. White runics and some son ofdull brown cloak may chus ha ve been uniform. Diodoros says Alexander issued his Companions with purplebordered cloaks after he took Persepolis; a tomb painting of chis era shows a Macedonian officer, perhaps of infantry, in red runic, white linen cuirass, :md sandy cloak decorated in purple, wbich fics admirably with the preceding refcrences.

Sarcophagus and mosaic show Boiotian helmets, one apparently painted wbite witb a golden wreath painted on; painted helmets were common (see figure 34). This and a second rider's helmet on tbc mosaic have white crests, which may also have been standard for Macedonian cavalry; a fallen Auic helmet on the mosaic has one (though this could be Persian), as did Pyrrhos, and compare Alexander'~ white feathers. One figure on tbe sarcophagus has a waist-length cui rass with one layer offringed preruges; Alketas has a similar cuirass wbich is certainly muscled pia te, whilc the Egyptian painting shows a bronzc cuirass with two rows of yellow-brown premges ac tbc waist.

Shields are never shown till well after Alexander's time, though he and bis Companions did use tbem when fighùng on foot, to storm walls or once to hold a hill; they were clearly not used in mounted combat. Main weapon was a spear, called either dor11 (the usual word for spear) or xys1011 (once a generaJ term fora spearshaft, but from thls era rarely used except for the spears ofMacedonian-style shock cavalry, who are sometimes called xystophoro1); Arrian equaces che xys1011 with the later long, heavy ko111os. Illustrations of Macedonian cavalry before Philip show tbe usual pairs oflight spears, but Alexander on the mosaic and the Poros coin is clearly using a single thrusùng spear about 12 feet (3.6m) long; al!achment holes on che sarcophagus show tbat the metal spears wbich once adomed ic were of similar length, and long spears are shown in severa! later illustrations. This long spear is clearly tbe xysron ofthe literary sources; it cannot be the sarissa as Markle recently suggested, since Asklepiodotos gives the minimum length for that as I Ocubits (I S feet/4.57m), while the bunspike surviving on tbe mosaic aJJows usto fix tbc length of that spear with certainty, though much of the shaft is missing. lt is gripped well back of the middle, so the buttspike may have been heavier than the head, for balance. The buttspike bere is based on a bronze example in Newcastle University inscribed "MAK" for Macedon. At 1Sinches (38 cm) it is longer than the usual hoplite spearbuu, yet shorter than the massive sarissa butt offigure 32, so may very well belong to an intermediate cype like this. The shaft of the xys1011 was of cornei wood - cornus mas, the cornelian cherry, apparently common in M acedon. lt was far more elfective fora charging cavalryman than the Persian javelins, in contrast to Xenophon's stricrures about earlier, flimsier thrusù ng spears. lt was stiU prone to breakage, and tben tbe buttspike would be used - it was preferred, it seems, to the sword by some, though Plutarch says Alexander relied on his sword as his principal weapon. Both st raight sword and kopis could be used.

31.

MACEDONIAN PROOROMOS

This man is from drawings of a lost· Macedonian tomb painting of about 300. He is riding down a Persian, so is presumably a veteran of Alexander's wars in Asia, wbile being unarmoured but carrying a long spear is probably from the prodromoi or sarissophoroi. His Thracian helmet is crimson, clearly painted, except for the polished bronze 104

31

fitting over the peak. Colour reproductions difTer in shades, but he seems to wear two runics, the outer one pulled up through the beh to shorten it and greyish in colour except fora pink shoulder-yoke and pale yellow sleeves. Skirt of the lower tunic is pink, cloak crimson-pink fastened with a purple ribbon. Beh is ydlow-brown, shoes dark brown, saddlecloth a skin, perhaps a lion, with the tail attached.

His spear has an iron buttspike shaped like a narrow spearhead; notali the shaft is visible and it may be roughly the length of the xyston. A sarissa should be longer, so the name san"ssophoroi may perhaps have been a nickname, contrasting these men with other light horse who used shorter weapons.

Alexander's sarissophoroi are now generally accepted as Macedonian, rather tban Thracian as Diodoros seems to say. There are nonetheless references to Thracians with sarissai; one is clearly a later infantryman. The other is a note in the late rhetorician Didymos to a Triballian who, pursuing Philip li after a minor Macedonian defeat, pinned him to his horse with a sarissa. It has been suggested that this Triballian must have been a cavalryman, to catch a fleeing mounted Philip, but this would depend on the terrain, which is unknown. It is thus possible that some Thracians used the sarissa, though the lack ofother references must mean theywere few, and even that Philip copied it from them. Ahernatively, since the word sarissa is probably older than tbe weapon (tbe -issa ending is thought to indicate an old pre-Greek linguistic stratum) it may once bave been a mere Macedonian dialect term for a long spear; so we may be dealing here simply with a Thracian spearman armed with a conventional weapon. This is purely hypothetical, but at least such troops are known from other sources to have existed.

32.

MACEDONIAN PEZET AIROS

There are no clear or certain contemporary illustrations of the equipment of Philip's and Alexander's pikemen.

Polyainos lists the armamene of Philip's as helmet, shield, greaves and sarissa. Swords are not mentioned but will have been carried. No cuirass is specified either, and it is commonly argued on both tact ical and economie grounds that, at least, no metal cuirasses are likely to have been worn. But a 4th-century seal showing a pikeman fighting a Persian cavalryman seems to show the infantryman in a non-metallic cuirass ofthe standard type, as worn by the next figure, or by the red-tuniced officer memioned under figure 30 (whose cuirass is white, with the portions here shaded in crimson). Armour may have varied between taxeis, or the rear ranks may have been unarmoured, since pan of Alexander's phalanx was regarded as more mobile, thus perhaps lighter armed, than the rest. A traditional Macedonian broad-brirnmed hat, the kausia, was worn instead of a helmet at least on the march; Darius' Greek genera! Memnon tried to disguise his men as Macedonians in kausiai, and one of Pyrrhos' cavalry wore a kausia even in bartle. In illustrations it is !inie difTerem to the petasos (figure 26a) though the brim usually rurns up; 32a and b are from Macedonian coins.

The long pike, sarissa, was introduced by Philip; it may be significant for the precise date of its adoption that Macedonian infantry were using javelins against Onomarchos in 353, that no sarissa-heads were found at Olynthos which Philip besieged in 349, but that severa! come from the battlefield ofChaironeia, 338. The shaft was of cornei wood, and a find of metal sarissa fittings from a tomb at Vergina indicates that the shaft was made in two pieces

105

32 a

b

joined by an iron tubular sleeve 61/2 inches (I ?cm) long. This tomb also cootained the bead and buuspike of the

sarissa, both iron, the head 20 inches (50cm) long, the buu 18 inches (45cm) long, witb four "wings". Their size shows they canno1 havc come from anything bui a sarissa. The contemporary wri1er Theophrastos says tbe longest 41h-century sarissai were 12 cubils, 18ft/5.5m long (Tam's suggestion that he means a sborter Macedonian cubie had lite le basis and is now gene rally rejected) while Asklepiodotos says the shortest were IOcubits, l 5ft/4.6m. Wbo used the shoner ones - pan ofthe ptzetairoi, hypaspists, cavalry sarissophoroi - is unknown. A late commentator on che lliad suggests different ranks had different length spears, so all 1he poiots were level with each other in front of the first rank. Apart from its late source, ibere is no supponing evidence for this, it specifically contradicts Polybios' account ofthe later phalanx, and it would make changing from 16 ranks deep to 8, as was done at Issos, rather difficult as men in the same rank would end up with different lengch spears. h is geoerally rejecced, if noc ignored. Markle calculated tbc weight of a 12-cubit sarissa ac about 14.5 lb. (6.5 kg), contrascing with the 2 1h lb., I. I kg (Markle's figure plus allowance fora buuspike, which he ignores) ofthc shorter hoplite spear. This figure bolds his sarissa as Polybios describes for the firsc five ranks, levelled at the hip with the left hand six feec forward oftbe butt.

The sbield used witb the sarissa was distinctively Macedonian rype, called aspis or occasionally (eg in Plutarch's account of Pydna)pt/te, as its size was like that ofthe traditional pelte. Asklepiodotos says ic was 8 palms, 24in/60cm, in dia meter, "not too concave", and made ofbronze. lllustrations reveal the concavity was nonetheless marked, as shown by figure 38. Thc bronze was a thin facing on a wood or leather core; exa mples have been found from Dodona and Pergamon, 1he lauer abou1 251/2 inches (65cm) across. The shield was rimless to lei 1he left hand grip the pike, and was carried by an armscrap like the hoplice shield's porpax, a baldric over 1he shoulder which would have 1aken some ofche weigh! ofthe pike too, and a handle at the rim which might conceivably ha ve been dispcnsed with when holding the pike.

The sarissa was very unwieldy for some oftbe tasks the pezetairoi undenook, such as fasi marches in hilly country or s1orming up scaling ladders. Tbere are in fact many references co Macedonians using javelins when not in picched banle - in naval fighting, executions, or a duci, for ins1ance - and one such instance, at the siegc of Halikarnassos, refers specifically to two men of Perdikkas' 1axis of the phalanx. The tomb at Vergina which contained the san¡ssa finings also held a smaller spearhead, which has been identified as from a hoplite-style thrusting spear; but as it lacks tbe usual buttspike it may perhaps be from a lighter 1hrowing wcapon. Ali this suggests thac tbc javelin, apparenÚy the weapon ofche Macedonian infantry before Philip, was stili used after his reform; tbc infantry would leave their sarissai with 1he baggage and take javelins to scorm walls or chase offinto the bills afcer barbarians. T bis belps to explain the greac mobility and versatility of Alexander's infantry. No other changes in cheir equipment would be necessary, chough perhaps greaves would be left offfor long marches. Onc Macedonian fought a due! with sarissa and javelin, but obviously it would not normally be practicable to carry both weapons at once.

106

33.

MACEDONIAN HYPASPIST

The armament ofthe hypaspists has been long debated with no universally accepted conclusion beiog reached; unfonunately it is never describcd by our sources. Gene rations of scholarly argument have produced two main hypotheses.

One view is that the hypaspists were armed and fought just like the pezetairoi ofthe phalanx. In favour ofthis, our sources do oot dilferentiate their equipment, implying it was the sa me as that of other Macedonian infantry, while the hypaspists' position immediately to the right ofthe pezetairoi in pitched battles is interpreted as meaning they were the righ1-hand unir ofthe phalanx - rhe righr ofcourse being rhe posirion ofhonour and rhus fitting a guards unir. lt has also been claimed that references 10 sarissai in pirched battle at Gaugamela and io Illyria refer ro hypaspists, though they seem iosufficently precise for this to be cerrain. One account of Alexander's killing of Kleiros says the murder weapon was a sarissa taken from a guardsman, presumably a hypaspist.

The alternative view is that their equipment was lighter, closer to that oft he peltasts. In favour ofthis their frequent use oo forced marches, hill campaigns and other mobile operations demaodiog mobility, and the oeed for such mobile troops before Greek mercenaries were available in any great numbers, in the early years of Philip's reign, argue that they were certainly capable offast marching and operatiog in difficult country, and thus may have been lightly armed. But recogniĂšon thai the pezerairoi could discard thcir pikes and fight just as lightly equipped removcs one ofthe main barriers 10 identifying the equipment ofthe two as the same; the hypaspists could be like the pezerairoi, fighting with javelins on detached duties and wi1h the sarissa in battle. lt has been argued that they had a mobi.le role on the banlefield as well, serving as a "hioge" betweeo 1he cavalry and 1he phalanx; but though their position in the line would sui1this,1here is no trace in the sources of1he sorr offas1 banlefield movement this role would imply. Philip's feigned re1 rear at Chaironeia, carried out by elire troops who are presumably the hypaspisrs, is however one example of a manoeuvre whlch perhaps could not bave beeo carried out by a rigid phalaox with the sarissa.

Since the literary evidence seems inconclusive, I teod 10 rely oo the figures showo on the AJe.xander sarcophagus.

They carry the old " Argive" shield, the rim ofwhich makes il impracticaJ for use with the pike. lt seems unlikely and unnecessary that any troops would be issued with two separate shlelds, so these must represent troops who did not have the sarissa at ali; the hypaspis1s are the only candidate. Their weapon could be either the hoplite¡style 1hrusting spear, which the use of the hoplite shield suggests, or the traditional Macedonian javelins. The rivai candidate for the murder weapon of Kleitos is a longche, and Plutarch records the hypaspist officer Neoptolemos claiming 10 have served Ale.xander with /011gche and shleld. Longche is ofteo used a1 this era for thrown spears or javelins. Ali this suggests that the hypaspists could ha ve used either javelins or thrusting spears - or quite possibly both. Certainly it is qui te practicable fora man with a hoplite shield to carry a spear in his right hand and one or two javelins in his left, and early hoplites did so. If we have to pick one weapon type onJy, then the association of thrusring spears with a hoplite style shield in Philip's tomb may make spears the more likely.

Otherwise the figures on the sarcophagus wear greaves, unusually fasteoed by straps at aokJe and knee, aod the common leather or linen cuirass, with the waistband picked out in darker paint. One figure in a muscled plate cuirass without pteruges, the only bearded mao on the reliefs, is an officer and perhaps a portrait. The helmet is typically M acedonian, of Thracian type but without the Thracian style long pointed cheekpieces. Most of the sarcophagus helmets ha ve fix.ing holes for the attachme nt of side decorations, made from separate metal pieces like the weapons on 1he sculpture. These could be feathers like figure 29, small crests like 34, or conceivably even horns like 34a; they probably denote rank or guard status, and are unusual without a centrai crest. The shlelds on the sarcophagus have yellow (for bronze) rims. In India, when the whole army received new equipment decora1ed in silver and gold 10 replace the old gear which was wearing out, the hypaspists seem ro bave acquired the name argyraspides, silver shields - indicating where some ofthe silver went.

34 a nd 35.

HELLENISTIC OFFICERS

These men typify the style of equipment worn by mos1 officers from AJe.xander's time to the fall of1hc Hellcnistic statcs, and copied under 1he Roma o Empire - pia te corselet, cres1ed helmet, wa.ist sash, short boo1s, and cloak. 34 is from reliefs in the Artemisioo at Magnesia oo the Meander, dared betweeo 206 and 130, probably based on Seleucid officers - terracotta figurines from Seleuceia on the Tigris are similar. The helmet however is a fine example painted in the late Macedonian tomb of Lyson aod Kallikles; peak, the section above it, and cheekpieces are pia in iron, skull either plain bronze or painted ye1low, decorated with one red and one black painted band. It has three trailing crests, the centrai one orange-red, the others white. Side decorations like this may ha ve been a mark of rank as they were sometimes worn by generals - see figure 29, while Pyrrhos ofEpeiros wore a helmet with tali

107

a

whitc crcst and two goats' horns - 34a is a possible recoosuuction based on a helmeted bust of Pyrrhos. Philip V of Macedoo also wore a horned hclmct once. 35 wears more ornate armour, based on a marble panoply found at Rhodes, probably 3rd cenrury; it is highly decorated io relief. Armour could by oow be iroo; in 305 Demetrios Poliorketes was given two unusually light Cypriot iroo panoplies, a 3rd century paintiog shows a Ptolemaic officer in iron crested helmet aod pia te cuirass, with dark red pceruges, while a Macedonian-sryle burial discovercd in 1979 in Epciros held ao iron muscled cuirass with gold fittings, two iron helmcts (ooe silvercd) aod a kopis with bird'shead bilt. Bronze would still be commoner, and helmets at least could be silvered or gilded which would conceal their materiai.

The cloak in the Ptolemaic painting mentioned is green; kings and perhaps other generals would wear purple, Philopoimco in 207 for instancc identifying the Spartan ruler Machanidas by his purple cloak and ornate horsetrappings. Pyrrhos at Heraclea, aftcr one Roman officer had almost killed him, exchanged his gold-cmbroidered purple cloak and ornate armour for the dun cloak, plain cuirass and felt kausia ofone ofhis companions; the scheme backfired when this man was killed and Pyrrhos' troops thought thcir king had fallcn! Purplc kausiai and cloaks were bestowed by Maccdonian kings as honours, while Dcmctrios wcnt so far as to wear purplc boots, which was considered a little excessive. Apart from the purple, tbe symbol ofkingship was the diadem, a narrow ribbon tied round the head, sometimes outside the helmet.

36.

HELLENISTIC MUSICIAN ANO STANDARDS

Greek troops used trumpets for signalling, while some, notably the Spartans and Arkadians, marched to the sound offlutes. The trumpets occasionally meotiooed in Persian armies may beloog to Greek mercenaries. Macedonian and Helleoistic armies continued to use trumpets. This figure is from a tomb-painting at Marissa, a Ptolemaic military settlement in ldumaea, ofthc late 3rd cenrury.

Standards were oot used by Grcck armies ofthe classica) era. M oving in solid phalaoxes, they had little need for standards which are chiefly useful as rallying points for small units which might scattcr. In the 4th century they do seem to have used similar devices for signals; Diodoros mentions a Theban officer in 371 with a ribbon tied to his spear to signal orders. Some son offlags were used to signal attacks at Sellasia in 222. T he first rcfercocc to what may be a unit standard is in Alexander's army in 324; after Hcphaistion's death his cavalry continued to be known as Hephaistion's, and to carry his emblem. The word used is semeio11, a signor badge, which carne to be tbc usual word for standards. H cphaistion's men used under their new commander the emblem they had carried under their own; this was clearly not normai, so standards changed with, and thus wcre probably chosen by, unit officers. It is 108

.d

~~~= a

...........,_,, \ --f.

c d

e

significant that cavalry were the first, it appears, to use standards as they were obviously more likely to scatter and had more use for them. Nonetheless tbey were later used by all arms. Standard-bearers, semeiophoro1: appear for instance in Ptolemaic papyri, in Livy's accounts of Rome's war with Macedon, while Asklepiodotos ascribes a standard to each 256-man symagma oftbe pbalanx.

There are only scattered hints ofthe form oftbese standards. The monument to Alketas includes what may be a standard as shown in 36a, an eagle witb a snake in its claws above a cloth hanging; presumably Alketas' own emblem, it is a motifnot uncommon in Greek art. 36b is the banner from Alexander's own funeral car, crimson with a large gold olive wreath; red felt banners also adorned Hepbaistion's pyre. 36c and dare from coins ofthe Seleucid military settlement ofBlaundos in Lydia, ofthe 2nd century. The sun-disc and crescent moon imagery is like that of 100 and derives from thc samc near eastern religious background; it is unlikely to have been used by Greeks or Macedonians not settled in Asia. 36e is a later standard, carried by a man in Palmyran dress from tbe tempie ofZeus Theos in Dura Europos, originally a Seleucid foundation; though 2nd century AD it has the same symbols. The purpose of c-e may conceivably have been purely religious, not mili tary.

37.

SELEUCID PHALANGITE

This figure is based on rwo terracotta plaques found in Campania, showing troops in a mixrure of Macedonian and Asiatic dress and equipment wbicb suggests they are from a Greek army in the east, probably the Seleucid.

Alexander's Persian phalangites, the Successors' Asian pamodapo1: Mithridates VI of Pomos' later troops, would probably be similar. The trousers are a sure sign of"barbarian" influence, and a figure on tbe other plaque adds long sleeves. The cuirass bends witb the movement oftbe body, showing it to be linen or leather ratber than plate.

There seems no warrant for the view tbat infantry of the Hellenistic kingdoms cbanged to metal armour, as is sometimes suggested. Asklepiodotos menùons corselets, as does Plutarch describing Pbilopoimen's Achaians, but neither say they were metal, and tbeir wording would be equally appropriate to leather. Conversely, illustrations of infantry in non-metallic corselets are common, while most ofthose shown in plate cuirasses wear officers' waist sashes. lt seems generai practice, then, tbat only officers wore metal armour. Confìrmaùon for the Seleucids is provided by reliefs of warriors in similar corselets to this figure's, from near Epbesos, possibly from tbe tomb of Antiochos Il Tbeos, who died tbere in 246. They have short sleeves and bare legs; most have crested Attic helmets in contrast to the uncrested Thracian type shown here and in 37a (from coins ofthe Seleucid usurper Tryphon, 142-139) though one oftbe Ephesos figures does wear the helmet in 37b. Perhaps the Ephesos figures are based on the argyraspides, tbe Seleucid guard infantry, who might be expected to present a more Hellenic appearance, with bare legs and greaves, and to wear crests. They fought "armed in the Macedonian manner" as phalangites. Except for an uncertain reference to a forced march in Bactria, they are not often found on mobile operations such as Alexander's hypaspists had been used for; they were not used, for instance, when Antiochos lii stormed mountain passes. There is thus no reason to believc that they had an alternative lighter sryle ofweaponry in Alexandrian style.

109

110

Those convened 10 Roman armament by Amiochos IV would resemble figure 133, with mail, two greaves, sc11tum, gladius hispa11ie11sis and pila; therc would seem 10 be no need 10 change their crested Attic or Macedonian helmets for Roman feathered rypes.

This man holds his sarissa at an angle, as did the rear ranks of the phalanx, hoping to break the force of enemy missi Ics. The heavy weighted bunspike, tbe "coumerweighted portion" of Polybios, is taken from the Pergamon reliefs. Length ofthe sarissa increased after Alexander; Polybios says it was 14 cubits (21 feet/6.4m) long in the 2nd ceorury, but had "originally" been 16 cubits (24 feet/7.3m). lfTheophrastos is righi (see figure 32) this was not the origi11al length; but an anecdote of Polyainos menùons 16-cubit sarissa i in the carly 3rd ccntury, which would scem to confirm that this length was experimemed with, but found too unwicldy.

38.

ANTIGONID MACEDONIAN PHALANGITE

An inscription from Amphipolis dating to the reign of Phìlip V (221-179) lists fines for loss of military equipment, whereby we learn that the later Macedonian phalaox carried sarissa, sword, shield, greaves, helmet and ko1Ch11bos.

This last is a rare word, identified by Feyel io bis analysis of thc inscription, via a variant spelling, as a son of Egyptian jerkin or girdle. The inscription requires officers, but not rankers, 10 wear cuirasses. That tbc rankers lacked cuirasses makes Feyel' s identification of the ko1Ch11bos as a light substitute for armour qui te likely; ifit was of Egyptian ori gin it may ha ve been ofquilted linen. But the only illustrations which show Antigonid armour are little difTerent from the usual linen or leather cuirass. This figure wears a very light brown version from tbe painted tomb ofLyson and Kallikles, shorter in the body tban usual and with the long preruges staning well above the waist, but otherwise unremarkable; 38a is tbc front view. lc has been suggested that the office rs required by the regulations to wear corselets include file-leaders, and thus that the phalanx presented a metal-armoured front rank to the enemy; and this idea has evcn been extended to Alexander' s phalanx, whcrc there is no direct evidence for it at ali. The word used in the inscription for officer is hegemo11, tbe exact scope of which seems to ha ve varied; but it usually denotes officers offairly high rank, for instance senior enough 10 anend a council ofwar, while the Anùgonids used il as a specific rank as the equivalent of chiliarchos, commaodcr of a thousand. I would thus suggest it is highly unlikely that the term would be used 10 deoote men as lowly as file-leaders, at least without any qualifying phrase. Filcleaders will thus bave beeo armed no differently from the rest ofthe file.

This man's helmet is a low-crested Macedonian style known from two bronze examples from Melos and one iron from Pantikapaion (Kercb) in the Bosporan kingdom. Ali are qui te ornate, with small busts beneath the crest and in the curved terminals ofthe peak, while the only cbeekpicce surviving, from the Bosporan example, sbows a goddcss in relief. The style is connected wi th tbe Antigonids by a painted example in tbc Lysoo and Kallikles tomb, with no decoration sculpted on, and with long T bracian sryle cheekpieces. lt is multi-coloured, indicaùng the originai helmet was painted; cheekpieces and cresi are bronze, peak and the area above it black edged bronze, skull red. The strapped greaves are from the Pergamon reliefs. Tunic was red. Plutarch mentions red tunics at Pydna; compare the red-clad officer memiooed uoder figure 30. Red was very likely standard for Macedonian-style infantry, considering that thcre was a long military tradiùon for red tunics, that Philopoimen's Macedonian-influenced Achaians wore red, and that madder red is a common and stable dye.

The equipment ofthe Macedooian guard iofantry, the pehasts, is controversia! io much the same way as that of Alexander's hypaspists. Peltasts were often used on Philip V's forced marches, so it has beeo suggested they were lightly equipped and mobile. However oo occasions (for instance in Aitolia in 218) Philip left them behind with the "heavier pan" ofthe army while the merceoaries coped with the difficult terrain. At Pydna, Livy uses the peltasts' defeat of the Paeligni as an example of tbe dangers of meeting a sarissa-armed phalanx frontally, while Plutarch's account specifically describes this fight as a victory for the two-haodcd sarissa. He also speaks of a forcc of picked Macedonians with sarissa i who must be the agema, thc peltasts' clitc battalioo, because the word he uses derivcs from agema. The conclusion must be that the peltasts fought as an ordinary Hellenisùc phalanx in pitched banlcs, ahhough they may conceivably have used lighter cquipment for forced marches, oo tbe model of Alcxandcr's pezerairoi. lt may seem odd to fmd troops called peltasts fightiog io a phalanx, but since the Maccdonian shicld could be called a pelte it is not inappropriate. The agema at least had gold-decorated equipment, and in generai Hellenistic armies seem to have been expensively and lavishly equipped, usually from royal arsenals.

This figure holds his sarissa on his shoulder as on the march; Polybios points out the difficulty ofthis with such a long and unwieldy weapon. H is shield is slung by a baldric; Plutarch speaks ofthe Macedonians at Pydna swinging thcir pelrai down ofTtheir shoulders as they levelled their sarissai.

Ptolcmaic infantry would probably be similar; the Melos helmets are probably Ptolemaic, while ifthe kottltubos was Egyptian some ofthem may ha ve used it. Thc only contcmporary painting of a Ptolemaic ranker I know has a white

lii

corselet ofthe usual linen or leather 1ype with waistband and shoulderpieces in a darker hue. He is barcleggcd wi1h sbort boots ins1ead of greaves. There is no direc1 evidence on tbe equipment oftbe Ptolemaic guard infantry, the

agema.

39.

BACTRIAN GREEK PHALANGITE

The source for 1his figure da1es after 1he fall of1he las1 Greek sta1es in India; he is from a coin of1he Kushan king Kujula Kadphises, who in thc lst century AD ruled much ofthe territory wbich had been Greek. The coin is rare, referred to as the "Macedoni:m soldier" type. The man on 1be coin clearly has a long spear grasped in bo1h hands (though because ofthe limited space ofthe coin, notali ofit is shown) and 1he small Macedonian round shield. Tbe helmet is related 10 the Bac1rian types dcscribed under 53, and he has long sleeves, bare legs, boots, and a cuirass which seems to have preruges, probably the usual leather or linen type worn by most Hellenistic infantry. What he is doing on Kujula's coinage is not clear, since it is unlikely that there were stìll Greek uni1s operating, :md although Kushan coinage was influenced by Greek this warrior is nota known Greek coin moti f. h seems cenain, 1hough, that he must derive from some Bac1rian Greek source, and represent one oftheir soldiers - the full beard and long hair suggesting he himselfis an Iranian rather than a Greek.

40.

HELLENISTIC SHIELD PATTERNS

Macedonian shield designs, and 1hose of1he various Hellenistic kingdoms that followed 1hcm, have two common fearures; the 8-pointed star (sometimcs elaborated 10 16) wbicb was also common elsewhere, bui seems to have become a Macedonian na1ional or dynasùc emblem in 1he 4th century; and symmetrical geometrie designs based on severa! crescents arranged round tbc circumference. A fragment ofbronze shield-facing from Dodona has a pattern of crescents worked in relief. This and the painted examples which are 40f and i here suggest that the un ics called chalkaspides, bronze shields, carried bronze-faced sbields with geometrie designs in relief; there is no sign 1bat any pan ofthe design was picked out witb paint. Tbe argyraspides, silver shields, probably had similar shields panly or wholly pla1ed in silver (or 1in, wbicb would be cheaper). Tbe Macedonian leukaspides, white shields, are less clear; but the bronze facing of a shield of phalangite rype from Pergamon had no relief design, so these may ha ve had shields with white-painted grounds, with any blazon painted on.

40a, after a clay model from a Macedonian tomb ofthe early 4th century, is the earlies1 known Macedonian sbield blazon, the star alone. lt probably predates Philip's reforms, and has the flat rim of1he hoplite shield. No blazons auributable to Philip :md Alexander are known; but 40b is from coins of Patraios of Paionia, probably Alexander's contemporary. lt is carried by an infantryman, in trousers and helme1, who is ridden down by a Paionian cavalryman. Mcrker associ:nes it wi1h the killing ofa Persian officer by Ariston, leader of Alexander's Paionians; but 1his Persian was a cavalry officer. The victim's combination oftrousers and Macedonian blazon may mean he is a Persian in Macedonian employ and the coin relates to an otherwise unknown incident in the wars of tbe Successors. 40c is from coins oflalysos, tyrant perhaps oflasos in Caria c.300. The centre device is not cenain. The shield is shown in 1hree-quarter view, and is clearly shown as the concave rimless Macedonian infantry type. 40d is from Aitolian coi ns, a trophy along with Celtic shields ofthe vic1ory of279; it may be a Macedonian sbield looted by a Gala1ian, or ha ve been carried by the Celts' Thessalian allies. 40e has the monogram of Pyrrhos of Epeiros, from a coin he minted as king of Macedon.

40f-m relate to the Amigonid kings of Macedon. 40fis from a painting celebra1ing Gonacas' marriage in 279; it is polishcd bronzc with thc design in relicf. 40g is from a coin ofGona1as; 1hc ccntrc shows 1he king as Pan. Similar ponraits ofkings or gods occur on other coins, including n and p here, but it scems unlikely they were used on real shields. In thesc cascs I havc lcft tbc centre blank; 40i confìrms this is possible. The Oodona fragment has crescents like this onc, without thc small s1ars; it was probably captured by Pyrrhos from Gonatas and dedicated at Dodona.

A coin of Philip V has a similar shield with 8-pointed stars instead of 6. 40h is from another coin of Philip, borne by Athene instead ofher usual gorgon. 40i and j are from the painted Lyson and Kallikles tomb; i is painted as polished bronze, with thunderbolts between the crescents, wbile j is, from 1he outside in; whi1e rim with red zig:iag lines; dark red with pale greco-blue wreath; four circles patte mcd like coiled rope, ali red, blending from dark to orange; white circlc; pale grccn-bluc centre with white star. 40k and I are infantry shields from the Aemilius Paullus monumcnt, commcmorating Pydna. 40m is from a coin ofthe dis1rict of Amphaxitis, probably from Perseus' reign; the whirlygig motif may be derived from the star. 40n is from a coin of Macedonia Prima, one ofthe Roman puppet republics formed after Pydna.

400 is Seleucid, from a coin of Antiochos I. That the anchor was carricd this way up is confì rmed by a coin celebrating a late Seleucid-Ptolemaic alliance; it has Athene's shield blazoned with the anchor and the Ptolemaic eagle. 40p bo re a portrait of Antiochos IV. 40q is from thc sa me source as 34. 40r and s are compiex designs from the Pergamene reliefs; like other equ ipment carved there, they may be Pergamene, but are more likely trophics, 112

probably Seleucid. 40t is Pergamene, from early coins of the town of Philadelphia, founded by Attalos Philadelphos, 159-138 BC. No Ptolemaic blazons are known. Coin symbols which might have served as such include an eagle, someĂšmes with a thunderbolt bctwccn his claws like later Roman lcgionary eagles, or a lion's head.

113

42

41

41.

HELLENISTIC MERCENARY THUREOPHOROS

In the 4th century the peltast was the typical mercenary, to the extent that "peltasts" became a synonym for mercenaries, even when it was not technically correct. In the 3rd and 2nd centuries there are numerous referenccs to "mercenaries" who are not further described or defined, as ifthe term mercenary itselfindicated a specific troop type; and these are usually acting as equivalents ofthe earlier peltasts. For instance, in Philip V's army they hcad his column ofmarch in rough country, screening the heavier troops; they are differentiated from the light troops in one passage, from the heavies in another. Soon after we learn t hey are lightcr armed than Spartan citizen phalangites.

Merccnarics oftcn support light troops, for instancc at Mantineia. In 195 a significant pan ofNabis ofSparta's infantry fights with javclins, and as his citizcn infantry wcrc pikcmen thesc must be his numcrous mercenarics. It may not be truc that al/ me rcenaries were thesc peltast¡equivalcnts; but mcrcenary phalangites only appear once, in the Ptolcmaic army at Raphia, and in thls case wc are clearly told that the whole army had been re-organised and reequippped "paying no attention to the way thcy had previously been armed", so this says nothing about the usual merccnary cquipmcnt.

Whilc the mercenaries wcre thus cquivalcnt to thc classica! peltast, most did not carry the pelte, and they borea new name. A fcw illustrations stili show small round pelta1~ but the ovai thureos replaced it almost completcly by the mid-3rd century. Tbc troops are usually called thureophoro~ when they are called anything but mercenaries.

Asklcpiodotos stili uses peltast to mcan troops intermediate between phalanx and skirmishers and when used in historics it may occasionally mean the same thing. But because the Macedonian infantry shield was sometimes callcd pelte, pcltast or peltophoros could be, and was, used for pikemen. The meaning of pcltast in the 3rd century and latcr may thus vary from piace to piace.

114

Thureophoroi are profusely illustrated. Many grave paintings from Alexandria and Sidon show mercenaries from the garrisons; there are many Anatolians as well as Greeks, and mercenary units probably varied considerably in nationality; Anriocbos Ill's "mixed auxiliaries of all natiooalities" ac Magnesia may be typical. The gear is also shown by several terraconas from Seleuceia on the T igris, mostly of Parthian date, perhaps showing the city militia whicb served Parthia. 4 la is a pleated tunic from ooe sucb statuette; most ha ve che simpler tunic ofche main figure, with bands of embroidery at sleeve and oeck. Clothing colours on the paintings include whice, red-brown, rose, blue and ochre; one Bithynian from Alexandria seems co bave a dark cunic with a broad wbite stripe down the front.

Some unics may bave been uniformed, especially garrisons, since these muse bave been standing forces, whereas troops hired for campaigns may noc bave been. Cloaks are not always shown and may not have been wom in battle.

Weapons are long spear, javelins and sword. Javelios are not always shown. Tbc ovai rhureos was adopted from the Galatians, and like theirs was wood covered with leather, with a spined boss and centrai handgrip. 41 b is a nearrectangular variant, from the stele of a Bithynian officer, probably early 3rd cenrury. He has killed rwo enemy, a Thracian and a Mysian, probably Lysimachos' men; one has this shield, tbc other a convcntional ovai. Shiclds are usually painccd plain whirc or a pale colour, witbout a design. Only other defence is an iron or bronze M acedonian helmec, somerimes red-crested.

T bose Greek citizen troops who adopted the rhureos would be armed much the same. Plutarch says Achaians with rhureoi were armed with spears, fougbt at a distance skirmishing like peltasts, which must imply javelins, and were unarmoured. Boiotians continued co wear their own distinctive helmec, shown with rwo rhureoi on a relief.

42.

HELLENISTIC THORAKITES

Thorakita1~ "cuirassiers", are mentioned by Polybios rwice in the Achaian army and once in the Seleucid. Thcy are used co support Lighc infantry, sometimes on sceep hills, andare either associaced with 1h11reophoroi or have rl111reoi themselves. Clearly they are an armoured but mobile infantry cypc noc fighting in rigid phalanx formation. T his figure from Sidon, whose fragmentary inscription suggests he is an Anatolian, shows an armoured version ofthe rhureophoros who fits che bili nicely. He wears a light version ofthc Ccltic mailshirt, shorter than usual and lacking shoulder pieces. He would probably add javelins to his spear, like the last figure.

H. Russell Robinson suggested the bulbous hclmet, shown in the originai without cheekpicces and with thc rim cut away in an invened V shape to show the deceased's face betcer, was a Celtic type. However the outline is equally appropriate co some Macedonian rypes such as the one shown bere, from the Pergamon rcliefs, and the inverted V shape suggests che Macedonian helmet's customary decoration.

43.

ST AFFSLINGER

Ofuncertain but probably Hellenistic origin, the staffsling was a sling on a polc, swung with both arms to project heavier missiles ehan che normai sling co a greater discance. Vegetius describes its use by Romans against elephants, apparently under the Republic. lt may perhaps bave inspired the later one-armed stone-throwing catapults.

43a shows the kesrros or kesrrosphmdone, first used by the Macedonians under Perseus, and more effective than most missiles against Romans sheltering behind their heavy shields. The surviving descriptions, by Livy and the lexicographer Suidas, are based on Polybios. lt was a spcarhcad two palms (about 6" / l 5cm) long attached 10 a wooden shaft half a cubit (9" /23cm) long and a finger thick, with three short wooden nights artached like the fletching of an arrow. It was shot from a sling with two thongs of unequal lcngth; one would be released at the moment of discharge, freeing the missile. The descriptions speak as ifthe kmros was shot from an ordinary handsling, but it is difficult to visualise such a sling having two thongs, or shooting such a large missile, and a staff-sling seems more likely.

44.

TARANTINE CAVALRYMAN

4th cenrury coins ofTaras show her cavalry using shields whilst mounted at a time when no other Greek cavalry did so, and it is likely Tarantine influence was responsible forche spread ofshields to other Hellenistic cavalry. Cavalry called Tarantines first appear as mercenari es in 3 I 7 when 2,300 who had "come up from the sea" fought for Antigonos in Asia. lfthey carne from the coast they could have come from ltaly, but Tarantine may already have denoted a style of fighting rather than a nationality, since only a generation later Pyrrhos found Taras soft and unwarlike, not a likely sotirce for many mercenaries. Later, so-called Tarantines fought for the Seleucids, Spana, Achaia, Athens, Boiotia, Elis and very likely other Hellenistic states too. Sparta's and Achaia's at Mantineia seem stili to have been mercenaries, though probably with no connection with Taras, but Athens' were citizens, and others may also have been. "Tarantine" now only referred to weaponry and tactics.

115

The Tarantine style thus popu1arised was that oflight cava1ry skirmishing with javelins and, one must assume from the coinage, shie1ds, which were the chief way the Tarantines differed from earlier 1ight horse (1ike figure 26) a1though Livy says they a1so had rwo horses each, rare in Greece. Tarantine coins of about 380·345 show small round cavalry shields as in 44a, only l 8n (45cm)or so in diameter, probably the same as the small bossed hide shie1ds ofearly Roman cavalry (figure 139). T he style may ha ve evolved from an Italiote Greek sporting evcnt, thc anabates, a race for horsemen carrying small shields, known from the Sth cenrury. Later coins show the much larger shie1d of the main figure. lt may have been ribbed like that offigure 46, the rib indicating ltalian or Ce1 tic origin, but a 3rd century coin shows a Tarantine rider with a ribless shield blazoned with the popular 8-poinc star. Crested Attic helmets are shown on some coins; others show riders barcheaded, but as they are "heroically" nude this may not be significant.

Taras also had a few heavy cava1ry, known from coins and sculprure, in the usual muscled cuirass, with probably the same arms as the lights.

45.

HELLENISTIC GREEK SHIELDED HEA VY CAVALRYMAN

This figure is basically similar to 25 witb a hoplite·type shield added, based on rwo relicfs from Thespiai in Boiotia.

Greek cavalry adopted shields probably in the 3rd cenrury, either after the appearance ofTarantine mercenaries in 317, or after Pyrrhos' ltalian campaign of281·75, when he met shie1ded Tarantine and Roman cava1ry and may bave brought shie1ds back for his cava1ry, and the contemporary incursion of sbie1ded Ga1atian cava1ry into Greece.

The appearance ofTbracian cava1ry without shields in the ear1y 3rd cenrury Kazanluk paincings (figure 68) inclines me to favour the later date rather than posrulating a rapid introduction after 317. Most Greek cavalry seem to ha ve retained their traditiona1 javelins rather than being re-armed, as the infantry were, after the Macedonian sty1e. For possib1e exceptions see under figure 4 7. The much praised Aitolian cavalry are sa id by Livy to have used spears, nor unfortunately further defined, and some at least wore cui rasses. In view oftheir generai poverty and rapid mobile tactics thcse may perhaps have been non-metallic, but some may well have bought or looted metal armour.

46.

ANTIGONID MACEDONIAN HEA VY CA VALRYMAN

This figure is from the Aemilius Paullus monument, which clearly shows that Macedonian cavalry, like the last figure, had adopted large round shields. These have a central rib Like the rh11reos indicating Celtic or ltalian origin, but no metal boss or rim is indicated, and unlikc the rhureos this shield is carried with the rib horizontal. Unlike Macedonian infantry, no pattern is visible on the shield face. Although Peter Connolly says this shield "is certainly neither Roman nor G reek" it was in fact used by both, whatever its origin. It occurs on Roman coinage (see figure 140) and as for the Greeks it is carried by a cavalry officer in typical Hellenistic short plate cuirass and knotted sash on a relieffrom near Daskylion, on the borders ofBithynia and Phrygia. The cavalry on the Aemilius Paullus monumcnt carrying these shields cannot be Cehs, as Connolly suggests, as we know enough about the orders ofbattle to be fairly certain no Celtic cavalry were present, on either side. In add ition, the clearest surviving cavalryman has a Macedonian sryle helmet, with low comb and cheekpieces, very like the si1vered iron examp1e from Epeiros, discusscd under figure 34, depicted here. One rider shown without his shield has thc usua1 cloak and short musei ed cuirass.

Macedonian cavalry secm to ha ve discarded thc long xys1011 aod reverted to javclins, since Livy describes thcir consternation when charged by Roman cavalry; the Macedoniaos automatically assumed cavalry would skirmish. He also me ntions Philip V himselffighting with missiles from horseback, outside Athens. The chaoge may bave come about because using long xys1011 and large shield together was asking too much of cavalry without saddle or stirrups.

The rcduced shock power wou1d ha ve been acceptable because of the declining role of cavalry. Pyrrhos, who as notcd may have been responsible for introducing cavalry shields to Greece, scems according to Dionysios of Halikarnassos to bave had his guards armed with xys1011 in Italy, but after his return was using a bronze sbield on horseback in his last battle at Argos (Pausanias records that the Argives kept it as a lrophy), so may have adopted javelins and shie1d for himself and his guards.

47.

HELLENISTIC XYSTOPHOROS CAVALRYMAN

Some Hellenistic cavalry retaincd the xys1011, but those who did so do not seem to have adopted shields, though ccrtainty is difficuJt. This figure is based on a Roman coin ce1ebrating the defeat of Macedon in 168. He may conceivably be a Macedonian, but as he is shown charging triumphantly rather than defeated and dispirited he is more likely a Romao ally, depicted to celebrate their assistance. Cava1ry from severa( Greek states helped Rome during this war, but the most important cqntingent carne from Rome's faithful ally Pergamon, and this figure is most like1y a Pergamene. Equipment is basically that of figure 30, but the Boiotian hclmct bccame 1ess common in the 3rd century, and Macedonian types deve1oped from the Thracian became standard. This man's crest and feathers may

116

indicate that he comes from a bodyguard unit. Clothing colours would stili usually be those noted under figure 30.

Seleucid and Ptolemaic art occasionally shows long runic sleeves.

Apart from Pergamenes, cavalry similarly armed seem to bave included Ptolemaic, early Seleucid, later Achaian and possibly Athenian uoops. Severa} concemporary illustrations show Ptolemaic cavalry with long xysca and no shields, the la test dating about 200 BC, some time after shields had been introduced elsewhere. Seleucid cavalry are called xyscophoroi by Polybios and Plutarch, and a relier or a shieldless xyscophoros, oruncertain date, comes from Ephesos. In generai, illustrations or Hellenistic cavalry with the long spear are quite numerous, and so are illustrations or cavalry with shields; it is not likely to be coincidence that the two are never seen together.

In Grccce proper, Plutarch describes Philopoirnen using a xyston, so he may have re-armed the Achaian cavalry, as hc did their infontry, though he rnay perhaps have been expressing persona! preference alone. Polybios, once generai orthe Achaian League's cavalry, regards large round shields as standard Greek cavalry gear, but this need not mean that thc Achaian xystophoroi used them. Finally, there is a relier orHellenistic date from Athens showing a cavalryman in Boiotian helmet thrusting underarm with a spear, perhaps a xyscon (though the old shorter cavalry spear or figure 27 is possible) so the Athenians too may have bccome xyscophoroi.

I 17

48.

SELEUCID CATAPHRACT CAVALRYMAN, 2nd CENTURY

Seleucid cavalry remained like the last figure through the 3rd century. lt is possible that some oftheir lranian units, notably the Medes ofthe agema, may have been ofa more eastern appearance, and may even have continued to use partial horse-armour in Persian style, but we have no direct evidence ofit. At the banle of Panion in 200, however, we hear for the first time of Seleucid cataphracts, very heavily armoured cavalry apparently copied from the Parthians met on Antiochos IIl's recent eastem campaigns. By Magnesia, if not before, all the regular cavalry except the two guards units had become cacaphracts. Polybios describes them as "men and horses completely armoured' ', Asklepiodotos mentions "very heavy equipment protecting both horses and men on ali sides with armour". This figure is based on a bronze statuette from Syria. Thc combination ofHellenistic muscled cuirass and pteruges with the laminated tubular arm and leg defenccs of later cataphracts suggests strongly he is a Seleucid cataphract. Thc armpieces are also shown on the 2nd century Pergamene reliefs, the turned edges at the wrist indicating thcy are metal. lt is not clear how tbey manage to bend at the elbow. They do not cover the hand, though Xenophon earlier described a leather armpiece which did, so pcrhaps leather gauntlets may have been worn with this equipment. T he helmet, with a metal mask in the form ofa bearded face, is also from Pergamon. The statuette seems to be wearing a similar helmet, but it is hard to be certa in and he could have an open-faced Thracian helmet with "bearded" cheekpieces. 48a is a similar bronze Hellenistic masked helmet from Jugoslavia, with provision for a crest and with the skull modclled to represent hair. The staruene has an officcr's sash, and an officcr would certainly wear a crested helmet. The resemblance ofthcsc catapbracts to later types, plus the literary descriptions, suggests they would have ridden fully armoured horses.

49.

SELEUCID COMPANION CAVALRYMAN, 2nd CENTURY

Livy describes the Companions at Magnesia "with lighter protection for riders and their mounts, but in their other equipment not unlike thc cataphracts". This implies they had adopted at least some protection for their horses, at the time of the cataphract reforms. Any reconstruction based on this brief description can only be tentative; this figures sbows a bypothetical but plausible lighter version of the cataphract. Hc has tbc masked helmet and laminated armpieces ofthe Pergamon reliefs, but no leg-armour. The rest ofhis dress, including the elaborately decorated pteniges and tasselled boots, is from a relief ofSeleucus I, a late copy of a Hellcnistic originai, from Dura.

His horse has a narrow bronze chamfron wi th a trans verse crest offeatbers, and a small poitrail of embossed bronze or moulded learher, also from Pergamon. The reliefs may indicate some Pergamene cavalry were similarly armoured, but the equipmcnt is probably trophies taken from tbe Seleucids just as the Galatian arms shown are trophies.

Equipment ofthe other guards unit, the agema, is never described. As noted above they may perbaps ha ve had horse armour even before the cataphracts were developcd, and would fairly certainly have had it after. As they and the Companions were of similar status and idencical numbers, and sometimes fought brigaded together, their equipmeot seems likely to have been broadly similar. The agema may thus have looked more like an lranian trousered version ofthis figure, with the horsc-armour of 167, for instance.

Seleucid cavalry were elaborately equipped, all those at the Daphnai parade in 165 (including Companions, agema, cataphracts and militia) wearing purple cloaks cmbroidered with gold, tbc devices including animals - probably bees and/or elephants, Scleucid emblems known from coins. The Companions' bridles were decorated in gold, the militia's in gold or silver; this may ind icate bridlcs ofmetal plaques as shown for this figure.

50 and 51

SELEUCID LIGHT CAVALRY

These two lightly cquipped cavalry probably represent the city militia type mentioncd by Polybios when describing the Daphnai parade of 165 BC, thougb eitber or both may altematively be variants on the merccnary Tarantine type. 50 is from the same source as 37, and wears a similar Thracian helmct and linen or leather cuirass - note the long slit armholes. One figure in the source adds a cloak; another in the background has a soft cap, and may be like 51. SO's large shield suggests his main armamene was javelins, though weapons no longer survive on the plaque. 51 is based on terracotta figurines from the early Seleucid capitai Seleuceia on the Tigris; he has a smallcr round shicld and would bave similar weapons to 50. Dernils of his dress are unclea r becausc the figuri o es are not of very high quality; thus his tunic migh1 have sbon slceves, and it is oot emirely clear that thls type wore 1rousers - some figurines which certainly do bave poioted cap and baggy uousers may be more closely related to Parthian types also found. S_ince the populaĂšons ofthe Scleucid citics ofSyria and Mesopotamia were very mixed, however, militia units would probably have varied considerably in dress.

118

51

119

52 and 53.

BACTRIAN GREEK CAVALRY

52 is taken from a Graeco·Bacuian silver dish. From his costume, he is one ofche Iranian nobles who must have formed the bulk of che Bactrian kings' cavalry. Lcngch of the spear is uncertain as ic is impaling a lion on the originai, but the two-handed grip suggests scrongly thac it is the 12 fooc (3.6m) konros oflater cataphracts. le is essentially che same as the Maccdonian xysto11 bue chis cxamplc at leasc lacks a bunspike. He also has a cased bow, and wears a musclcd cuirass but no belmet.

53 is more Greek in appearance, from a coin ofHermaios, last Greek king in India. He has a bow the case ofwhich also holds a spear; this appears to be fairly sborc but the limited space oftbe coin may have affected it. The ends of che royal diadem are visible under tbe belmet brim ofche originai, indicating tbe rider represents che king hi mself.

The equipment is probably tbat of a royal bodyguard more Greek in nationaliry, dress and equipment than che rese ofthe cavalry - mucb like tbe Seleucid Companions. Menander's 500 Yonakas may represcnt such a farce. Tbe helmet is of a type apparently common in Bactria and India, sbown on many ofthe kings' coins. lt seems to be of Boiotian type, witb tbe addition ofdecoration on the forchead resembling Macedonian styles. A few belmets bave the division between peak and neckguard much more sharply emphasised, to give a hclmet very like the Macedonian style but lacking cheekpieces.

Bach these cavalry ride una rmoured horses, and 1here is no direct evidence for any Bactrian Greeks using borse· armour. Since their Saka, Panhian and Seleucid neighbours ali did, bowever, it is possible 1ha1 some Bactrians may have copied thcm.

54.

LYKIAN INFANTRYMAN

Lykia was part oftbe Persian Empire at che start ofthis period, but though ic enjoyed considerable autonomy, and its aristocracy was closely connected with the satraps and other local Persian nobles, most ofits cities surrende red to Alexander without a fighe. In tbe Successors' wars Lykians fought brigaded with Pamphylians for Antigonos and Demetrios and then for Lysimachos. This figure is in fact based on 5tb cenrury sources (tomb paimings from tbe Karaburun II tumulus near Elmali) but equipment is likely to bave been much thc same in tbe 4tb. ln the 3rd century Lykians, like other Anatolians, may have adopted tbe chureos; the Lykian mcrccnarics known from tbc Ptolemaic garrison on Cyprus may have been cl111reophoroi. The Lykians, Pisidians and Pamphylians in the Seleucid army at Magnesia are called peltasts, and may ha ve been chureophoroi or similar; but as they fought in the skirmish line they may more likely have been light javelinmen witb small round shields like 1he pelte.

This warrior bas a short light blue cunic, plca1ed below the waist, and a shorc white cloak knotted on the lef1 shouldcr. He bas whice leggings with black or red garters - He rodocos says che Pisidians bad similar crimson leggings, and a Lydian wears tbem on one Persian relief. Shoes and belts are red. Elsewhere on the paintings red is a clothing colour, along with green and purple. Ma in weapon is a shorc spear wich iron points at both ends. It is used co chrusc, but from tbc length may be che last of a bunch of javelins, the rest of which ha ve been thrown . A short dagger with a black hilc is in a red sheath. Shield is ofthe hoplice's Argive cype, wich a reinforcing pane! u nder the arm, and tbe inside of tbe shield is painted in dark red, blue and white.

55.

KAPPADOKIAN CAVALRYMA N, c.320 BC

Kappadokian cavalry foughc for Darius ac Gaugamela. The country chen passed co a Persian noble, Ariaratbes, who raised a mixed native and merccnary army of30,000 fooc and 15,000 borse. He was killed by che Macedonian regent Perdikkas, and Eumenes ofKardia took aver the province. For his campaign against Kraceros and Neopcolemos in 321 Eumenes raised numerous Kappadokian cavalry, and cbey probably provided his lacer bodyguard.

This figure's dress is based on earlicr Persian representations ofKappadokians. He has Median tunic, crousers, shoes and cùtra, the crown stitched in three folds and the cheekpieces knottcd behind the head. He also wears a short cloak with weighcs or cassels ac the corners, fastened by a Phrygian fibula. The scaled cuirass and kopis could already bave been wom by Darius' Kappadokians, but chese would bave carried javelins while Eumenes, since he was raising cavalry as a subs1ituce for Macedonians, would probably bave given at leasc his bodyguard che xysco11, as Alexander bad already done wich Persians. As noted under figure 5, 1he role of Darius' Kappadokians suggests some nobles may have caken co riding armoured borses and wearing parapleuridia; if so Eumenes' guards will no doubc bave done tbe same. Most cavalry would probably have been lighter.

120

52

121

56.

KAPPADOKIAN INFANTRYMAN, 2nd CENTURY

A native dynasty claiming descent from the first Ariarathes later regained Kappadokia, and king Ariarathes IV seot Antiochos IV 2,000 infamry "equipped in che Galatian sryle" who fought at Magnesia; the Kappadokians found later assisting the Galatians are probably the same force. This figure has the Kappadokian dress of the last, but carries a rypicaJ Celtic sword and shield. he might add spear or javelins. Herodotos' earlier testimony has Kappadokians armed with short spear, javelins, and small shicld; adopting thc Celti e longer sword and larger shield may not have changed their fighting methods greatly.

The nonhem pan ofK appadokia became the kingdom of Pontos, whose army co ntained Kappadokians, locai hill tribesmen, Gal3lians and Greek mercenaries. The prominence of archers in later Pontic armies suggests some Kappadokians may have been chus armed.

57.

ELYMAIAN ARCHER

The people known variously as Elymaians or Kissians inhabited the hills of ancient Elam becween Mesopotamia and Persia. Their archers fought for the Seleucids at both Raphia and Magnesia. As early as Herodotos' time they dressed and fought much like the Persians, though wearing a fillet round the head rather than the tiara. Persian sculpture confirms this, showing Kissians at court in the same long formai robes as the Persians, wearing the fiUet and also wearing shon laced ankle boots instcad oflow Persian shoes. In battle they, like the Persians, would have worn Median tunic and trousers. The resuh should have been like this reconstruction.

58 and 59.

ARAB ARCHERS

The Arabs appear on che fringes of Hellenistic history chieny as mercenaries, though one Arab state plays a role.

The Nabataean kingdom emerged in south Palestine in the late 4th century, beating off Antigonos One-eye, and in the early Ist cenrury BC defeated both the Maccabees :md the declining Seleucids in batti e, killing the Seleucid king Antiochos XII and capturing Damascus.

M any poorer Arabs would be very ill-clad, probably like 58 with no more than a loincloth beneath their shon, dyed , behed cloaks, but Pcrsian and south Arabian sculptures shows the richer men wore long ca)f. or knee-length runics like 59, sometimes with pattcrned hems. The carefully draped robes sometimes worn on top are clearly not battle wear. Sandals were generai. The main weapon seems to be the bow, which Herodotos earlier described as long, assuming a reverse curve whcn unstrung, and carried on tbc right. Strabo adds spear, sword and sling. T he Arab mercenaries defending Gaza against Alexander fought mostly with missiles, but one anacked the king with sword and shield. Figure 58's sword, on a baldric, is from Persian sculpture. Diodoros describes the Nabataeans beating off Antigonos' expeditions with both arrows and javelins. The infantry of"Arabs and neighbouring tribes" serving Antiochos nr at Raphia were so numerous - I 0,000, the largest of ali the Seleucids' Asia tic levy contingeots - as to suggest thcy were the whole avai lable manpower ofthe locai tribes, so probably mixed spearmen, archers and others.

Arab camel-riders fought, or rather failed to fight, for the Seleucids at Magnesia. They were equipped like 59, as archers but carrying long thin swords, 6 feet ( I .Sm) long, so they could reach the enemy from their high perches.

This is particularly interesting as it suggcsts thesc Arabs cxpected 10 fight from camel-back; later Arabs more often rode to battle but fought on foot. The carnei harness is based on a seal from Babylonia, probably of Persian date.

Therc is no rein or bridle, and in fact the carnei is guided che sa me way the Numidians conuolled their horses, with a stick plus a coil of rope round the throat. The rider si1s on a saddle-pad atop a chequered cloth.

Strabo says the Arabs had no horses. This may have been true when the earlier source be took it ITom was written, but by his own day in the lst century BC was not. A few representations of ridden horses apparently date as early as the 2nd ce nrury BC. Cavalry would stili probably be rare in ou r era, and mostly armed with spears or javelins.

60 and 61.

MACCABEAN INFANTRY

Therc is very little contemporary evidence on the appearance of Judas' and his successors' troops, and these reconstructions largely use later evidence for dress and the Rute of War for armament. Cosrume is that found in the Bar- Kokhba cave finds, dating to the 2nd century AD and known from other sources of roughly the same era.

Costume ofthe Maccabean era was probably similar, but this cannot be certain. Tunic is wool or linen made oftwo rectangular sheets sewn together at the shoulders, with a hole left for the head, apparently left unstitched down the

122

59

sides but presumably belted. T unics are often white with purple stripes down front and back, otherwise red, saffron yellow, indigo blue, or black. Also worn is a rectangular mantle, decorated with a notch-ended band of colour and probably a fringe. lt usually secms 10 have been drapcd 10 !cave the right shoulder uncovered and held in posi1ion with the left hand. lfworn in baule i1 would obviously have to be fastcned by a pin or brooch, and might perhaps have been worn like 61, which does leave 1he right shoulder free, or else like a conventional cloak. Leather sandals are worn.

60 is a slinger; these are mentioncd in the Ru/e and perhaps hinted a1 in I Maccabees. He is very likely typical ofthe early guerilla forces, while similar light Lroops would also serve in later armies. In the ea rly stages ofthe revolt Judas' men were very short ofarms, "who nevertheless had neither armou r nor swords co their minds". Captured Seleucid cquipment supplied most ofthe deficiency, Judas himselfusing till bis death the sword he had taken from the defeated Seleucid generai Apollonios.

123

61 is the infantry type who is the mainscay orche forces described in tbc Rute. He has no armour but carries a shield 2 112 cubits long by 1112 broad (45" tl .14m by 27" /0.68m), a spear 7 cubits (IO' 6" /3.2m) long, javelins and a sword. lt will readily be seen that he is armed just like the common mercenary 1l111reophoros (figure 41) of Hellenistic armies, whicb is hardly surprising ifthe Rute does indeed reflect Maccabean practice, as they would start by using caprured Hellenistic equipment, later no doubt copying it. Attempts ha ve been made to suggest Roman parallels for this passage, identifying the shield with the Roman sc111um, but the lack or armour is in marked contrast to Roman practice, as is the long spear, and che Hellenistic parallel is far closer. The Rutedescribes tbc sbield as bronze-faced and che sword ornamented in gold, silver and gems, but these are probably literary conceits reflecting no real practice.

62 and 63.

MACCABEAN CAVALRY

Again chese are taken from the Rute. 62 represents thc younger cavalrymen assigned to support the infantry, unarmoured and armed with bow and javelins.

63 is one oftbe older men deployed on the wings, and while pans orhis equipment have clear Hellenistic parallels be does present some curious fearures, and is perhaps the most questionable ofthese reconstructions. The Rute says "tbey and their mounts shall wear breastplates, helmets and greaves; they shall carry in tbeir bands bucklers and a spear 8 cubits ( 12 feet/3.65m) long." Unfortunately the manuscript is damaged at the references to horse armour and co tbe length orthe spear, so che reading here is uncertain. Irthe lengtb ofthe spear is correct, however, it would clearly seem to be the Hellenistic xys1011. Possibly dubious reatures are che use of a shield with such a long spear, in apparent contrast with Hellenistic pr:ictice, and the wearing of greaves, unusual for cavalry. I bave shown che belmet as a typical Macedonian type such as might bave been taken from tbe Seleucids; the cuirass is not, and a Hellenistic pi ate cuirass could be substituted. The style or armour I bave sbown is a cuirass ofHellenistic cut, with shoulder-pieces and pteruges, but made oflaced metal tamellae, from a Palmyran sculprure orthe lst century AD. Lamellar armour was widespread under the Assyrians. What became ofit after their fall is uncenain, but it cannot have died out, as this example shows. Hellenistic forces knew lamellar borse armour at least (figure 167), and it is possible that lamellar cuirasses may ha ve continued in use throughout the Persian and Hellenistic periods, perhaps even in common use in Syria and Palestine, and that the Palmyran sculprure may represent a style orrespectable antiquity, which would seem a likely candidate for Seleucid and Maccabean use. Tbe books or the Maccabees unfortunately make lictle reference to armour, and though judas is described as wearing a corselet its type is not mentioned.

64.

PAIONIAN CAVALRYMAN

The Paionian trĂŹbes had by the mid-4th century formed a small kingdom nonb ofMacedon. Throughout our period it was dominated by its southern neighbour, and both under Lysimachos and under the later Antigonids was directly incorporated into Macedon. Paionian cavalry fought for Alexander, and infantry for Perseus. This man is based on the coins orking Patraios which also provide 40b. The creases ofthe tunic bave been identified as purugts, and hence the figure said to be armoured; but enlarged pbocographs clearly show be is unarmoured. The runic, square-necked wi th elbow-lcngth sleeves, falls into heavy folds (or possibly is patterned in stripes?) both above and below che waist. He has a crested Attic helmet and tbrusts underarm with a spear about 7 feet/2. l 5m long; he would probably bave a sword on a baldric tbough none is shown on the coin. His large saddlecloth cxtcnds round thc horse's chest and has a pebbled finish probably indicating furor sbeepskin.

The Agrianes were a Paionian tribe not included in the Paionian state, but with a king orchcir own. Thcy provided light infantry to Maccdon, but no cavalry are mentioned. They may have dressed similarly co this figure.

65 and 66.

THRACIAN INFANTRY

Thracians were typically red-haired and grey- or blue-eyed. Their traditional dress was thc patterned tunic and foxskin cap offigure 67, thc fawnskin boots of26, and a long cloak, che zeira, geomet rically pauerned in bright colours. Their main traditional infantry rypc was the peltast armed with thrusting spear or javclins, sword or dagger, and the pelte (see figure 20). lnfant ry with this dress and armament are common in 6th-5th ccntury Greek art, and stili described by Xenophon in the early 4th. Tbey would probably stili be in use at the start of our period; the costume is stili worn by the Thracian warrior goddess Bendis on an Athenian relief or about 350.

Later infantry are scii! essentially javelinmen - spearmen are last mentioned in 322, when a Macedonian generai sent some to arrest tbc Atbenian orator Dcmosthenes - but style ofbotb cosrume and equipment has changed. 65 is from the late 4tb or early 3rd cenrury Odrysian tomb-paintings at Kazanluk. He is clean-shaven while most earlier Thracians are shown bearded. He wcars low yellowisb, or red- brown sboes with uprurned toes, and his tunic could

124

be shon-sleeved, or slecvelcss in which case it was probably fastened at the shoulder with pins, as it is occasionally shown leaving the right shoulder bare like the exomis offigure 17. Clothing colours are white, off-white, cream, rose, light blue, light green and a variety of reds and red-browns. Most infantrymen wear bronze Attic helmcts, one with a pale blue crest. One figure however is bare-headed and long-haired; anotber, who wears a cloak, has the curious headgcar shown in 65a, painted in tbe ochre used for bronze. Tbc hclmct of66, tbc typc gcncrally callcd Thracian, was perhaps commonest. In contrast to earlier fashions, cloaks and tunics are usually plain, though the tunic ofthis figure, red-brown with one white stripe on each sleeve and two down each side, is worn by a servant elsewhere in the tomb.

Weapons are two javelins and a curved iron biade, shown here as a sword of a type found at Vinograd, though it could perhaps be a shoner version of66's weapon. Swords would probably be worn from a baldric. The shicld used is the ovai thureos, either the shape shown here or the squared-offCeltic type. Tbc face sbows tbe same metal strip boss that the Celts used; the centrai spine is not clearly shown, though may be represented by slightly darker paint.

However the Bithynian stele discussed under figure 41 confums that Tbracians did use the thureos witb the usual 125

spine. Since the paintings probably predace che Galacian invasions ofThrace, Macedon and Greece, it seems likely that che Thracians adop ted the tl111reos before the Greeks did, either from earlier concacts with the Cehs, or from the Illyrians, who used the same type ofshield (figure 73). The Kazanluk shields are painted in light ochre; Thracians serving Perscus of Macedon carried whice tliureoi. Round shields were also used; 65b shows one with a centrai silver plaque, dating c.350-300; similar round plaques have also been found. 65c is carried by che goddess Bendis on a Bithynian coin ofNikomedes I (279-255), along with two javelins and a straight sword with scabbard and baldric.

The shield is either decorated wich circles ofrivets, or pcrhaps has an embossed metal facing.

66 represents one of Perseus' Thracians at Pydna, with black tunic, whlce tlzureos, greaves, and che enigmatic rliomplraia as his ma in weapon. This is the only recorded instance of a Thracian unit apparently wearing uniform, and contrasts wich the varied colours of che Kazanluk paintings. The rlromphaia (or mmpia, in Latin) is first mcntioned in Thracian hands by Livy, who describes Philip V's Thracians in 200 finding it long enough to be inconvenicnt in woods; be also describes Perseus' men at the Kallinikos skirrnish in 171 using it to hamstring horses and to carry sevcred Roman heads as trophies. Plutarch, using the eyewitness account ofScipio Nasica, describes Thracians ac Pydna "brandishing scraight rhomphaias, beavy with iron, on their righe shoulders". In che !se cenrury AD, Valcrius Flaccus' Argo11awica mencions Bastarnae carrying a bark shield and a rumpia with an iron biade the samc length as the wooden bafc. This scrongly suggests che identification ofthe rhomphaia wich the billlike weapon shown here, used on the 2nd century AD Adamklissi reliefs by warriors ofcen idencificd as Bastarnae. lt has the righe proporcions ofiron to wood, is suitably long and heavy, and ic is swung wich both hands. Trus would entail slinging the shield on che back; earlier Thracian pelta i were certainly thus slung. This weapon seems to fic che various ancient descriptions bener than any ocher, with che possible excepcion of Plucarch's "scrnight"; bue che curvature ofthc Adamklissi blades varies greatly, and examples like that carried by this figure are straight for mosc ofcheir length. A biade wrucb might be from such a weapon was found ac Olynthos on che Thracian coast, probably dacing from the siege of348; at l 5.3" /39cm it is shorter than that ofthe main figure, but comparable to some ofthe more sharply curved Adamklissi examples. 66a shows this biade mounted on an Adamklissi shaft. If, however, this Olynthos biade actually belonged to some agriculrural implemenc, then tbe rlromplraia may ha ve been copied by the Thracians from the Bastamae, or invenced, in the 3rd century.

Of Perseus' three Thracian infantry concingencs, the rhomphaia-wielders were probably the "free" Thracian mcrcenaries; the Odrysian infaotry were more likely to be skirmishers anachcd to the elite Odrysian cavalry, and the Thracian settlers, from their association with Paiones and Agrianes, were probably light-armed javelin skirmishers. Philip V's Thracians carricd javclins as well as che rhomplraia. This man has a daggcr as an auxiliary weapon; both straight and curved daggers have been found.

a

67

68

126

67.

THRACIAN ARCHER

This figure is from a probably 4ch-cenrury silver belt-plaque. He has che craditional beard, fox-skin cap (alopekis) and pattern-edged runic. The quiver hanging from his waiscbelt is caken from a 6th-cen rury vase-paiming of a similar kneeling archer; che dagger is conjeccural. Small bronze socketed arrowheads with three barbs, found at Olynthos and elsewhere, may be Thracian.

Ocher lighc infancry - javelinmen and a few slingers - may bave dressed similarly, wearing che cap inscead ofthe heavier cypes' helmec. Thracian wicker shields (gerrha) were among che Roman trophies from Pydna. This could indicate chat che Thracian rh11reoi were made of wicker, rather chan che usual heavier wooden construccion, but is more likely co mean tbat light infantry stili carried the old wicker pelrai.

68 and 69.

THRACIAN LIGHT CAVALRY

These are typical ofche majority Thracian cavalry cype, unarmoured and armed with javelins and probably sword, usually the kopis. 68 is from the Kazanluk paintings; he has a plain cloak rather than the earl ier patterned type, and wears a typical T hracian helmet. The saddlecloch is red with yellow decoration, except chat the tassels on the rear points are white. Others are straight-edged and plain red.

69 is from a carved sarcophagus from Ladschane in Bulga ria, probably ofthe late Hellenistic era. His saddlecloth, to judge from the ragged edge, is probably an animal-skin; unusually it has a scrap for the horse's rump as well as che breast. T he rider wears long sleeves and che high laced boots ofearlier Thracian costume. This figure indicates that Thracian cavalry, like their neighbours, adopted the shield, probably some cime in che 3rd cenrury when che Greeks and Macedonians seem to bave taken it up.

70.

GETIC HORSE ARCHER

The Gecai, a north Thracian people ofthe Danube valley, roseto prominence in the late 4ch cenrury, fighting the Macedonians under Alexander and Lysimachos; their power seems to bave been broken by the Galatian invasion.

They and their neighbours were under heavy Scychian influence, and many fought as moun ced archers in Scythian style. This figure is based on Achaemenid Persian reliefs of their Thracian subjeccs, and bis Scythian coat and trousers suggesc he represents one ofchese northerners. The originai reliefhas a pointed cap ofScythian type, but I bave substiruted a bronze Thracian helmec, a 4th-cenrury example from a later Getic grave. Footwear is unclear on the relief, but the toes turn up. The bow would be the Scythian composite type in a leather goryros; Getic sites have produced numerous small bronze three-edged arrowheads, as well as iron spearheads, curved swords and knives, similar to southern Thracian types. He rides on a simple saddle as discussed under 72.

127

71.

THRACI.AN NOBLE HEAVY CAVALRYMAN

At the beginning of the 41h century Xenophon saw a Thracian prince with a retinue of cuirassed cavalry, and che richest nobles throughou1 this period would no doubt be armoured. T his figure is rypical ofthe early and mid-4th century. His bronze piace cuirass is a late improved version ofthe "beli" rype, a variant worn in G reece at the beginning ofthe 5th cenrury, but found in Thrace Ăšll che middle ofthe 4th. Unlike the early beli cuirass, this sryle has only a narrow out-turned flange at the waist, and instead ofa high collar to protec1 the throat, the neck is cut low leaving the upper chest exposed. Unlike Greek examples, which were worn with pterugts probably attached to an underga rment, the Thracian cuirasses have a row of holes along the edges to take a Liaing, and this probably indicates they were worn without pceruges. The exposed throat and upper chest were covered by a crescent shaped gorge1 of silver-plated gilded iron, decorated with bands of relief ornamentation; it has an upstanding collar to protect che th roat, and a narrow h inged strip fastens round the back ofthe neck. The collars seem to have stayed in use longer than 1he cuirasses, as two examples date from che second half of the 4th cenrury (from the Maltepe tumulus and Vurbitsa), and a gorget of similar shape and date, but of bronze scales on leather, comes from a Macedonian grave. T hese examplcs are not associated with cuirasses, and in fact it is notatali clear what armour was worn when the late beli style went out .offashion. One grave (lst century BC, from Beroe south ofthe Balkan mountains) produced scale atmour, but generally this was confined to che north (next figure). Possibly Linen or leather cuirasses wcre worn (there are earlier examples of metal fiu ings for Greek-style aoa-metal cuirasscs) or perhaps the Helleaistic cavalry muscled cuirass, though no traces ofth is type survive.

This figurc's helmet is a horned bronze example, with cresi and iron cheekpicces, from a 4th century grave at Bryastovets near the Black Sea. T his spleadid example probably belongcd to a man ofhigh rank, and the plainer style of 68 would no doubt be commoner. At least one iron helmct has been found, aad one leather example with bronze neck-and cheek-guards and ornamental appliques. Otherwise this man is similar to the light cavalry, and the noble heavies no doubt adopted shields at the same time. Saddlecloth is aga in a Kazanluk example, brown on cream.

72.

NORTH THRACIAN NOBLE CA V ALRYMAN

Based chiefly on a group of mid-4th cenrury silver-gih plaques showing mounted heroes, from Letnitsa in the Triballian country, this man represems 71 's counterparts among the Triballoi, Getai and the other northern tribcs.

Hc wcars loag sleeves, here rolled up, and trousers. Armour is a bronze or iron scale cuirass, with scaled pcemgts, short sleeves, and a small collar of upright plates. Scale armour in Thrace is in generai confined to the northern arcas, and is probably a product of Scythian influence. One 4th or 3rd century Getic grave held a cuirass of iroa strips linkcd with wire rings, perhaps related to the bronze strip limb-defences described under 77.

One ofthe Letnitsa plaques shows the hero hunting, wearing greaves with a foce at the knee; elaborate panly gih silver greaves with similar faces have been found at Triballian Vratsa and Getic Aghighiol. These are expensive display items, but the plaque suggests cheaper versions may ha ve been used for war and the hunt, so they are shown here. Also shown is the elaborate partly gilt silver helmet found with the Aghighiol greave; it has gilded eyes on the 128

forehead and Ùle cbeekpiece bears a horseman similar to the Letnitsa heroes except that his hair curls and his cuirass is shown as smootb (linen or lcather?) not scaled. Again, cheaper versions ofthis expensive helmet may perhaps have been used in action, as bronze Greek and Thracian sryles cenaioly were. But ali the Letnitsa warriors are in fact bare-headed, mostly with shon hair but occasionally with tbe topknot of 72a. 72b and 72c are two gold ceremonial helmets from Cotofencsti and Baiceni respectively, ofsimilar sryle to the Aghighiol example, though both bave their sl.."111ls covered with styliscd curls (or flamc symbols). An cxample from thc Iron Gates is cvcn closer to the Aghighiol sryle.

The Lernitsa plaques show che warrior armed with a spear, like most Thracian cavalry, and no doubt a sword would be added. They also suggest a simple saddle.One ofthe horses from che Kazanluk paintings has a low dark brown saddle on a cream and brown cloth, shown here. This suggests the southcrn Thracians were aware ofthe simplc Scyrhian saddle, having no doubt learned of it through rheir norÙJern cousins.

73 and 74.

ILLYRIAN WARRIORS

Generai appearance aod dress of these figures are from earlier artistic sourccs, no representations of lllyrians apparently surviviog from our era. Latest ofthese sources is a decorated uro from Ribic, perhaps early 4th century.

A bronze belt-clasp from Vace io Slovenia, 5th cenrury or earlier, shows similar warriors. They are clean-shaven, barefoot, and wear short, sleeveless, unbehed tunics, decorated with a fondness for venical stripes.

73 is from one of tbc nonhem tribes, like the Ribic and Vace warriors; they were distinguished from the tribes nearer to Macedoo and Epeiros by their preference for ovai shields Like che Italian scii/a (thcre seem 10 have been dose connections betweeo Ùle early cuhures of lllyria and north ltaly). 73a and 73b are alternative rypes.

Southemers favoured a small round coovex shield similar to the Macedonian type, sometimes decorated in the sa mc style. lt may thus have been ofsimilar bronze-faced construction. The Macedonian-style star-and-crescent patterns occur at times and places where thcy do not apparently refer to Maccdonian rulc, so may well be part of a joint heritage shared by Macedonians and lllyrians from carly timcs. 74a and 74b are from coins ofthe coastal town of Skodra, probably late 3rd century; 74c is from astone rclicffrom Baush. The main figure has a shicld pattemed with concentric circles, from an earlier bronze model.

Weapons could be spear, javelios, sword and light axe. Spears may have become more comrnon with the apparent increase in hand·to·hand fightiog. Some seem to ha ve been qui te long, to judge from the sheer size ofsome of the spearheads; Appian describes the oorth Illyrian Iapudes using long spears against Augustus, probably the sibina of earlier Roman poetry. The typical sword was a short curved weapon with che cuning edge on the inside, similar to the weapon of 65; it was also used in Epeiros. d

74

129

73 wears a Negau helmct, a conica) bronze rype with a narrow thick brim, sometimes crested, also used in nonh ltaly and Etruria. 74 has a late example ofthe Greek helmet sryle normally called Illyrian, obsolete in Greece in che early 5th century but lasting in Illyria perhaps as late as the 2nd. This one, from near Apollonia, lacks che characteristic double crest ridge of earlicr models. Other early helmet types characteristic oflllyrians may stili ha ve been in use at the start of our era, but soon disappeared, ifthey had not already. These included the broad-brimmed pot of74d and the light Smarteja rype of74e, ofwicker reinforced with metal studs and plates. Other armour was vcry rare and probably limited to chieftains, occasionai sculptures showing Greek greaves, while cuirasses of studs made in the sryle of74c may also have existed.

Light infantry includcd slingers, javelinmen and perhaps a few archcrs; unlikc the other rypcs archers are not mentioned in battle accounts, but bronze arrowheads ha ve bcen found. They may perhaps ha ve been solely hunting weapons. Light infantry would no doubt look like poor versions ofthese warriors, but lacking helmet and perhaps sword and shield. Many slave warriors would also look much like poor Illyrians, but others might wear remnants of their originai national costume - Illyrian slaves including many Greeks :md ltalians.

75.

ILLYRIAN CAVALRYMAN

Never numerous, Illyrian cavalry were unarmoured javelin skirmishers. The Vace clasp shows them without shields. The later Romans caUed shielded javclin cavalry "Illyrian", which implies they adopted shields at some time - perhaps in the 3rd century BC when their Greek and Macedonian neighbours also took them up. This fĂŹgure's helmet is a bronze Macedonian type, found in Lake Ochrida on the Illyrian-Macedonian fron tier, inscribed with the name ofa 3rd-century Illyrian king, Monunios. lt has a flat metal crest, with a hole in it, riveted on.

76.

SCYTHIAN HORSE ARCHER

The Scythians worc their hair very long. They are usually shown with long beards, and with high cheekbones and broad flat noses. They may have worn tattoos on tbc arms and hands, as men ofthe similar Altai culture did. T his figure is the basic Scythian light cavalryman in the costume best depicted by Greek precious metalwork made for che Scythian market in the 4th-3rd centurics. He wcars a crossover-frontcd coat cut short in che back bue coming to a long point in the front, baggy trousers usually tuckcd imo soft, low boots, and a pointed cap though Scythians are quite often shown barcheaded. Clothcs wcre ofsilk, leather, felt, wool or linen. Coats could be lincd with ermine, sable or other furs, and were trimmed at the edgcs with fur, patterned felt applique, or rectangular gold plaques dccorated in relief. Thcre might also be decoration a round the upper arm and down tbc centre ofthe back. Trousers too were highly decorated. As for colours, one coat dyed red and green has been found, while the Altai tombs at Pazyryk ha ve produced a shirt and a cap in white edged with red, and othcr tcxtiles in various reds, crimson, sand, brown, yellow, cream, blue, purple, black and whitc. These bright colours would usually be dulled by dirt as the Scythians did not wash (probably for religious reasons).

130

a

b

e

Main weapon is a short recurved composi te bow ofhorn, wood and sinew, with a sinew string. It shoots very light arrows, which in the Altai had their shafts painted black and red, with small socketed heads, rypically casi bronze though bone and iron could also be used. Bow and arrows were carried together in a case called agoryros, hung from the bel! on the left, although as che Scythians are shown shooting in the convenrional right-handed manner they must have used the right hand to draw the arrows. The case could hold 200 or more arrows. Rich men would bave the gorytos faced with a thin sheet of giJded silver, highly decorated with Greek designs in relief; poorer archers would make do with leather no doubt decorated witb dyed felt or leatber appliques. Herodotos says human skin, from enemy hands, was favoured for covering the bowcase because ofits whiteness. As well as hands, the ScyLhians used enemy skulls for drinking cups and scalps to decorate their bridles.

Secondary weaponry would be dagger or axe, and occasionally a javelin. Daggers or short swords were ofthe Persian aki11akes type, usually worn on the right in Persian style though one illus1ration shows. the scabbard on the left.

Again hilt and scabbard could be decorated with gold sheet.

By the 2nd cenrury Celtic weapons were occasionally used, under the influence ofthe Bastarnai then dominating rhe Dniester valley. One grave from rhe Crimean Scythian capitai conrained the helmer shown in figure 124i, Celtic swords and a Scythian bowcase.

77.

SCYTHIAN NOBLE

This man is based on the cent rai figure ofthe 4th-cenrury gold Solokha comb. Such heavy equipment, Like that found in princely graves, may bave become commoner under the influence ofrhe armou red Sarmarians. Certainly in 31 O che Crimean Scythians had a body of picked cavalry well enough equipped to charge straight through a Sarmatian army. This figure wears normai Scythian dress plus Greek greaves, old-fashioned Corinthian helmet and a scaled corselet. 77a shows another Greek bronze helmet, an "Illyrian" type (figure 74) with the characreristic double crest ridge ofthat type. lt has been cur down and had a peak riveted on. It was found in the Crimea along with a fragmenr ofbronze scales on a leather hacking, no doubt from a cuirass, and another armour fragment of1hin bronze s1rips on lea1her, like much later "splinred" arm and leg defences. Otber scale armour was iron; one iron cuirass was decora1ed wi1h copper scales at shoulders and the lower edge. This man's cuirass has shoulder-pieces but not pteruges. Greek bronze plate was occasionally used.

Weapons are bow and javelins, and a short sword has also been shown. He has a wicker shield slung on his back; 77b shows 1he shape, 77c is a similar shield of pelte shape, carried by a second cuirassed figure on the comb, who since he stands by a fallen horse is probably a dismounted cavalryman. In earlier cenruries sma([ round shields faced with iron or gold were used, but they seem to havc fallen out offavour.

131

78 and 79. SCYTHIAN INFANTRV Although the Scythians were famous for their cavalry they could produce large numbers of infantry. Some may ha ve been subjected peasams but tbese two are true Scythians. 78 is an archer no difTerent from his cavalry counterpart. He carries an axe, whicb could be decorated io gold. These are light-headed long-hafted weapons usually showo carried by dismounted bowmen (tbough some may be men who would mount to fight). One Scythian is illustrated using the eastem sagans (see 84) but this more conventional axe was obviously far commoner. 79 has two javelins and the characteristic Scytbian infantry shield of hide stretched over a wooden frame. 79a is a slightly different sbaped shield, showing tbe grip, once again from tbe Solokha comb.

80.

MAIOTIAN INFANTRYMAN

This is ooe of two figures sbown on a gilt goryros-place of the 4ch century defeating Scythians. Though his axe, shield and bowcase are similar 10 Scythjan types, and his companion has a Scythian shon sword and coat, this man is scripped to the waist in very un-Scythfan style, wbile both men are beardless, wear their hair shoner chan do Scythians, andare depicted without the Scythians' high cheekbones and flat noses. RostovtzefTtherefore suggests tbey are members ofone oftbe indigenous cribes oftbe nonhern Black Sea coast, the Sindi, Tauri and Maiotians, who were probably relaced co the ancient Kimmerians and to che Thracians. The T auri were independent till the 2od cenrury, tbe Siodi and Maiotiaos subject 10 the Bosporan Greeks. Similar tribesmen will have provided many oftbe peasant subjects ofthe Scytbians aod Sarmatians, who may have fought for their masters as infantry.

81.

BOSPORAN lNFANTRYMAN

The new army which emerged in che Bosporan state in the 3rd century was not very Greek in apppearance. Greek equipment Like Sia (a gilded bronze helmet, late 4th century, wich high Thracian skull bue wich a nasal) gave way 10 Sarmacian-style armour and weapons among che cavalry. This figure is based on a number ofterracotta figurines ofinfantrymen. He has a runic decoraced in Scythian sryle, loose trousers, andatali cap closest to Thracian rypes.

His shield is tbe Hellenistic rhureos, and weapons are javelins, perhaps spear, and short sword. Others were archers.

82.

EARLY SARMATIAN HORSE ARCHER

This figure is from a gold openwork plaque from Czar Peter the Great's Siberian collection, probably eastern Sarmatian work ofthe 3rd-2nd centuries BC. Sarmatians are almost always shown beardless, but occasionally with moustaches. He has a Scythian-style decorated coat and tighter trousers than usual. Clothing materials and decoration were probably similar to Scythians, but where gold plaques were used they werc not the rectangular figured Scythian types but small crescents, rosettcs, and othcr geometrie shapes. Thc western Sarmatians also adopted cloaks from che Greeks. Main weapon is a Scythian-scyle composite bow, wìth an unusual combined bowcase-quiver; the Scythian goryros was also used. Arrowheads were small, usually three-edged and barbed, of bone, bronze or iron. Swords were worn on the right and varied from short akùrakes types slightly longer than Scythian models to long slashing blades up to 4 feec/ l .2m in lengch. They could be highly decoraced with gold, at this period often with polychrome and jewelled inlay.

83.

RHOXOLANI CAVALRYMAN

The Rhoxolani were a major western Sarmatian tribe who moved west across the Don in the 2nd century BC and towards the end of that cenrury attacked the Bosporan kingdom. Strabo describes them at that period as lightly armed wìth helmets and cuirasses of raw oxhide, wicker shields, spears, bows and swords. This seems to be the equipment of the cavalry who made up the majority of at !case thc efTcctive troops (Sarmatian infantry being sometimes numerous but ncver much use) and grave finds confirm that most Rhoxolani were lightly armed horse archers unti I the 1st century AD. Cuirasses were probably of rawhide scales, perhaps lacquered red; an alternative possibility is the laminated hide armour shown on the base ofTrajan's Column and as figure 88a in TheArmies a11d Enemies o/ Imperia/ Rome. The helmets were probably leather with a metal frame like those shown in later art. The shape ofthe wicker shield is not specified by Strabo; the Scythian styles of77a and bare possible. The bow is kept in a birchbark quiver, and arrowheads were usually iron. The sword shown is a rclatively short typc with a ring pommel; a longer sword with spherical pommel was also in use. Strabo uses /011gcl1e for tbe spear, implying a fairly light spcar or a javelin.

Some Sarmatian cavalry however were already more heavily armoured, carrying a longer lance, as shown in figure 88 of The Armies a11d E11emies o/ fmperial Rome. This style emerges in the 4th and 3rd ccnturies in the Sarmatian Prokhorovka culture of the southern U rais, probably under the influence of the Massagetai. Tbe noblcs of tbe Rhoxolani probably fought in this way. Among the Siracae, in the Kuban valley, the nobles fought io this stylc, as armoured lancers with long sword and bow; the bulk ofthe people lacked armour, but unlike tbe Rhoxolani they did

132

133

use the long lance. Equipment ofthe armoured lancers included conica I bronze or iron helmcts and armour for bo1h rider and horsc, usually of bronze, iron or horn scalcs, bui iron pla1e seems to ha ve been used occasionally. The Siracae also used numerous ligh1 infantry, notably archers, bu1 Lhese may have been subject peasants ofa differem nationali1y, ali che Sarma1ians proper thus being to some ex1ent an aristocracy.

Some ac leas1 of1he Sarmacian womcn fough1alongside1heir men. Herodo1os records in che 5Lh cenrury 1ha1 a giri was not allowed 10 marry untiI she had killed a foe in battle. Female graves continue co include weapons through our period, both thosc of ligh1 horse archers (a grave in che Dnieper valley had sword and quiver of arrows) and of armoured lancers (a burial perbaps ofthe Sarmatian wife ofa Scythian ruler held scale armour and lanceheads as well as arrows). Such female warriors wore the same costume as the men; women usually wore thcir hair long and braided.

84.

SAKA HORSE ARCHER

This man is from a reliefofche Persian king Artaxerxes Ili Ochos (358-338). Costume and equipment are vcry similar to bis western cousin, 76. Tbe cap is caller, and Persian artis1s use clifferent cap s1yles co denote clifferent 1ribes. His coat is cui higher in front than a11he back, almost like a 1ailcoa1. His loose uousers are 1ucked imo shon leggings, tied at 1he top. The main weapon is a composite bow ofthe Scythian 1ype, but hc also carries a bronzeheaded sagaris. This axe-like weapon has a point like a pick, balanced by a small hammerbead. lt was uscd by che Saka and neighbouring nomad 1ribes in cenLral Asia :ind Siberia, and adopted from 1hem by Sarmatians and Persians, who used it occasionally. From his belt, which fora noble could be covered in gold plates, he wears an aki11akes thonged 10 hjs righe leg. The weapon has an iron biade and a bronze hllt wi1h 1he pommel in the shape of a crouching ibex. One Saka on a Persian relicfhas his aki11akes worn under his coat skircs, with only the chape of1hc scabbard visible.

Some Saka tribes at least could raise infantry, archers like this man and spearmen with light round shields, who are illus1ra1ed later in the an ofSaka India.

The British Museum has a small bronzc figurine from western Iran, drcssed very like 1his figure, who may be an early Parthian, 1he Panhlan state of course being founded by a Saka tribc. Tbey originally used 1he Scytbian s1yle ofbow, bui somewhen before 1he mid lsl cemury BC had replaccd it wi1h a longer but less deeply recun-ed bow very like 1hose previously used by the Persians. An example was found in 1he Yrzi necropolis on the banks ofthc Euphra1es. lt is 58" / 1.4 7m long. Parthlan bows seem 10 ha ve been built longer and heavier 1han 1heir Persian ances1ors, and 1be effec1s reponed in accounts of Parthian batti es with Rome in the I st cenrury BC suggests 1hey had grea1er penetrative power.

85.

EARLY SAKA ARMOURED CAVALRYMAN

Herodotos records the Massagetai, a major Saka people, already used armoured cavalry riding panly annoured horses in che Sth cenrury BC. Similar armoured Massagetai foughc for Darius at Gaugamela. This man's long sword comes from a Massagetic grave, but otherwise he is a fairly tentative reconstruction. The troops a1 Gaugamela wore more persona! armour tban Greek cavalry, as well as having horse armour. This implies somethjng additional to a cuirass; given the equipmen1 of their Persian neighbours and of the next figure, parapleuridia seem likely. Helmets may have been worn instcad ofthe cap shown. An armoured Saka cavalryman shot by a Macedonian catapult-bolt at che Jaxanes banle had a shield; 1hese are noc mentioned for the Massagetai a1 Gaugamela, alchough they were not normai cavalry eqwpmem at the time and rojghc bave exci1ed attemion ifused.

This figure carries a light reccangular shield made of35 round stjcks passed through a thjn leather shee1, in such a way as to produce a lozenge pattern, and carried by a broad leather loop. 85a is a similar, larger shield. Both are from the Altai culture's tombs at Pazyryk, perhaps the eas1ernmos1 outpost ofthe Saka peoples. T hey seem 10 have been hung at thc horsc's saddlcbow on the right when not in use. Apart from the sword he is armed with spear and bow.

Spears were probably stiU quite light at the starr of our period, bue were soon replaccd by the next figure's longer lance.

86.

LATER SAKA ARMOURED CAVALRYMAN

This figure is in fact based on a coin ofAzes, a lst cemury BC Saka king in India, but the equipmen1 ofthe las1 figure was replaced by heavier sryles like this well before the end of our period. The long hcavy lance, called kontos by the Greeks, was adopted in thc 4ch ot 3rd century. Ac thai Ăšme it appears in the graves ofthe Prokhorovka culture, whlch represems an intrusive eastern element, a branch ofthe Massagetai, living among and probably ruling, the eastern Sarmatians. The Massagecai may conceivably have copied the long spear from the xysro11 of the Macedonians, but dating of the graves is not suflĂŹciently precise to celi whether the ko11tos was in use before the 134

a

85

Macedonians arrived. Shortly before 200 BC the Seleucids adopted the complete cataphract panoply from the Parthians; their Saka neighbours and relatives will have been similarly equipped.

Azes' defensive arm'our is sirnilar to that of later Kushan cataphracts illustrated on frescoes from Khalchayan in Uzbekistan, not surprisingly since the Kushan replaced the Saka in India and neighbouring areas. I have used the Kusha.n example for details such as the collar of small upstanding lamellae, which seem to be indicated but not clearly shown on the coins. Azes wears what may be an ankle-length cuirass but is more likely a shorter cuirass and separate parapleuridia made oflarge lamelJae - long narrow armour plates laced together. His arms are protected by the tubular segmented defences offigures IO and 40. No bow or indeed any secondary weapon is shown; che Kushan figures similarly lack bows, and the arm defences may bave been restriccive enough to ma.ke archery impractical. He would probably carry a long sword and/or sagaris.

A (probably early) Parthian terracotta in the British Museum shows a similar figure, with the helmet shown in 86a, thrusting underarm with his spear held in both hands. He too lacks a bow. Neither of these figures has horse armour, but since the Saka had it much earlier, and since the similar Kushan cavalry have horses fully procected with leather scales, it is safc to assume that at least some troops ofthis type would ride armoured horses.

87, 88 and 89.

INDIAN INFANTRY

87 and 88 are based chiefly on the 4th-century descriptions of Alexander 's admiral Nearchos, preserved in Arrian's Indike, and on the I st century BC reliefs from Sanchi which indicate that equipment changed little throughout our era and beyond. The northern Indians were descended from Aryan invaders who had imposed chemselves on the previous population; by now the population was largely homogeneous, tali and fairly light-skinned (Greek writers compared them to Egyptians) though the lower classes might stili tend to darker skin. Some lndians wore beards but most are shown clean-shaven. Hair was usually long, often piled into a topknot. lt was tied with a strip of cloth, a simple turban, sometimes decorated with a spherical ornament. Clothing was a white conon kilt reaching to mid135

caJf, usually gathered up at the front, with the ends ofthe clorh and ofrhc sash used to fasten it hanging down; rhis could be quice ornate, though Persian sculptures show Indian tributar ics in simplc, shorter kilts, perhaps a locai variacion from che Persian domains in the Punjab. Sandals might be worn but common soldiers are often shown barefoor. Men invariably wore earrings, bui rankers rarely wore much other jewellery.

87 is the commonesc rype, the archer. The lndian bow was a long, heavy, powerful bamboo weapon, as long as che archcr was tali, with a hemp or sinew string drawn to the ear. Arrows were 3 cubits (4ft 6in/l .37m) long, of cane or reed, flighted with vulture feathers. Heads were usually iron, sometimes horn. Poisoned arrows are occasionally mentioned, usually co be condemned as unethicaJ especiaJly by Brahmin writers, though Greek sources record the Brahmins themselves using poisoned arrows againsc Alexander; presumably chey did not feel inclined to observe the usual courtesies against a foreign invader. Arrows were kept in a long quiver wom venically down the centre of the back, suspended by crossbelts. Archers on the Sanchi reliefs are shown with handfuls of arrows kepc ready in their right hands. A leather bracer was wom on the left arm, plus some kind offinger protection, perhaps an archer's ring for the right thumb. The Indian bow was very powerful, Arrian saying no shield or cuirass could stop its arrows, but it did not apparently meet with unqualified approvai from Alexander's officers. Curtius says bow and arrow werc too hcavy to be accurately aimed, and both he and Arrian record that the bow had to be braced with the left foot when firing. Curtius adds that the muddy footing at thc Hydaspes hindered this, thus contributing to the Indian defeat, but this is thrown in doubt by Arrian's contrary asserrion that the main barde was fought on ftrm dry ground. The bracing with the foot they describe is not shown in Indian art, which pictures the conventional stance shown here; obviously putting the bow against the foot would lead to the arrow pointing downwards, unless the archcr crouched, which no source describes. lt is likely that the bow was only thus braced to string it, which would probably be necessary with such a powerful weapon. In a way it may conceivably ha ve contributed to the inefficacy of the archers at the Hydaspes; the bow was strung only immediately before use, and since Poros' army was deploying hurriedly from column ofmarch when the main baule staned, some ofhis archers may stili have been struggling to string their bows when the Macedonians attacked.

88 is a javelinman, the other main type. ~earchos says they carried a light shield of raw oxhide, nearly as long as the wielder bui not as broad. lndian sources also mention shiclds of hide (peri or canna), cane and leather (kirika, perhaps le:nher on a cane framework), and wood (hasrikama). The shields at Sanchi are not ofNearchos' type, but shoner and broader (88a). The pattern could conceivably re present strengthening bars or the cane framework of a kirika, bui this is made less likely by the resemblance to pauerns on standards (see 95 a-e). Shields decorated with sun, moon and stars (perhaps like 95 dande) are also described. This man's spear is from Sanchi; 88b is a longer wcapon with three·pronged head, perhaps the trim/a ofthe texts, from 2nd century AD reliefs at Amaravati. The Anhasas1ra mcntions an all-iron spear, sakti, 4 cubits (6f1/I .8m) long, with leaf-shaped head and blunt butt.

89 is from 2nd century BC sculpture at Bharhut, and his cu rious dress has raised questions about bis identity. He wears thc usual lndian kilt but has unusual heavy boots and tied, long·sleeved jacket with braided edges. He also lacks earrings and wears bis hair short, both contrary 10 Indian cus1om - though not unique, as shown by 89a, an armed gatekccpcr from Sancbi, who however wcars usual Indian dress. An Indian tribe ofthe north·western hìlls, the Kamboja, were called "shaven-headed" because of thcir barbarously short hair - that this necd not mean completely shavcn is clear becausc the Greeks were also callcd "shaven·headed". This man's boots and sleeves would obviously suit tbe colder northcm mountain climate, and such an origin would explain his short hair, and mixture oflndian and unorchodox dress. Herodotos also mcntions north·western Indians whose customs rescmble those ofthe Bactrians, which would suit sleeved tunics like this.

89's sword is probably the type Nearchos says ali lndian infantry carri ed, with a broad biade 3 cubits (4ft 6in/ I .37m) long, used for powerful two·handed cutting blows, broughc down from above the head. lt is ca rried in a scabbard worn from a baldric over tbe right shoulder; che ends ofthe baldric pass severa! times round the scabbard andare sometimes knotted at the end. Although Nearchos says ali infantry carr ied chis sword, it is not in fact uni versai in Indian :m. As well as the Bharhut soldier it is carried by Indian tributaries in Persian sculpture and by one soldier ot Sanchi; but most ofche Sanchi figurcs have shorter swords as carried by 87 and 90, with a broad spoon-ended biade. h is usually wielded in one hand but can occasionally be used in two. The Arrhasastra speaks ofthree sword typcs; thc 111a11dalagra is straight with a round tip, thc 11is1ri111sa has a curved tip, the asiyasi is long and thin. The manda/agra is obviously ofthe type carried by 90. The 11isrri111sa has been variously identified; bue ifits "curved" tip is to be distinguished from che 111a11dalagra's "round" tip, it must indicate forward-curved swords modelled on the Greek kopis, which became quitc popular in India and stili exist in che fo rm ofthe Gurkha kukri. (This would suggest prolonged contact with Greek arms, and thus imply a date for this pan ofthe Artltasastra after the Greek invasions ofthe 2nd century.) This would leave the asiyasi which could be the big two-hander, though calling this "thin" seems odd. Asi is a common word for sword in the Mahabharata, with a variant cali ed maltasi, "great asi", perhaps appropriate for the two-bander. lndian swords were of good steel with hilts ofhorn, ivory, wood or bamboo root. Scabbards could be cowbide, goacskin or tigerskin.

136

87

I¡

a b

Other weapons included the club or mace, ca rried by epic beroes for whom mace-fighting was a highly developed art. Clubs could be used io ooe band, or two, or thrown. Alexaoder was attacked by a Mallian clubman, and his officers described the Sibi oftbe Punjab using clubs. 89b, probably wood bound with metal, is from tbe Amaravati sculptures. Others had faceted iron heads. Some were decorated witb gold. Axes were also used. Tbe Arrhasasrra mentions stones, both hand-hurled and for slings, while the Mahabharata mentions hillmen usiog stones. Probably the sling was popular io India as elsewhere with herdsmen to protect their charges from wild beasts, since ammunition was freely availablc io hill country, and thus it may havc occasionally bcen used io battle.

90.

INDIAN ARMOURED INFANTRYMAN

How common armour was io India at this time is somcthiog of a problem, as different sources give differenl imprcssions. In thc cpics, bcrocs aod nobles always wcar cuirass and belmet, often described as gold or silver, or more rcalistically as copper or iron sometimes decorated with silver or gol d. Tigerskins were also used as protection.

Ordinary soldiers are also sometimes said to be armoured. The Artltasasrra mentions both armoured cavalry and infantry, and lists armour of metal rings or plates, horn plates, or padded cotton. But this profusion of armour is not in evidence in early lndian art, where even chariot and elephant warriors, the nobles and leaders, are unprotected.

Litcrature and art seem to follow contradictory conventioos, and so the Greek literary evidence is espccially valuable as il is not bound by thesc conventions. The Greeks confirm kiogs wore armour; Poros is said 10 have worn "a corselet exceedingly tougb and closely fitting", decorated witb gold, which lcft only his rigbt sboulder unprotected. But they make no reference to armour for ordinary soldiers; even the mercenaries hired by the Assakenoi, professional troops of higb standing whose fighting is describcd in some detail by Diodoros, have javelins and shields but not armour. Nor does Megasthenes describe the Mauryans using armour. Io generai it seems armour was very scarce, the epic mcntions of it may largely be poetic convention, while the armoured infantry and cavalry ofthe Arrltasasrra werc eithe r very rare or else may belong to a later period entirely.

137

This figure is based on one ofthe very few early lndian representations of armour, a group ofsoldiers at Sanchi with bows and short spoon-bladed swords. Thc armour is offive bands, perhaps leather, protecting the lower torso but leaving the up per bare. lt could be the jalika, "hcadless armour" of the Arthasastra. To sec armoured archers is surprising because tbe Artlzasasrra distinguishes between "armoured men" and archers, implying the former were not archers, presumably the javelinmen.

When armour does become common, in Kushan art ofthe lst cenrury AD, much ofit rcsemblcs Kushan lamellar styles, so is not Likely to be a guide to earlier lndian armour. One exception is the quihed armour shown in 90a. lt has pteruges presumably inuoduced under Greek influence in the 2nd century BC or later, but otberwise may be like the padded couon ofthe Arrhasastra. One king's armour in the Mahabharata is "decked with a hundred suns, a hundred circles, a hundred spots and a hundred eyes"; it recalls the much later Moghul "coat ofa thousand eyes", a sort ofstudded brigandine, and may represent similar quilted armour reinforced with studs.

91.

INDIAN NOBLE

In view ofwhat has been said abouc che uncenainties concerning the form and frequency oflndian armour, it will be realised that this reconscruccion of a king or greac noble is in some respects hypothetical. He has a turbanwrappcd helmet, and a cuirass bascd on armour scyles shown in later art, of scalcs which are laced together and do noc overlap horizontally. lf Poros' armour left only his right shoulder bare, it could be because he carried a shield on his left arm.

Ocherwise che figure does show some feacures of nobles' appearance. Cloching would be dyed rather than che commoners' plain white, in bright colours such as red and green. lt could be striped, chequered, or flowered. The Malzabharata mentions royal bodyguards in red. Beards were also dyed, in white or purple or red or green; wcaring beards may only bave been common among the nobles. As well as che usual earrings (ivory for tbc rich) jewellcry was worn in abundance. Nobles worc whitc lcather sandals wich multicoloured built up woodcn or lcatber soles, to increase tbeir beight.

A man ofchis rank would usually ride a chariot or elephant - the Arthasastra says "An clephanc or chariot should be che vehicle for the king, guarded by cavalry. He sbould mount chat vehicle ofwhich the army mosĂšy consists or in which hc may be trained". Ac the Hydaspes Poros rode an elephanc, Alexander's ally Ambhi ofTaxila a horse.

138

92.

INDlA N NOBLE' S PARASOL-BEARER

lndian nobles were invariably accompanied by an at1endant carrying a parasol. They are often shown riding on chariots or elephants to shade their masters. The parasol usually seems to bave been white, wi th golden or silver ribs. Other anendants carried fans or flywhisks, but these may not have been carried on the batrlefìeld.

93.

INDlAN CAVALRYMA N

Cavalry were of fairly low status in lndian armies, chariots and elephants being the prestige arms. Most, if notali, were unarrnoured javelinmen. Nearchos says they carried two small javelins "like those called sa1111ia ", which since sau11io11 is used fora wide variety of different javelins is nota lot ofhelp. The Arrliasasrra mentions cavalry spears called kuma, bltindipala and liaraka, while rhe Mahabharara has cavalry throwing spears called prasa, somctimes cailed barbed. Short swords would also be carried. Nearchos says cavalry carried shields, smaller than those ofthe infantry but not otherwise defìned. This fìgure's shield is carried by a cavalryman at Sanchi. 93a seems 10 be a development ofthe same shape, perhaps with the aim of catching enemy blades between the "horns". lt is carried at Amaravati by an infantryman with a short sword. Although Nearchos irnplies ali cavalry had shields, there are severa.I illustrations ofhorsemen without them, including some at Sanchi. Cavalry guided their horses by a whip strapped to the left wrist.

il is possible that some cavalry may have used bows, as there is a reference in the Maliabltarara to princes shooting bows from horseback. The usual long lndian bow was probably 100 unwieldy to be used from a horse, but occasionai references 10 horn bows suggests the handier composite bow was known. However Indian mounted archers who appear in later art wear Saka-inspired costume, including trousers and pointed cap, although their features show they are Indians and not foreigners. This suggests that mounted archery in India was associated with the Saka, and was adopted from them after their invasions of the early lst century BC, the epic references being fairly late accretions to the tradition.

Whether any Indian cavalry wore arrnour in our era is uncertain. Thcrc are rcferenccs 10 armourcd cavalry in the epics, and in the Ar1'1asas1ra - "a centre of armoured horse, flanks and wings of unarmoured". But of cou rse the dates ofthese passages may be dubious, and they contrast with the absence of armoured riders in Greck dcscription s o r in lndian art. They may therefore be late passages inspired by the armoured cavalry of thc Greek and Saka invade rs. There is also a mention ofhorse armour in thc Arthasastra; even ifthis is rcliable (I have seen so many translations from Grcck and Lat in rendering ge nerai terms for arms and equipment as "armour" that I am vcry wary about translat ions whose originai I cannot check; and the list of armour types in the Arrhasasrra makcs no mention of protection for horses) it may referto chariot horses, who do seem occasionally to have been armoured (see figu re 177). On the whole the low status of cavalry makes me doubt whethcr any werc arrnoured, at least with a nything bcttcr than quilted couon, with the possiblc exception ofthe royal escons mentioned under 91.

93 94

a

139

94.

INDIAN PALACE GUARDSWOMAN

Many Indian kings kept female bodyguard units; thc Sclcucid envoy Megasthenes records Chandragupta Maurya attended by armed women, some Tamil kings had similar femalc guards, and Lhe 2nd cemury AD sculptures from Amaravati show severa! armed women, on one ofwhom this figure is based. Her headgear is probably cloth, wound around ber piled-up hair. The sword is che only wcapon shown. Guardswomen are assigned by later dramatists to many early, legendary kings, but Chandragupca mny have becn the first to use them in reality. They are conventionally called Greek, but I doubt ifChandragupta's wc re; Megasthenes, a Greek himself, would surely bave remarked upon such an unusual sight as armcd Greek women at a fore ign court. Greek women, like male Greek mercenaries, may bave been first hired as the Greek military class sought employment as their kingdoms broke up in the lst cenrury BC. Megasthenes describes Chandragupta's guardswomen escorting him on the hunt, riding horses, elephants and chariots. lfthey actually fought in baule, which is perhaps doubtfu l, thcy may thus have becn mounted.

95.

INDIAN ST ANDARDS

The comrnonest rype ofstandard in early lndian art is a flag suspended from a cross bar, carrried on a relatively short pole by a standard-bearer squatting on an elephant's back. 95a is from 2nd century BC sculpture from Sarnat h; the main figure and 95b-e are from Sanchi.

The Mahabharara often mentions chariot standards; every hero has his own ensign on a pole rising from the back of the car. The only representations of chariot standards I have been able to find in art, 95f and g, are much later than our era (6th century AD) but some ca rlier standards may havc been similar images on top of poi es. Arjuna's standard for instance had the figure ofa monkey atop the shaft, with flags bearing other animals below. Some ofthe other emblems described may be blazons ofnags - moon and stars (pe rhaps like dande here), or a water-jar on deerskin.

Othcrs sound more like precious metal images on poles, like a jewelled peacock, silver lion, golden deer, silver swan, golden p loughshare, silver boar, golden parrots, jewelled elephant. A golden lion's tail or elephant rope may be the prototype ofthe streamers of95fand g. Early Tamil poems mencion Mauryan chariots with white pennons.

The Mahabliarara also has a reference to cavalry flags, but I bave found no description oftheir form, or that of infantry standards.

M usical instruments were importane; Nearchos says that no nation was as fond of rnusic and dancing as the lndians, and they went into baule to the sound ofdrums and cymbals. Gongs and conch-shell trurnpets were also used. Their main purpose seems 10 have been to boost morale, but the Arthasasrra mentions the use of drurns and flags for signalling.

96.

INDIAN FOREST TRIBESMAN

Based on descriptions in the Ra111aya11a, this man represents the "wild tribes" ofthe jungles, remnants ofthe preAryan peoples sometimes employed as auxiliaries. Short and dark, he is barefoot, dressed in a skin kilt with feathers in his hair. Main weapon seems to havc been the bow. These primitive tribes were probably short of metal, so arrowheads may have been ofhorn, and swords are unlikely.

97 and 98.

CARTHAGINIAN CITIZEN INFANTRY

At the Krimisos in 341 the Sacred Band of Carthaginian citizens fought as a wcll-d rilled phalanx, with iron cuirasses, bronze helmets and large round white shields. Stelae frorn Carthage, many ofthem unpublished, are described as showing similar troops in muscled cuirasses with pteruges, conica! helmets and round shields. 97 is such a warrior. I doubt however ifthe cuirasses at the Krimisos were acrually the same muscled type, as the mid-4th century seems too early for iron plate, even if Punic armourers were in advance ofthe Greeks. Scale armour (sec figure 100) is a possible alternative. Another is lamellar; there is no direct evidence for tbis, but the Etruscans alone in the western Mediterranean used it, and it is tempting to suppose that tbey adopted lamellar armour from thc Canhaginians, with whom tbey had dose Lrading contacts, and who could have brought it west from their Phoenician homeland, where they were in contact with lamellar-using Assyrians. Linen armour was also uscd (scc figure 102). An iron conica! helmet, with ears embossed at the sides, was found in a 2nd¡century BC Numidian prince's grave at el-Soumaa. Although this example is Numidian, it is very like Assyrian and othcr eastcrn styles (compare thc helmet of figure 5) and must have been introduced by the Canhaginians. No doubt it is this type shown on thc Carthaginian stelae, though depicted in rather crude triangular form. Greaves do not seem to be specifically mentioned, but are likely for the heavily-equipped Sacred Band.

140

The large round shields at tbc Krimisos will be of tbc cype depicted on monumental reliefs at Cbemtou and the Kbor Klib, perbaps representing Zama and perhaps later. ln fact thcy are probably the shields of similarly-armed Libyan troops rather than ofCartbaginians, since tbe Sacred Band seems 10 have disappeared by this time and othe r citizcn troops, as we will see, did not use round shields; but the style is probably the same. The shields are about a metre (39") in diameter, at the high end ofthe size range for Greek "Argive" shields, which is wby Greek observers thought them large. Tbey are quite like the Argive sbield in shape, with a broad flat rim, but the centre is slightly recessed. The shield would be ofwood covered in leather, perbaps elephant hide, with the rim faced in bronze. 97a shows the likely cross-section, with the bronze facing denoted by tbe thick black line. The sculprured shields bave blazons, but the descript ion of the Sacred Band 's shields as wbite may mean they were pia in white witb no dcvicc.

Earlier Phoenicians carried light rimless shields; some are shown rising lo a point, probably a spiked boss, but this seems highly impractical for phalanx warfare, at least for ranks bchind the first as the spikes would press uncomfortably ifnot dangerously into the backs ofthe men in front.

Although the Sacred Band's equipment and formation was similar to that ofGreek hoplites, tbey do not seem to have used the sa me long spear, since Punic stclae show armoured infantry with broad-bladed spears only about their 141

own height, while the Libyans anned in a similar style also used short spcars (see figure 102). Swords were sbon and straight, about 26" /65cm long, and similar to early Celtic and Spanish types. The Kbor Klib reliefs also show swordhilts ofthe Hcllenistic kopis type. Swords would be worn on baldrics, since the Canhaginians traditionally did not beh their ruoics.

The usual Carthaginian dress was a knee-length or longer tunic, loose and full, normally with long sleeves though sleeveless garments are occasionally sccn. From about 300 (when Canhage began to adopt some elements of Hellenistic culture) shoner Greek¡stylc runics were occasionally worn. Silius ltalicus mentions troops "from Carthage" in red tunics, but these were skirmishers with light round shields, and I suspect them ofbeing Libyans like 103. Carthaginian dress also included some form ofunderwear, a custom that the Greeks, who did without, regarded as ludicrously effeminate. Some Carthaginians were bearded, espccially older men, but many shaved.

The Carthaginian citizen infantry at the time ofthe chird war with Rome, however, are described by Appian and Strabo as having swords, /011gchai (light spears or javelins) :md tl111reoi (ovai shields). This is essentially the same style of fighting as that of the Phoenician marines in Xerxes' flect in 480, except that the western tl111reos adoptcd from Gallic or Spanish merceoaries has rcplaced the older Phoenician light round shields. The use of thureos and javelins is confirmed by Punic-influenced Numidian stelae (figure 108). Probably only che Sacred Band was ever more heavily equipped, and after this disappeared about 300, ali Carthaginian citizen infantry would have used javelins and thureoi. lt is highly unlikely that such troops would use metal armour, though the traditional linea cuirasscs are perhaps possible. 98 is such a soldier. The Canhagioian citizen troops in the second line at Zama would be like this, as well as the troops ofthe third war, and no doubt the citizen leviesofthe other Puniccities that occasionally appear on the banlefield. In the Phoenician homeland, Sidon's numerous "wcll trained and strong" citizen soldiers fought against Persia in 350, while T yre's militia put up a good defence against Alexander. They too would probably fight with javelins, but would keep the old light round shields instead ofthe western thureos.

99.

CARTHAGINIA N CAVALRYMAN

A terracotta disc, probably in fact dating before our era, shows a Canhaginian cavalryman who provides the hclmet, shield, spear and long tunic ofthis figure . His torso is hidden by the shield, but the armour types discussed under 97 are likely. T he helmet seems to be a hemispherical bowl with simple cheekguards hanging from the rim in Assyrian style. The shield couJd perhaps have a recessed centre like 97's. The citizen cavalry was recruited from the nobility, so would probably be elaborately drcsscd and equipped. Silius ltalicus mentions a purple saddle clocb.

100.

CARTHAGINIAN OFFICER WITH ST ANDARDS

This man's hclmet and standard are from a crude and incomplete carving ofa borseman sometimes calied "the stele ofthe Punic Mars" whicb has been dated to the time of the wars with Rome. The helmet has a conventional classica! crest, but its large crest-knob, bell-shaped skull and thick rim indicate that it is ofthe Celtic Montefortino type. This was so common in the Western Mediterranean, used in Gaul, Spain and ltaly, that it must almost inevitably have found its way into the Canhaginian army, as H. Russell Robinson suggested. Silius ltalicus mentions scale armour for officers, and also mentions senior officers at least resplendent in purple and gold; Polybios confirms purple-clad geoerals. Whilc Silius is often rejected as an historical sourcc, and cenainly throws generous amounts of myth, fantasy and rhctoric into bis narrative, many ofhis references to armour aad equipment can be confirmed from other sources, and he may bave used reliable sources for such "loca I colour". He also mentions red, yellow and white helmet crests in the Carthaginian army; some helmets have feathcrs, some have more than one crest.

Thc stnndard this figure carrics is from the " Punic Mars" stele. lt has been described as a ring, but there is a clearly visiblc break at the top; it seems in fact to be a crescent, with the ends slightly rurned out, and with a separate short crossbar to which the dangling ribbons are probably attached. 99a, from a reliefofa Carthaginian warship, adds a sun-disc to the lunar cresceot; this was common Punic religious symbolism. The standard is also shown, somctimes without the ribbons, on a number ofgrave stelae.

There are also references to Carthaginian trumpetcrs. One stele from Cina shows an instrument like the Roman cor1111 {see figure 141) which is probably the trumpet meant.

101.

LIBYAN CAVALRYMAN

Based on a terracotta figurine of a north African cavalryman. His hair is arranged in ticrs of ringlets like the Numidian style (see figure 105) but he differs in severa! respects from Numidian cavalry. He is seated on a 142

saddlecloth; he certainly wears shoes, and a ridge at his right wrist indicates long sleeves. Visible muscle detail on his chest and a sharp out-rurncd ridge at the waist suggest he is wearing a Hellenistic muscled plate cuirass. In addition his shicld has a rounded boss and raised rim, unlike the Numidian style. He wears a crescent pendant around his neck, recalling the Libyan omament of figure 104. I suggest therefore that he is a Libyan or Libyphoenician heavy cavalryman io Cartbaginian service.

Like 99, he rides an unarmoured horse. Recently Petcr Connolly, followed by other writers, has illustrated Carthaginian cavalry with lamellar horse armour ofthe type worn by figure 167, depicted on the Pergamene reliefs. h cannot be stressed too strongly that chere is no evidence for this. No horse armour is shown in the sources for this figure and 99, and it is also abseot from the literary evidence. The suggestion that Cartbaginian troops used some Hellenistic equipment and will therefore have used Hellenistic horse armour is unfounded, since the horse armour was not in generai use among the Hellenistic armics but is unattested outside Asia. In the very unlikely case of any Canhaginian cavalry usiag horse armour despite the lack of any evidcncc for it, the Graeco-Oscan style offigure 175 is far more likely than the Asiatic style, as the Carthaginians may have encountered it used by Campanian cavalry.

102.

LIBYAN SPEARMAN

These troops fought io a phalanx like the Carthaginians thcmselves, aod their equipmenr must have been broadly similar, sin ce the Grecks at the Krimisos, seeing a large phalanx of mixed Carthaginiaas and Libyans, assumed they were ali Carthaginian. (This need not imply that the cquipmcnt was identica!, as the Carthaginians were at the front of the column of march, and the Libyans behind would not havc been too clearly visiblc even from the Greeks' hilltop position, there stili being some mise about.) This figurc's shield and cuirass are from the Chemtou relicf.

Thc shield is of the type discussed under 97. Most of thc recessed middle portion is covered with a panern resembling overlapping scales or feathers, edged with thongs. This is the pattern used by Greek artists to represent the aigis of Athene; and the comparision is important because Herodotos says the aigis was derived from Libyan gar ments offringed goatskin. lt seems likely, therefore, that the pattern represents goatskin here too, and indicates a Libyan origin, and that the shield is a type carried by regular Libyan troops and decorated in a style reminiscem of goatskin shields carried by Libyan tribal levies. The centre device ofthe shield is damaged and it is not clear what it represents. Other Chemtou shields have Greek-style blazons including eyes, clubs or thunderbolts. Silius ltalicus mentions a horse's head device somewhere on Hannibal's armour, perhaps a shield¡blazon; the horse is a known Carthaginian emblem. Suo-disc and crescent moon, lotus, lion and palm tree are othcr popular Carthaginian symbols which rnight fearure as shield blazons.

The cuirass associated with this shield is ofHellenistic cut with shoulderpieces and pteruges ac the waist. lt is shown with a smooth surface, but Peter Connolly suggested it is actually mail, mailsbirts often being sculptcd smooth and 143

then painted 10 indicate che materiai. However short mailshirts with pteruges do not seem 10 be known 1his early, or indeed before 1he lst century AD when che Romans start to use them, while the fastening ofthe shoulderpieces (a metal ring anached to the shoulderstrap and another attached to the body ofthe cuirass, joined by a shon thong or strap) is not found on other mailshirts of1his period, but is reminiscelll ofthe fastening ofGreek Unen or leather armour. Linen armour was certainly used at Carthage; Pausanias saw three linen cuirasses at Olympia which had been caprured from che Canhaginians by Gelon of Syracuse, probably in 480 at the H imera. (In 1he same year Xerxes' Phoenician marines wore lincn corselets.) Since Hannibal fitted out his Libyans with captured metal armour, they were presumably lighter armcd before this, and the linen armour would be appropriate. Even ifthe monument represents arms taken at Zama, when Hannibal's ve1erans presen1 would have bcen in their metal armour, the linen armour could stili be that ofthe Libyan levies ofthe second line.

This figure's helmet is a Hcllcnistic type similar to that of63; it is shown on a coin s1ruck by Hamikar Barca in Spain soon after 237 BC, sowas probablya common type in the Punic armyduring thewars with Rome. 102a shows a possible alternative type, a simple Attic he lmet without crest or cheekpieces, shown occasionally in Punic art and notably on a scarab from Utica, worn by a warrior with rimmed shield, shon sword, and spear. lt is not of course cenain 1hat he is a Libyan, bui 1he helmet is of a simple type that could cheaply and easily be mass produced fora large army.

Pe1er Connolly has rather curiously suggested that the Libyan or Libyphoenician infantry formed a Macedoniansryle pike phalanx (despite having himself suggested that they carried the large rimmed shields of the Chemtou monuments, which are noi compatible with the use ofthe pike) and has again been followed by other authors.

However there is no evidence for this. Connolly seems to be under the impression that Polybios refers to pikemen in Hannibal's army; despite the Loeb translation, be does not. Polybios speaks of /011gchophoroi, which the Loeb 1ranslator renders pikemen, and it is clear from the roles Connolly assigns to his pikemen that he accepts these /011gchophoroi and the Libyan heavy infantry as one and the same. But /011gchophoroi are troops armed with the /011gche, a word which is used fora number of dilfereot cypes oflight spear, usually thrown. Thus light cavalry are cali ed /011gchophoroi(see note to figure 12} while lndian infantry, or Macedonians fighting duels or naval battles, are specifically said 10 throw /011gchai. lt is also used to mean the Imperia! Roman /a11cea, a related word describing a similar light spear. Polybios' /011gchophoroi are found screening deployment, seizing hills, and performing other light infantry tasks, often in conjunction with slingers. They are called light infantry (euumo1) in the account of Hannibal's famous ruse with the oxen, and differentiated from the Libyan heavy infantry at the deployment for Cannae. Clearly the /011gchoplroroi are the conventional light javelin skirmishcrs ofthc army. (They need not be the same typc of soldier as 98, although both carry /011gchai; while the /011gchophoroi seem, especially from their association with the slingers, to be conventional skirmishers, the large shiclds ofthe citizen troops suggest they are closer to rhe Hellenistic tl111reoplroroi.) So there is no evidencefor the Libyans being pikemen; good evidence against is provided by a passage of Plutarch's life ofMarcellus, which says that they had difficulty in dose combat against the Romans because they did not throw javelins (probably sa id with pila in mind) and their spears were shorter than the Romans' weapons.

103. LIBYAN VETERAN SPEARMAN lt must often have happened that ancient armies finished campaigns looking vcry difTerent than when they started, having rcplaced or supplemented their originai clothing and equipment with articles looted from the enemy or the locai populace. Hannibal's army in ltaly is the best documented example. Crossing the Alps they were supplied with clothing by a friendly Gallic chief. At Cannae the Libyans appear ali armed in Roman style, with the best ofthe equipment captured at previous victories. Only the Libyans were thus armed; the Gauls stili appear stripped to the waist, the Spaniards with their brilliant white tunics unobscured by armour. Ifthere was enough loot available to ofTer to them, they presumably preferred not to hamper 1heir native agility.

This man therefore wears a Gallic runic and shoes, having discarded in the ltaUan beat the trousers and cloak he would also have worn when crossing the Alps. He has a bronze Montefortino helmet, iron mailshirt (with a waistbelt to gird it in tightly so that ali the weight does not drag on tbc shoulders; probably a necessity despite the Libyan aversion to belts) and strapped bronze greaves, ali Roman 1001. A very finely omamented Oscan triple-disc cuirass, ofthe rype sbowo in figure 147a, found in north Africa, was probably a similar anicle ofloot brought back from ltaly by one of Hannibal's veterans. He also ca rries a large Roman scutum. lt has been suggested that the Libyans would keep their round shields to avoid being mistaken for tbc eoemy, but this seems unnecessary as there was little danger of such confusion when rwo formed bodies of heavy infantry met. They did not, bowever, apparently adopt the pilum, as the passage from Plutarch cired under I 02, indicating that the Libyans used shon spears, refers to a period after the assumption ofRoman armour ar Cannae.

144

102

104.

103

104

LIBYAN JA VELINMAN

The javelin was the t raditional Libyan weapon, used in Xerxes' army in 480 and shown in Persian sculprure. When che nacionaliry ofCanhage's javelin skirmishers is meotioned, chey are usually called Moors, but it is likely chat some were Libyan. The Libyan cribesmen who sometimes foughc for Kyrene, and Carthage's disaffccced subjeccs who sided with Agachokles ofSyracuse and tacer joined che rebellious mercenarics, will mostly havc bccn like Lhis man. He is partly based on a descripcion of skirmishcrs in Canhaginian service, probably Libyans, barefooc, dressed in red, with short swords and light round shields. Like Carthaginians, the Libyans wcre noced for noi wearing bclts. Herodocos says they wore leather cloches, somecimes of goatskin with the hair rcmovcd, dycd red.

Persian reliefs show Libyans in ankle-length robes with cloaks which may be of skin, but shortcr cunics are perhaps more likely for war. This man's head is shaven excepc fora plaited cresc with a crescent ornament over the forehead; this is from a grave scele ofthe 2nd ccnrury AD, but in the Sth BC Hcrodotos describes the style as characteriscic of the Makai, so it would certainly have been in use in our era. He also says the Makai used ostrich-skin shields.

Curious hairstylcs wcre characteristic of othcr Libyan tribes too; thc Machlycs worc their hair long at che back of thc head, the Auses at the front, while the Maxycs shaved the left side ofthc head but grew the hair long on the right, perhaps in a plaited scalplock. Libyans on Pcrsian reliefs have short hair and shon beards. Some Libyans wore red body-paint or tattoos; Herodotos says che Gyzantes and Maxyes scained their bodies red, while Silius ltalicus mentions paintcd Moors. Much earlier Egyptian are shows Libyans wich cauoos down the :irms, as shown here, so this may well be what is meant.

105.

NUMIDIAN OR MOORISH CAVALRYMAN

Thc Numidians, "by far thc besi mountcd croops in ali Africa'', and cheir Moorish and Gactulian neighbours were lighcly armed and unprepossessing in appearance, "horses and riders tiny aad lean; che horsemen were wilhouc armour, and withouc weapons, apan from the javelins they carry" (both quoces from Livy). They :ire shown on Trajan's Column and in less decail bue with che addition of a cloak, on a coin ofHannibal's ally Syphax. They were somecimes bearded, wore their hair in ciers of1igh1ly curled ringlecs, and had dark but not black skin; they were, like the Libyans, Berbers perhaps wich some Negro admixrure. Dress was Limiced to a shon, baggy, sleeveless runic, probably undyed wool, pinned on che shoulders and behed. A pan from bis yew-shafted javelins with iron heads and buttspikes, this man carries a stick to guide his horse and che rypical light, round bossless leather shield, slightly convex wich a narrow rim. Hand weapons are noi usually mentioned, so probably nothing more th:m a knife was usually carried.

145

e a

110

d

146

e

b

106.

NUMIDIA N PRINCE

A Roman coin ofthe end ofthe 2nd century BC shows Bocchus of Maurctania surrcndering Jugunha ofNumidia to Sulla. Both African kiogs wear calf-leng1h sleevcless robcs with large armholes open to che waist, where a broad belt or sash is wom. Bocchus adds a short cloak fastcned on the right shoulder. Rome's client Massinissa was presented with purple cloaks by the Senate; Juba I in che I st ccntury BC wore a rcd one. Syphax and later rulers wore the white diadem, che Hcllenistic symbol ofkingship, ticd round che hcad. There are severa! rcferences to Numidian leaders carrying swords, Massinissa prcscnting an ivory·dccora ted sword and scabbard as a gift. Some princes wore armour; Romc gave Massinissa rwo cavalry panoplics, with cuirasses. Thc 2nd century BC el-Souma tomb held a sword, spears, the helmet worn by 97 and a corroded iron mailshirt.

107 and 108.

NUMIDIA N INFA NTRY

107 is essentially the same as his mountcd countcrpan. Most Numidian and Moorish infantry were javelin skirmishers like this; the few archers and slingers would probably look similar. However experiments like Syphax' Roman-rrained infantry represent an attempt 10 create close-fighting troops, and they were probably more like I 08.

He is based on a Punic-inOuenced stele of perhaps thc 2nd ccntury BC from the Numidian city of Cina (Constanti ne) which provides his ovai tlmreos with spine and winged metal boss, two javelins, sword, and helmet, which is ofthe conica! Punic type with a thick rim. Thc stele has been identified with that of a Spanish mercenary but bears the Punic name Abdasart. He is either a Phoenician in Numidia or, more likely, a Numidian who has becn inOuenced by Punic religion and culture to the extent of adopting a Punic name (probably for religious reasons; it means "servant of Astarte"). The impact of Punic civilisation on Numidia was great, and continued eveo after the fall ofCarthage. Syphax' Rom:m·trained troops and Jugunha's latcr army at !cast uscd standards; there is no record oftheir form, but they may well have used Punic religious symbolism likc those shown in figure 100.

109 a nd 110.

IBERIA N S CUTARll

Sc11tarù~

so named by the Romans for their large ovai shields, werc the majority lberian infantry type. Costume was a short V-neckcd linen or woollen tunic, usually with short sleeves but occasionally long. Polybios and Livy describe Hannibal's lberians in "dazzling white" tunics edged purple; Polybios says this was nationa1 costume, rather chan a Carthaginian issuc uniform, and Spanish art often shows tunics with various styles ofedging. Some woollen tunics may instead ha ve been black - see note to figure 113. A broad leather beh, which for tbc rich could be covered with gold or silver plaques, was worn. Infontry are not usually shown with cloaks. Some warriors are barefoot, most wcar shocs or calf-lcngth boots.

Thcsc two figures are bascd on thc most decailed ancient representations of scutarÌI~ from the earlier phase of sculpture at che lberian fonrcss of Osuna, 3rd or 2nd century BC. Headgear is che sinew helmcts described by Strabo, onc with a horschair crcst.

Sc11tarii carricd various hcavy throwing spcars. Most discinctive is thc slender aU-iron javelin of 110 and I IOa, called by Diodoros sa1111io11 (n word used for severa I different kinds of javelin) and by Livy soliferreum. le was used ali over Spain, vari es from l.6·2m (Sft 2" ·6ft 6" ), has a small barbed head and capers co a point at the buct. The centrai shaft broadens to forma grip, usually square or hexagona1 in section, somecimes more elaborate. 109's weapon has a 36cm ( 14") socketed head like n light pilum. I IOb is afalarica as described by Livy. lt has a 3ft (90cm) iron head, and a fir shaft whi~h is round in seccion cxcept that it is square where it joins the head; probably the head was sunk into the shaft and riveted chrough the broad square portion, like the Roman chick pi/11111. The square scction was bound with pitch-soaked 1ow which was set alight before the spea r was thrown. Thrusting spears were also common in Spain, and some infantry may havc carried them, but sou rces only speak of a shower of thrown weapons followed up with thc sword.

I 09 carries the commonest Spanish sword type, a variant of the Greek kopis cali ed by modem writers espada falcata. h was uscd ali over Spain from the Sth century but was most characteristic ofthe souch and south-east. Unlike the kopis the back edge is sharpened near che poinc, so the sword can thrusc as well as cut. Some examples have a bar joining pommel and guard to protect the hand. The biade was 35-52cm (13-20") long. 110 instead has a straight sword with a 52cm (20") biade and long tapered point. This is a more characteriscically Celtiberian weapon, the gladius Jiispa11iensis of the Romans, also a cut-and-thrust weapon. Scabbards were worn on the left from a ba Id rie; 1hey were usually wood or leather with metal edging and bands co hold the suspension rings, but some were ali iron.

A small knife could be attached to thefalcata scabbard with its biade the same shape as the sword, as shown by 115 and 116. Daggers were common, varying from short triangular blades to small versions of thc sword shapes.

The shield is Oat like the Gallic type, but with metal edging and with che centrai handgrip parallcl 10 thc long axis

147

112

a

b

ofthc shicld. I !Oc-e are lberiao shicld pauerns, c and d from 2nd-century Valencian vases. Another shield on the same vasc as d has the line of"S" shapes down both sides. 1 lOe is a cavalry shield ofthe 4th-2nd century.

111.

IBERIAN CAETRATUS

Based chiefly on a vase from Murcia. Some caetrati are shown bareheaded. Hc carrics 1he caetra from which these light infantry took their name, a round le:nher buckler 1-2 feet (30-60cm) in diame1er, with a rounded or conica! .metal boss, held by a centrai handgrip :md a shoulderstrap by which it could be slung on the back. Armament is javelins andfalcata. 111 a is the ma in figure's shield pauern, bis from a bronze horseman figurine, c is a trophy from a Roman coin.

112.

CELTIBERIAN SCUTARIUS

Thc Ccltibcrians fought chiefly in Spanish style, bui with some Celtic or 01her distinctivc features. Diodoros says 1hcy wore black doaks, probably thc Celtic woollen sagum, sinew greaves, and bronzc helmets with crimson crests.

The ubiquitous Gallic Montefonino hclmet is oftcn found in Spaio, and while many ofthe examples are no doub1 Roman, the Celtiberian bronze helmet can hardly be of aoy other typc. Both greaves and helmets with Monteforcioo-style crest-knobs are shown worn by two caetrati who also carry 1honged javelios and short straight swords, on a Celtiberian vase painting. Diodoros meotioos round shiclds, thc usual caetra, aod also sc111a which he calls Gallic, perhaps indicating slight differeoces from the Iberian type.

Though Celtiberians used the falcata, various straight swords were commoner. Il 2a and b show a short sword, little more than a dagger, developed locally from antenna-hilted swords ofthe early Celtic Halstatt culture, charac¡ teristic ofthe beginning ofour period. The biade is 81/2-14 inches (23-36cm) long; the knobs on the hilt can be iron like the rest ofthe sword, or bronze. The longer gladius hispa11ie11sis offigure 110 was probably derived from the antenna types. This figure however has a long CeItic slashing sword. Celtiberian s1eel swords had a high reputation for strength aod flexibility; this may noi be true of ali Spanish blades, reports of some analyses suggesting many wcrc very poor. Diodorus notes Celtiberiaos carried daggers with a biade a span (9 inches123cm) long, used in very dose combat, no doubt because the long sword was unwieldly dose to.

113and114.

LUSITANIA N INFANTRY

Ali Lusitanian warriors carried the caetra; both Suabo and Diodoros describe i1, Diodoros remarking on 1he dexterity with which they whirled the shield round to parry blows, while ali the Lusitanian foot who joined the Romao rebel Sertorius were caetrati. Statues from Portugal, unfortunately headless, show warriors with caetra and dagger. A little confusingly, however, their relatively heavy weapons and armour suggest some Lusitanian caetrati were dose equivalents to 01her peoples' scucarii.

148

Stra bo says they carried kopis and dagger, Diodoros gives them Cehiberian swords, archaeology confirms that both falcata and straight types werc used. Strabo says each man carried severa( javelins and some had bronze-headed spcars; bronze weapons wcre stili used alongside iron in the west ofSpain, and some iron swords had bronze hilts.

Diodoros gives them the barbed all-iron soliferreum. These may be Strabo's javelins, or he may be thinking oflighter weapons of skirmishers - "true" caetrati. Stra bo also says most wore sinew helmets and linen cuirasses. lt has been suggested that the usual linen tunic was stiff and heavy enough to be protectivc, and if so the lincn "cuirass" may be the ordinary tunic. But some runics in art crease and bang like ordinary lightwcight cloth, so I suggest rhe stiff appearance of others is the result of the stiff conventions of some anistic styles. One caetrat11s from Osuna has a layered garment that I ha ve here interpreted as a quilted linen cuirass (though the Osuna warriors are on rhe wrong side ofSpain to be Lusitani). A few Lusitani wore iron mail shins, no doubt adopted from their Celtic neighbours, and uiple-crested helmets. These are differentiated from the sinew rype worn by the majoriry. No known example survives, but as Diodoros says the Lusitani wore Celtiberian helmets I suggest that the triple crest was worn on bronze Celtic helmets, of which examples bave been found in Portugal. Infantry, perhaps only the armoured rninoriry, also wore greaves.

Again according to Strabo, Spanish highland peoples including the Lusitani, Callaeci, Asturi, Cantabri and Va~cones, wore black, mostly black cloaks, the qualification suggesting some tunics may have been black. Strabo calls the cloak sagon, which usually means the Gallic cloak, but Spanish art usually shows a different type, woro undcr the left arm. lt is sometimes folded to produce a double rhickness, and can have a fringed or parterned hem. lncideocally Silius ltalicus twi ce mentions Vascones as not wearing helmets; he uses galea, strictly a leather rather than metal helmet, so may mean che Vascones did not use the common Spanish sinew headgear.

115, 116and 117.

SPANIS H CAVALRV

Ali these are from Iberian sources, and the locai differences mentioned above would apply. Cavalry were more likely to wear helmets and cloaks than infantry. When not the black bighland type, cloaks could be red; a fragment of sculpture shows a red cloak, while Turdetania in southern Spaio exported scarlet and red ochre dyes.

6th-4th cenrury art shows shieldless cavalry likc 115, based on a 4th-century Valencian bronze; thereafter they are usually, though not always, shielded, most carrying the small roundcaetra like 116, from another Valencian bronze.

Both these havefalcata, two javelins or light spears, and a close-firtiog helmei which in l l 6's case I have interpreted as the sinew type. 115 however has a high stilted crest, which seems unlikely on a sinew helmet; his may be a bronze helmet ofthe same shape, or a crude representation ofa Greek helmet, some being imported.

149

A few sources show cavalry with thc ovai scmum like 117. Il 7a is a round cavalry shield larger than tbc caerra, from an lberian coin; the size suggests its bearer is closer in type to the sc11111111-ca rrying than the cae1ra-carrying cavalry.

Such men usually carry a single longer spear used 10 thrust over· or under-arm. Spanish spearheads are distinctively long and slender, up 10 22 in eh es (55cm) long; bunspikes can al so be very long. Some at least ofthese heavier cavalry were probably armoured; Silius ltalicus mcntions cuirassed Scdclanian cavalry, and scale cuirasses are seen in art, though some are worn by dismounted men. Ccltiberian and Lusitanian cavalry would no doubt favour mail, while one Cehiberian site produced metal discs probably sewn 10 a leacher or linen jerkin as armour.

118.

SPANISH STANDARD BEARER

This rider carries a boar image on a pole, like Gallic standards, from an l berian coin. I have taken the opponunity to show an alternative armour style, a disc brcastplate with a snarling wotrs head from a 4th cencury scuJpture. lt is probably bronze on a leather hacking, likc the ltalian defenccs it closely resemblcs. The sbouJder straps, probably also bronze, are embossed witb acoms as good luck symbols.

For musical instruments, Spaniards used various sons of trumpet, including the curved typc of ligure 141. Some were ofbaked day.

119.

BALEARIC SLINGER

The Balearic islands produced famous slingcrs, who were sa id to train from infancy, a chi Id not bcing allowed lo eat till he had hit his food with a slings hot. According 10 Strabo and 01hcrs, the islandcrs seem to have becn primitive and very poor when the Canhaginians lirst comacted them, goiog barefoot and dressed in skins. Those who did not light with lhe sling used lire·hardened javelios, as they had few metal spearheads, with goatskins draped over 1heir arms as improvised shields. The Carthaginians taugh1 them 10 dress in broad·bordercd 1unics, probably thc Spanish style. They may have issued 1unics as part oftheir mercenaries' pay; Diodoros says that the islanders did not use money but wcrc paid in kind, including wine and women. A poem ofabout 300 calls their tunics "wretchcd and diny''. Bronze ligurines from the islands sometimes show sandals, but most are barefoot.

The slings wcre made ofhair or sinew, and Strabo says each man had three, with thongs ofdifferent lengths for different ranges. Looger thongs would allow the missile 10 buìld up grcatcr velocity as it was whirled in a wider are, but accuracy would suffcr. Those slings not in use were tied round thc hcad. Diodoros says Balearic slingers sho1 sto nes weighing a 111i11a (15.3 ounces/436g). This seems exagge ra1ed, bui Jarge lis1-sized slingstones have been found in mainland Spain, where slingers wcrc also used, as havc lcad and even iron bullets. \Ve are no1 told how the ammunition was carried; slingers on Trajan's column, with stones carried in folds of their cloaks, have been identilied as Balearic, but this is far from ccrtain, and given the islanders' apparcnt poveny one wonders whether they all had cloaks to use.

One figurine has a small round shield; he is nota stinger and we ha ve no definite evidence thai slingers used shields, but it may be signilicant that his shicld is ca rried on armstraps, nota Spanish centrai handgrip. Thus it could hang on the foreann whilc the left hand was free to load the sliog. The ligurines ha ve helmets, mostly ofGreek type, while severa! Montefortino helmets have been found. lt seems un likely that these belonged co slingcrs; apart from considcrations ofpoveny, a helmet would interfere with thc stowage ofthe sling! Probably most ofthcse helmets belonged to Roman and Canhaginian garrison troops, though no doubt Balearic chieftains used chem.

Bull symbolism, probably religiou~, is common in the islands' art. 1 I 9a shows a bronze bull's head restored as a standard, which it may have been.

120, 121 a nd122.

CELTIC INFANTRY

T he Celts were markedly taller 1han the Mediterranean peoples, muscular, fair-skinned and mostly fair·haired.

Most men wore long flowing moustaches and a few added short beards. Hair was smeared with a lhick wash oflime, which bleached it and_made it stiff and spiky, and then combed back frorn the forehead to stand out like some wild animal's mane. I 20 lighcs naked, except for the gold or bronze torc and armlets which mosc warriors wore. Nude lighting was common and perhaps uni versai among early Cclts, but died out during our era; at Telamon io 225 only the mercenary gaisatai stili fought nude. A small statuette found near Rome, ofa nude warrior with torc and horned helmet throwing a spear, may represent one ofthesegaisata1~ from his helmet pcrhaps a chic( The Galatians in lhe cast apparently kept up the practice longer, as they stili fought nude against the Roman invasion of 189. The mocive seems to have been an invocation for divine aid. 121 typilies most ofthc western Celts of our period; 1he Boii and

150

118

lnsubres at T elamon, and the Gauls with H annibal, fought s1ripped 10 the waist bui wearing trousers and cloak.

122, in breeches and tunic, is from the Gundestrup bowl, found in Denmark bui made somewhere in northern or eastem Gaul somewhen in 1he 3rd-lst cenruries BC. C lothing was usually wool, and cspccially for the rich was often woven or embroidered in striped or checked patterns, and brightly coloured. Trouscrs varied in style; an suggests the tribes living in ltaly usually wore long tight trousers with the long leather laces from the shoes tied around them at the ankles, but looser trouser styles were also worn, while the Gundestrup cauldron shows breechcs reaching just below the knee, clearly tbc ancestors ofthe style wom by Imperia( Roman troops. The heavy woollcn cloak or sag11111, usually dyed, was fastened with a brooch on the right shoulder, and may sometimes have been hooded. Tunics were fairly short and had long slecves.

Most warriors were armed with sword, shield, and one or two spea rs. The sword was the primary weapon. 120 carries the fairly short pointed weapon of the La Tene I culture of thc 4th ccntury, the biade usually 22-26" (55-65cm) long. Short daggers were commoner at this carly pcriod than later. 121 carries the long, blunt-cndcd slashing sword ofthe 3rd cenrury and later, so often described by classica( writcrs, brought down from abovc 1hc head in wild blows "throwing thc whole weight ofthcir bodies into tbc blow like hewers ofwood". Bladcs were 29¡ 35" /75-90cm long. Polybios claims Celtic swords were useless after the first blow, as they bent and had 10 be 151

s1raighrened wiLh che fooc; bui 10 01her writers Celtic swords, like Spanish, were a byword for strength and flexibiliey, and some surviving examples are s1ill superbly flexible. Polybios may be confused with the ritual bending or "killing" of weapons for burial wi1h their dead owners, but a simpler explanation may be that the quality ofswords varied considerably, wi1h the different ores orsmiths' skills ofa specific locality. Swords were worn on the righ e from a leather or chain wais1belt, in scabbards which were oftea ealirely of iron. Despite the curious statements ofsome early 20th ceniury wri1ers, swords were 1101 used in two hands.

Mos1 warriors had spears, bui as che sword was clcarly 1hc primary weapoa most spears were probably lhrown.

They are usually illusera1ed abou11he hcigh1 ofthc warriors wielding Lhem, but could be up to 8 feet/2.5m long.

Spearheads varied greacly in sizc :md shape, bui were often long and broad, 1he sides curving inwards becween the broades1 point and 1he 1ip. Sockc1s are shown reinforced with wire binding.

The characteristic Celtic shicld was ovai, of lca1hcr-covcred wood, very like che Icalian scurum from which ii is presumably derived, bui flat instead of convex. Shape varies around 1he standa rd ovai, some having squared-off ends and 01hers nearly s1raigh1 sidcs. Shields were quite large, one exam ple from La Tene being 4 I" long by 21" wide (l.05m by 54cm) while one from Denmark is shoner but wider, 341/2" by 221/i'' (88 by 57cm). Nonetheless both Greeks and Romans thought they were inadequaee procection for an unarmoured man. Surviving shields are of oak planks covered with hidc; Caesar describes Gauls in an emergency improvising shields ofbark or wickerwork covered with leather, and poorer men or skirmishers may bave used such cheaper, ligheer shields even in normai circumstances. The shield rarely had a metal binding, and the thick raised rim often depicted is probably the leather covering doubled over 10 reinforce the edge. The shield was held by a single handgrip across the short axis, usually covered on the outside by a long wooden spindle-shaped boss hollowed out to cover che hand and reinforced by an iron or bronze boss-plate. Some circular bosses have been found and may belong 10 shields wiLhout 1he wooden rib, like Lhat of 122. Pergamene statues of defeatcd Gala1ians and other Hellenistic an some1imes shows Celtic shields held by armstraps in Greek s1yle. This has been dismissed as anistic convention bue there is no good reason why some Galatians should not have chosen 10 adop1 the Hellenistic grip.

123.

CELTIC SKIRMISHER

Lighc infoncry were noi common in Cel1ic armies; some are memioned in 1he Gallic army ai Telamon, thc Bascarnai had some, but there are no rcferences to Galatian skirmishers. Most werc probably javelinmen, with dagger rather than sword and a lighc shield. Archers and slingers however were used in small numbers by the lst century BC so were probably around carlicr.

124 and 125.

CELTIC NOBLES

These mai led noblcs are 1ypical of kings and chieftains, most chariot-riders and an unknown but perhaps large proportion of1hc cavalry. I 24's helmet and mail shirt are taken from a 3rd cenrury grave ac Ciumesti in Romania.

125 is based on evidence from north lialy and sourhern Gaul.

Mail, armour ofinierlocking iron rings, was invented by the Celts perhaps in che 41h ceniury. They used two main s1yles ofmail shirt. 124's has an overhanging cape-like shoulder defence. 125's shirt has shoulder pieces cut like chose ofGreek cuirasses, which do not overhang the shoulders. They are fastened by a bar which is fĂŹrmly fixed to che chest ofche shin and hooked to the shoulder sections. M ail was 1he only armour used on any scale by the Celts in our era. Thcre is a reference to bronze-armoured Galatian infaniry ai the "elcphani" baule in 273; ifthis is corrccc, chey must have been wearing caprured Hellenistic armour.

Bronze or iron helmets were commoner ehan mail, worn by most cavalrymen and perhaps even the occasionai ordinary infantry warrior. The commones1 sryle is Lhe Moniefonino, a bronzc hclmet named after the cernetery of the Gallic Scnones in north Italy where many were found. It has a low bell-shaped or slightly bulbous skull, rising to a knob usually drilled out to take a crest; a thick lower rim; and a small neckguard projec1ing a11he back. The cwo sryles of cheekpiece used wilh it, the triple-disc sryle of 125 and the scalloped pauern of l 24a, are both Icalian. 126 shows 1he basic helme1, which was used ali over lhe Celtic world from Spain to the Crimea and became che commonest Roman helmet type. 125 is a 4th-cenrury example from Moniefortino itself, bronze with an iroa cres1holder, che outer arms of which may have held ltalian-style uprigh1 feathers. I 24a, from a grave near Parma, has shee1 bronze horns anached; Lhese may bave been limited to ltaly - see fĂŹgures 129 and l 56b. I 24b is 4th century, from Amfreville in France, and is bronze covered with iron shee1s worked in reUef, with Lhe centrai bands gilded, and with a lyre-shaped ornamene on che sides. 124 wears an iron variant common (without lhe bird crest) among Celts in the Balkans and illustrated among che Galatian trophies on the Pergamon reliefs. le has a simplified versi on ofthe Italian three-disc cheekpiece and matching orna meni on the sides ofthc skull. The Ciumesti example shown has a large bronze bird with hinged wings as a crest. This is the only surviving example ofthe animal-crested

152

153

helmets mentioned by Diodoros, but compare l 24h; anothcr helmet from that source has a bird crest. I 24c is a late 5th century tali spired bronze helmet from the Marne arca, perhaps the prototype ofthe Montefortino style. Some spired helmets, though not as tali, remained in use later; I24d is iron, 3rd century, from the Alps. I24e is the simple round-capped Coolus type, used from the 3rd century to thc I st and aga in the ancestor of a school of Rornan hclmets. 124 f¡h are from the Gundestrup cauldron; the helmets on this seem closest to the Coolus style. l 24i, from eastern Europe, shows heavy Hellenistic influcncc. I 24j, very sirnilar except for thc horns-and¡whcel crest, is from the Arch ofOrange ofthe early lst century AD. 125 carrics a javelin with a long sockctcd iron head, from the Montefortino graves. lt is clearly derived from the

pi/11111 or similar ltalian socketed javelins.

126.

CELTIC C AVALRYMA N

Many Celtic cavalry used round shiclds, with thc spine carried horizontally (like figure 46; see J 30c), though ovai shields were uscd too. This man carries a thrusting spcar with typically Celtic head. As mentioned in the last note,

126

a

154

b

mosc would wear helmecs, as do all che cavalry on che Gundescrup cauldron (buc only one ofthe infantry) and many mighc wear mail, as cavalry were rccruited from che nobles and cheir retainers. Celcic horsemen used short prick spurs.

127 and 128.

CELTIC MUSICIANS

127 is a horn-player, caken from a carved stone from northern Italy. 127a is a larger bue similar horn wich the addi cion ofa carrying-strap, from the famous Pergamene stacue of1he dying Galatian. 128 is from che Gundescrup cauldron. He plays che mosc discinctive Celcic inscrumenc, che long camyx. This has a bronze head shaped like an open-mouched animai.

129.

CELTIC STANDARD-BEARER

From che same source as 127. He wears a horned versi on of the Negau helmet (see figure 73) with a Jow comb on che skull. The standard has an elaborate pierced spearhead on che butt; severa! ofchese have been found. The standard itselfincorporaces sheet bronze horns like those used on helmets and a small figure of a fis!h. l 29a, from che Arch of Orange, is a bronze ligure of a boar. Ocher animai scandards included cockerels and the bronze bull on which the Ci mbri (fairly cenainly Celts, not Germans) swore oachs, and which chey carried into banlc. l 29b, also from Orange, is a cloth vexilfum hung from a crossbar.

130.

CELTIC SHIELD PATTERN$

Diodoros says che mocifs which dccoraced Celcic shiclds were ofbronze. Some ceremonial shields may bave had bronzc cmblcms, Like the elaborate bronze-faced Bricish ceremonial shields, but baule shields had painted decoration. l 30a and b are carried by warriors shown on a bronze scabbard of about 400 BC. l 30c is among che Pergamene crophies. So is l 30d, clearly used by a Galacian from ics associacion wich other Galacian equipment, bue nota Celcic cype. lt is che round, rimless, convex Macedonian infancry shield, presurnably either decorated with a leaping wolf(? the originai is rather crude) by a Celt who has capcured ic, or made in imitation of Macedonian styles.

The Gala1ians probably used much H elleniscic equ ipmenc from che start; whcn chey first crossed co Asia half of chem were unarmed, and chcy were equipped by their ally Nikomedes of Bitbynia. l 30e is carried by a Gallic horseman in Hannibal's army depicced on a Roman coin. l 30fis from a warrior starue from Mondragon in southern France, Isc ce ntury BC. l 30g-i are from Orange.

155

131.

LIGURlAN WARRIOR

The Ligures were a people speaking an Indo-European language related both to Celtic and Italic, who had absorbed the earlier inhabitants ofnorth·west Italy and south-east France. In our period they were restricted to che hills of modem Liguria, Piedmont and tbc northern Appennines, having bcen displaced by che Gauls from the richer plains Lhey had once occupied. They were grouped in scvcral loosc tribal federations. Most were very poor, and so supplemented their hunting, herding and forest ry by raiding and piracy. A few port towns, notably Genoa, were more advanced and prosperous. Ligurian mercenaries fought for Carthage and Syracuse; in che 2nd century they quite often co·operated with tbeir Gallic neighbours against Rome, and in 171-168 some fought for Rome against Macedon. After our era tbey were commoner in Roman service, in Marius' day. The Ligures were a short, wiry people toughened by their harsh mountain life and considcred " tough fighters'', "warlike and proud", with more endurance than their rather fickle Celtic neighbours. Strabo and Diodoros say they fought almost exclusively on foot, because ofthe nature oftheir country; but their phrasing implies that cavalry were not entirely unknown, and two recently discovered Ligurian graves bave included harness fittings. Strabo indeed assigns thc Salycs, north of Massilia, a substantial cavalry force, but they were, as he points out, one of severa! mixed "Celto·Ligurian" tribes, and the cavalry probably reflect lhe Celtic element. Ligurian infantry were good both at dose combat and as skirmishers. They preferred raids and skirmishes, somctimes losing contro! and scattering in scarch of plunder, but could fight doggedly hand·to·hand when necessary. Before Pydna, the Romans uscd their Ligurian auxiliaries brigaded with the velires for chasing ofT Macedonian skirmishers.

Most wore their dark hair fairly short and had short beards. Clothing was mostly wool, Strabo mentioning woollen runics and cloaks ofcharacteristic Ligurian style as a locaI export. Tbe tunic seems to have been round-necked, longsleeved and slit at tbc sides. Legs were bare; feet could be too, or bave Gallic-style leather shoes. Furs were used, and alsoexported. In rerum richer foreign textiles were import ed and would be worn by chiefs. A broad leather belt with bronze clasp was usual. Sidearm would usuall y be a short, straight sword or long dagger, some using longer Celtic swords instead, and these usuaUy hung from the belt, occasionally a baldric. The typical shield, the sc111um ligus1ic11111, seems to have been the same as the Gallic flat ovai or square-ended scutum. The main offensive weapon was a bundle of4 feet (l .2m) javelins, the iron heads often three-cdged. Axes were obviously common because ofthe importance offorestry, and migbt bave been uscd in baule. Some use was made ofbows and slings.

Chiefs would often bave Celtic torcs and might wear Celtic or Etruscan helmets. Silius Italicus mentions Ligures with bristling conica! helmets, co11is horridis, perhaps indicating Montefortino or Etruscan Ncgau helmets with crests. 13 la shows a rarity, a bronze helmet ofremarkably old-fashioned 8th-cenrury Villanovan style. Severa! such helmets ha ve appa rently been found in Ligurian graves of our era. Whether they were handed down through the centuries or made more recently to an extremely ancient design is unclear. However, their use may throw some light on an obscure remark ofStrabo's. He says thnt some carlier authors had considcred the Ligurians to be ofGreek ancestry, because they used bronze shields. While not accepting the obviously baseless idea of Greek descent, Strabo does apparently ueat the bronze shiclds as fact. Two explanations suggest themselves. One is that some Ligures had adopted the round bronze-faced Argive shield, though probably not the full hoplite panoply, from the Greeks of Massilia or the Etruscans ofthe Po valley. But an alternative suggested by the Villanovan helmet is tbat some warriors had round bronze shields of Villanovan type, held by a centrai handgrip and decorated wilh concentric ci rei es ofraised bosses. These would be purely ceremonial, tbc Villanovan examples being merely a thin bronze facing with no wood or leather backup, too fragile for use in battle. This would explain why we bear only of sciaa in battle. Perhaps the ancient helmets were also ceremonia1?

132 a nd 133.

ROMAN HASTATI OR PRINCIPES WITH PILA

The pilum was a heavy javelin which became standard Roman infantry armament. lts origin has been variously explained, some ancient writers bclieving it was Samnite. Archaeologica1 evidence suggests it was developed from earlier Italian javelins, perhaps by the Etruscans (figure 163). Tbe Romans were using i1 by the early 4th century, pila being prominenc in accounts ofCamillus' victories over the Gauls after the sack ofRome. Tbc hascati were the first 10 use pila. This at first seems curious, as liasta usually means a thrusting spear, so it has been suggested this was their originai weapon. But the explanation seems to be that liasta could originally include thrown spears; our earliest surviving evidence, in Ennius' A1111als, speaks of"hastati wbo hurl hastae"; their name tbus reflects the period wben tbey alone used pila. In the mid·2nd ce ntury Polybios decribes the pri11cipes too armed with pila, but they staned our period with thrusting spears, as Dionysios ofHalikarnassos, in the context ofthe Pyrrhic Wars, says they used "cavalry spears gripped in the middle with both hands". The two-handed grip, which would nor be practical with a Roman shield, may be a corruption, but tbc rest ofthe passage probably goes back to a reliable con temporary Greek historian; the "cavalry spears" be is tbinking ofwould be the long Macedonian xysca. In 223, however, a Roman generai fighting thc Gauls distributed 1he spears ofthe tria rii to the foremost maniples, and it

156

a

b

132

would seem that by this time only the rriarii stili had tbrusting spears. Roman coinage berween the Pyrrhic and Punic wars sbows shield and sword, emphasising them as if the sword was now the main Roman weapon. T his would be che case if che majori1y of che infantry bad now 1aken up che pilum, because a legionary wi1h pila is essentially a swordsrnan, while the sword is a secondary weapon to a spearman. This sugges1s the pila and swords of the has1a1i were rnarkedly more efTective against Pyrrhos' phalanx 1han 1he spears of the pri11cipes, whose equipment was reformed as a result.

Polybios disĂšnguisbes two 1ypes of pilum, "thick" and " 1hin", saying each man had both rypes. Surviving examples from Telarnon (pcrhaps dedicated to commemorate the victory of 225) and from Numantia in Spain confirm two basic rypes of construct ion. Borh bave a small barbed head at the end of a long iron shank, fitted 10 a wooden shaft. One rype has the shank socketed. The other has a wide flat iron tang riveted to a thickened section of 1he wooden shaft. Thc riveted type must be Polybios' " thick" pilum, referring to the broad joint ofiron and wood.

This broad section can be either square or round in section, and is strengthened by a small iron ferrule. The iron shank varies considerably in length. Many examples are about 30 inches (70cm); the longest example at Numantia, one ofthe socketed type, is 37112 inches (95cm), even with its head missing, while the tanged Telamon pila are only about 10 inches (26cm). Polybios says the wooden haft was three cubits (54 inches/l.37m) long, the iron thc samc length, which would seem to be an exaggeration. Probably the slightly lighcer "thin" sockeced pilum was thrown first, then the other, and the legionary followed up with the sword. The pilum could pierce armour or shicld, sometimes pinning overlapping shields togethcr. A pilum-pierced shield was vinually useless, as the weapon's own weight would bend the thin iron shank, leaving it trailing along the ground 10 weigh tbe shield down. The barbed head made it difficult to remove, while the bent shank meant it could not be 1hrown back by the enemy.

Pila thrown, tbe legionary was a swordsman. By the end of our era he used the Spanish cu t-and-thrust sword, the gladius hispa11iensis (figure 11 O). A later lexicographer, possibly following Polybios' lost account of the Celtiberian war, says this was adopted from 1he Celtiberians after the war with Hannibal. It was certainly in use by 200, when Livy describes the Macedonians' shock at thc 1errible wounds it inflicted. The army in Greece in 200 may tbus ha ve been che first to bave this sword as standard issue, chough individuals in Spaio probably adopted and popularised il earlier. The biade (I 32a) is about 20 inches (52cm) long, slighcly waisted with a long point. Its Celtiberian ori gin was reflecced in ics good sceel, perhaps the mosc important difference between this and ics predecessor in Roman hands. This predecessor is never described, but must have been a short thrusting sword as Romans are described ducking under wild Celtic slashes and stabbing up at the stomacb. This early sword must be the Greek sryle widespread in ltaly, sometimes longer than the Celtiberian, with a biade up to 271/z inches (70cm) long, "leaf-shaped" so it was broadest near the point. This broadening gave maximum weigh1 at the point ofimpact ofa cut, but the long 157

point of1heg/adi11s '1ispa11iensis must have been better fo r chrusting, and handie r since the cenere ofgravity would have been further back. Swords werc worn high on the right from a broad waistbelc. Spanish daggers like small versions ofthe swords were also adopted.

The main defensive item was thc shicld, a large ve rsion ofthc ovai sc11111111 co mmon in Italy. Some Roman infancry had always used scura ofsome type, and about 400 this style became universal. lt was 4 Roman feci long by 2 112 wide ( l .2m by 77cm), convex, made oftwo thicknesses of wood covered with caovas aod calfskin. As with Celtic shields it had a long centrai wooden spine, metal boss-plate and single handgrip. The top and bottom rims (not, Polybios implics, the whole rim) were reinforced with iron binding. No shicld blazons seem to have been used at this time.

The only comemporary painting, in the tomb ofthe Fabii on the Esquiline hill, shows a plain light brown shield, whilc carly sculpture like thc Acmilius Pau llus monumenc shows plain sciaa, in contrast both with the 1\ \acedonian shiclds on the same monumcnt and with later Roman sculpturc, which has shield-blazons in relief. Silius Italicus describes Picentes (armed in Roman style as they use pila) with rcd shields.

132 wears the regulation armour Polybios describes, by his day issued by the Sme to soldicrs 100 poor to buy their own, the cost being reclaimed from pay. His torso is protected by a small bronzc breastplate, pectorale, a span (9" /23cm) square. It rescmbles the Oscan style of 148, bui is a little smaller, and Polybios makes no memion of a backplate. lt is clearly desccndcd from bronze age defences, which were often decoratively embossed, but as chis is now chcap government issue it will not ha\'e been very ornate. Severa! round breastplates from Numantia may be Roman, but if Polybios is right to say only squa re ones were used at Rome, the discs may belong to ltalian allies (see ligure 153).

One greave is worn, on the left leg. Polybios clearly refcrs to one only, and Arrian conlìrms him, saying the ancient Romans used to wea r one grcavc only, on the leading lcg, 1hc left. As well as protcction, it would prevent the shield· rim chalìng the leg. Silius ltalicus also mencions Sabines with one grea ve. No doubt many of those who could a(ford li would actually have a pair. Italian greaves were ofcen fastened by straps, not their own elastic ity.

Polybios says bronze helmets were wom, without describing them (there is probably no truth io lhe story Camillus issued iron helmets in the early 4th cenrury), bue the Attic, Moncefonino and Etrusco-Corinthian scyles worn by 132, 133 and 134 respectively were all popular in ltaly at this time and were probably ali used, as they cenainly ali were by later Roman troops. An unclear helmet on thc Esquiline painting may be Montefortino. This type was probably commonest; m:my 3rd -2nd ce ntury exa mplcs come from ltaly and Spain. They are slighcly plainer :md cruder than thc early Gallic cxamples, no doubt mass-produced for the larger Rom:m armies ofthe era. Polybios says helmets were crested with a crown or ci rclet of feathcrs and chree uprigh t black or crimson feathers a cubic ( 18" /46cm) cali. Paul Coussin incerpreced chis as indicating a horsehair cresi plus fcathcrs, following the practice of other ltalian nalions (see 144a), and other writcrs ha ve followed him, but this is 1101 wh:n Polybios qui te clearly says, and the Romans are under no obligation to conform to cheir neighbours' practice. Polybios implies this stylc of crcst was uniform.

133 is a richer citizen; the only relic of earlier org:misation, whereby troops wcre equipped according 10 their wealth, was that the richesc class had to provide :m iron mailshirt. He is perhaps more likely 10 be a pri11ceps than a lras1aws, as the former class, being older men, will on average have been better ofT. Scale cuirasses, though not mentioned in this era (except once for an officer by Silius Italicus) may havc bccn worn.

Cercainly scale cuirasses cut like the mail shirts are shown on monumcnts from the lst ce ntury BC on, and fragmencs ofbronze scale survive which may date to our era. Onc cui rass, with a broad ltalian belt, is said 10 havc been found in Lake Trasimenc, lost perhaps by a victim of Hannibal.

Clothing, a woollen tunic, was also state issue by the 2nd cemury, and e\•en earlier towns in the theatre of war had been obliged to providc clothing 10 Rom an armics. Such issue runics may alrcady have been dyed a uniform red, as this seems to havc bcen a popular colou r with othcr Italian warriors, as well as with Hellenistic armies.

134.

ROMAN PRINCEPS OR TRlARIUS WITH SPEAR

The long spcar, llasta, survives from the era whcn thc Roman army was a hoplitc forcc. lt was stili carried by the pri11cipes at the start of our period and by thc triarii at lcasc unti! Polybios wrotc. They probably kept it unti! the disappearancc ofmanipular 1actics oblicerated the di!Tercnce betwcen 1he types oflcgionary, at the end ofthe 2nd century; but some Roman writcrs imagined a period whcn llas1a11: pri11cipes :md triari i ali had pila. This may be a groundless anachronism, buc could perhaps indicate thai tria rii adopted pila fora shon while before they lost their identity.

158

This figure's position follows Livy's description of the criarii waiting in reserve, kneeling "with the left leg advanced, their sc111a leaning on thcir shoulde rs, their spears fi.xed in the ground and pointing obliqucly upward, as iftheir line were protected by a bristling palisade". As the oldest men of the army, the eria rii would ha ve the highest proportion able to alford mail.

135.

ROMAN ACCENSUS, 340 BC

As noted in the account of Roman organisation, the nature and role of the accemi are problematica!. But if the interpretation put forward therc, that thcy wcrc an emergency levy ofthe poorer classes, is correct, this figure shows likely equipment. Their rote at the baule ofSuessa, successfully impersonating the criarii, implies that they carried the same spear and scuwm, and had the usual helmet crest (iftbe distinctive Roman crest was in fact in use by 340).

However tbey would not be able to alford armour. lt would be surprising to find the state issuing armour at this early date, and it not certain either that Rome could alford to find large quantities of extra armour, or that there would have been time. The absence of armour would not be clear behind the sc11111111, so would not tip the enemy olf that these were not real triarii. As a parallel, the proletarii called up to defend the city walls during thc war with Pyrrhos were issued with sword and shield, but no mention is made ofarmour. This figure thereforc wears only a cheap and simple version ofthe Attic helmer.

136.

ROMAN LIGHT INFANTRVMAN, 4th CENTURV BC

Livy describes the /eves ofthe 4th century army as ca rrying spear and javelins, with no reference to sword or shield.

This is the same equipment he allots to light infantry ofthe earlicr hoplite¡based army, and would probably also be the armament ofthe rorarii, ifthesc werc indccd light troops. This style may not havc becn rcplaced by that ofthc next figure unti! 21 1 BC.

137.

ROMAN VELES

Polybios describes the equipment ofthc velites (ofwhich ve/es is the singular) in the 2nd century. His only armour is a helmet with no cresi, sometimes covcred by a wolfskin or similar, both for additional protection and so that an individuai ve/es could be recognised ifhe distinguished himself. He carries a strong round shield (par111a) 3ft/90cm 159

in diamecer. \Veaponry is a Spanish sword and severa! javelins with 3ft/90cm shafcs and long thin iron beads, of which examples have been found at Numantia. The large shield and effcctive sword gave the ve/es an advamage in close combat over most other light troops, and chey were generally prepared to come to close quarters with such troops.

The earliest mention ofshielded Roman light infantry is in 211 when Livy mentions light infantry, each with seven javelins and a round shield, co-opcrating with Roman cavalry. He claims that as a result of their success light infantry were incorporated for the lirst time in the legions. As noted in the section ofRoman organisaĂšon, this cannot be wholly correct as the legions had long contained lighc infantry, but may reflecc a genuine rc-organisation ofthe corps connecced with an increase in their numbers. lt seems likely cbat shields are mentioned for the lirst time because they and swords were introduced at the time ofthis reform. Tbis would explain wby swords and shields, stressed in severa I later accounts ofRoman light infantry's exploits (for instance bcforc Pydna and in Galatia in 189) are noi mentioned before 211, and why Roman skirmishe rs pu1 up a poor show in the early phases ofHannibal's war while discinguishing themselves in lacer conflicts. Of course if swords were introduced in 211 they would presumably be ltalian rather than Spanish weapons a1 lirst.

138.

ROMAN PENAL LEGIONARY, 216 BC

In 1he desperate days after Cannae Romc tried 10 replace some ofthe losses suffered 1here by arming 6,000 debtors and criminals, pardoning their offences in re1um for military scrvice. They were armed with Gallic weapons, captured at Telamon and dedicated as trophies. There would have been few cuirasses available, but probably enough helmets, swords, spears and shields 10 cquip all 6,000, though che assoned Gallic spears and javelins would not be exact subsituces for pila or the long spears of 1he triarz'i.

139.

4th CENT URY ROMAN C AV ALRYMAN

Polybios says early Roman cavalry were lightly equipped "the rcsult of whicb was 1hat chey were able to dismoum and mounc aga in with grcat dexterity and facility, but were exposed to greac danger in dose com ba e". They wore no armour, though helmets are likely. Protection was a small oxhide shield (panna eques1ris) which Livy describes in use as early as the late 5th centu ry. lt was too light for adequate pro1ec1ion at close quarters, and tended 10 rot in 1he rain. Polybios compares its shapc to a type ofround bossed cake; it may be the type shown in 44a, used in T aras, with a flat rim and convex cenere. M ain weapon was a slender thrusting spear, also unsatisfactorily ligbt. lt tended to shake in the charge, making aiming difficu lt, broke easily, and ifbroken was useless as it had no buttspike. A shon sword would no doubt be carried on the right.

137

139

160

140.

2nd CENTURY ROMAN CAVALRYMAN

Because ofcbese deliciencies, Roman cavalry adopced the Greek-influenced equipment ofchis ligure, wbo is from an Impe ria! copy ofthe 2nd century BC relief of Menius Cunius, ac Rome. Unfonunately Polybios does not give che dace ofthis reforrn, but as he says the equipment worn "anciently" was "soon" replaced, tbc change probably occurred during or soon afcer Rome's lirst clash with Greek cavalry, in the Pyrrhic war. The lacer wars wich Macedon and che Seleucids seem coo dose co Polybios' own day, as he implies che reform was well in che past. One ahemacive chac has been suggesced is chac Scipio Africanus introduced the new equipment during tbe second Punic war. While Roman cavalry may have been more effective at the end ofthac war than che beginning, Polybios' dose ties with cbe Scipios surely mean chat he would bave claimed for them anycredit going for this reform, while he does specilically mention Greek, not Carthaginian, influence. On balance che Pyrrhic date seems most likcly.

Thc rclief shows a sbon Hellenistic plate cavalry cuirass and Anic helmet. Alternatively, the cavalry on the Aemilius Paullus monument bave mail. The shield is now larger, but though Polybios says che Romans adopted "lirm and scout" shields from the Greeks, Mettius CurĂšus' shield does not match any known Greek type, but seems to be a variant ofthe oxhide parma equescris, slightly larger tban the last ligure's and with a metal boss recessed in the centre. This stylc isstill shown on a coin ofAugustus. The ribbed scyle of 46 and 130e was also used; 140a is the shield of 130e's opponent, a cavalryman shown on a Roman coin. lt has a curious 'M'-shaped blazon ofuncertain signilicance; when I lirst saw it I thought it was an eagle, but if so it is remarkably crude even given the limitations ofthe medium. Silius ltalicus mentions a consul whose shield was blazoned with Romulus and 1he wolf, but this may be mere poetic fancy.

The thrusting spear is now "steady and strong" and carrics a buttspike. lt is clearly modelled on the Macedonian xys1011, but tbe relief shows the Roman spear was shorter, about 6 feet (I .8m). A Spanish sword would be carried,

and the cavalry on thc Aemilius Paullus monument cercainly had swords ac the rigbt oftheir waistbelts.

Cavalry very like chis ligure are shown lighting naked Gauls on a relieffrom Lecce in Apulia, 3rd or 2nd century BC. They are probably locai allies, indicating thac this ligure's style of equipment replaced chat of 151-153 and 158-160 in t he south.

141.

ROMAN MUSICIAN

This ligure is from the later phase ofthe sculptures at Osuna in Spain, actually lst century BC, but an Etruscan relief ofthe 2nd century shows very similar musicians. Though from Spain he is identilied by bis ltalian greaves as both Roman and military. Tbe originai carries no sword. Note he does not wear t hc animal-skin headgear of Imperia! musicians and standard bearers; tbough the resemblance of these co the velices' wolfskins suggests they migbt have an early origin, they d"I not yet seem to be in use. The curved trumpel is a bronze cornu.

161

142

a

b

e

a

e

162

142.

ROMAN STANDARDS

No illustrations of standards or standard¡bearers of this era seem to survive. As noted above, standard bearers probably did not yet wear animal-skin headdresses, or indeed any distinctive dress.

142a isa vexillum, a fringed flag, probably red, hanging from acrossbar. A large red vexillum would be hoisted as the sign to join bartle. They were also used for signalling. Cavalry had vexilla , probably one per wr111a. In the 4th century legion each unir of tria rii, rorarii and accensi was called a vexil/um and had such a flag. At this stage hastati and pri11cipes had no standards, their ma111'p11li regrouping only on the standards of the rear line; they were thus referred to as amesig11a11i, "before the standards". T his may be why 142b carne into use; this is the standard traditionally associated with the 111a11ip11/11s, a handfuJ of straw tied to a spear. hs rudimenta.ry nature suggests it may ha ve been improvised to fili the mam'p11/11s' need fora standard ofirs own as a rallying poiot. By Polybios' time each ma11ipulus had two standard-bearers on the strength, though he does nor describe the srandards. Early Imperia! maniple standards secms to have been topped by a hand- 111a11us.

Pliny says that by the late 2nd century BC the legion had five srandards with animai images - eagle, boar, wolf, horse and minoraur. Marius, abour 107 BC, confirmed the eagle as the pre-eminent standard, but he was only formalising an existing arrangement as the other four had for an unspecified time previously (thus possibly before the end of our era) been usually left in camp. This strongly suggests they were addi tional 10 the manipular sta ndards, as these had a definite use to mark their unit's positions, and would not have been left behind. As forerunners ofthe later eagles, these standards we re probably small silver statuenes like 142c. ltalian allied troops had ar least one vexi/111111 per cohors, and may have had standards for smaller sub-units 100.

Standards were already the object ofa quasi-religious veneration; Plutarch says the ltalians thought it "criminal and impious" to lose a standard. There are instances of officers, both Roman and allied, using this ro inspire a desperate charge, by hurling the unit's standard into the midst ofthe enemy and challenging the mento get it back. The Paeligni seem ro have been especially fond of this; a Paeligni officer at Pydna for instance threw his cohort's standard imo the Macedonian peltasts, but bis men 's rash charge after it got them impaled on the Macedonian sarissai and roured.

143.

ROMAN OFFICER

This figure is based on what may be tbe only represemation of a Roman officer of our period, in tbe paintings oftbe tomb ofthe Fabii. He carries the usual infantry scumm, and wears a feathered, probably Montefortino, helmet; a cuirass apparently ofplate, worn without pternges as was common in ltaly; two greaves; :and a cloak. Art ofthe l st century BC shows officers in Hellenistic panoply, muscled cui rass with fringed pteruges and shoulder straps, crested helme1, and waisc sash. This ca rne 10 be ge ne rally associated wi1h high rank, and by the end of our era many officers migh1 be similar 10 35, though ca rrying che Spanish sword. Since officers would provide cbeir own equipment, chere might be considerable va riation. Paullus fought withouc helmet or cuirass at Pydna; another genera i fought wi1h a hunting spear.

Generals wore red cloaks, tbough Silius ltalicus memions a consul in purple, which was cenainly worn for ceremonial purposes. A consul would ha ve a retinue of 12 lictores ca rrying thefasces, the rods and axes emblematic ofhis authority. Tribunes and above may have worn white tunics edged purple, matching the rogas which distinguished the senatoria! ariscocracy from which 1hey were drawn. Cenrurions may already bave wom helmec cresrs running transversely, from ear to ear. These were worn under tbe Empire; tbere is no direct evidence for our era, che earliest apparently a relief of about 70 BC. However the Etruscans used such crests (see figures l 62a and b), so chey may have been adopced by Roman centurions at an early date.

144, 145 and 146.

OSCAN JA VELINMEN

Oscan warriors are profusely illustrated, parcicularly in 4th centu ry tomb paintings from Capua in Campania and Paestum in Lucania, and on Campanian and Lucanian vases. There are no representations of Samnites excepc a statuene from Sicily, which is probably Samnite work, and some lost paintings from the Samnite town Allifae, su rviving descriptions ofwhich suggest they were similar to che Capua and Paestum art. Strabo confi rms chat che va rious Oscan peoples had once had distinctive sryles of dress and armamene, rhough che se bad disappeared by bis day. What evidence we ha ve suggests the difTerences between them were not great, although the coastal peoples had adopted more Greek equ ipment, and therefore that the ill-documented Samnites and Brunii would gene rally look similar to the bener illustrated Lucani and Campani. Figure 144 is based on a srarue made ac Taras abouc 280, probably a Lucanian or Samnice. hs missing head has been replaced by a Lucanian helmet from the Paesrum paintings. 145 is from paintings ac Naples, 146 from Paescum.

163

Oscans wore their black or dark brown hair short, usually shaved, but occasionally wore short beards, most often without moustaches. They wore a short wool or linen tunic, and sometimes a cloak; this is rarely shown worn by infamrymen though one at Paestum uses bis cloak for an improvised shield. Tunics are brightly decorated, on a base colour which is most often red or white, but could be crimson, red-brown, ochre, light grey or blue-grey. Ali tbese colours plus black, yellow and blue were used in decoration. There is little evidence ofuniform. Livy in one passage has a Samnite unit uniformed in whĂše, another in multi-coloured tunics, but this is probably worthless as the whole passage seems to be fictitious. He also describes the elite "linen Jegion" in white, which may be more reliable but suspiciously echoes the previous passage. Most men are shown barefoot, some in sandals or low black shoes. A broad leather belt covered witb bronze sheet, fastened with elaborate hooks, was uni versai; it seems to have been a socia! or religious symbol ofmanhood rather than purely military.

Occasionai warriors are shown bareheaded, but most wear bronze helmets, and the Attic style is by far the commonest, replacing the earlier wide-brimmed pot helmets in the 5th century. lt was usually crested with horsehair and/or feathers. The horsehair crest would be red, black or white, or alternate two or three colours. lt could either be directly fixed to tbe skull or raised on a stilt. Feathersare black or white in most cases, though 144's are yellow, fixed in bronze tubular or spring holders. Crest variations not shown in the main figures are in 144a (seen on both Campanian and Lucanian vases), b from a Lucanian vase, and c from Paesrum. 144d is a bronze helmet ofthe CelĂšc Montefortino type with a five-branched iron feather-holder. This type ofhelmet was adopted in Samnium as elsewhere in Italy. One bronze Thracian helmet also comes from Samnium.

Bronze greaves were sometimes worn. Livy says the Samnite wore one greave, but he is probably describing the equipment of later "Samnite" gladiarors; one greave is certainly possible, as the Roman example (figure 132) shows, but Oscan sources, including our solitary Samnite statuette, show rwo if any.

Main weapon is the javelin; paintings show up to four, 3-6 feet/l-l .8m long, with throwing loops, iron heads but no apparent buttspikes. Even tbe longest ofthese are light weapons; the Romans had a tradition that they had adopted the pilum from the Samnires but there is no evidence any of the Oscans ever used this weapon, and convincing nonOscan explanations ofits origins. 146 also has a short thrusting spear witb heavy naring iron butt, which is gripped well back presumably because tbc heavy butt ensures a centre ofgravity well behind the middle oftbe shaft. Swords are rarely shown in art, and when they do appear they are slashing weapons of kopis type, presumably adopted by the coastal peoples from Greeks or Etruscans. 144's is similar, though unusually short and broad. The suggestion that Oscans had no swords at ali till they adopted them from the Greeks, however, scems to be refuted by tbc presence of short straight swords in early Samnite graves, though many contain only long daggers. Tbese early swords, too, are wom at tbe left hip, probably from a baldric Like 144's and 145's later weapons.

Most Oscan warriors probably used the ovai scutum, concave and ribbed like the Roman shield, but smaller.

Dionysios calls the Samnites at Asculum 1hureopl1oro1~ meaning they had scuta (Greek thureos equating to Latin sc111um). 144 has an ovai scuwm; the blazon is from a Campanian painting, where a shield ofthis shape is carried as a trophy. lt is yellow with black stars asymmetrically placed. 144e, from Paestum, has the same stars in black on brown, with a lighter brown rim; it is unusually broad and lacks the centrai spine. So does 145's, which is deeply dished; it is white witb a red star. Livy auributes to tbe Samnites tbc curious shield shown in 144f, with a flattened top but narrowing to tbc bottom for lightness, Like a mediaeval kite-shield. This style ofshield was carried by the class of gladiators called Samnites, and again Livy may merely be describing gladiatoria! equipment. However, the shield is a perfectly reasonable developmem ofthe ovai scutum, uimmed at the top to make it lighter and handier, and indeed Imperia! Roman troops similarly trimmed both tbc top and bottom oftheir scuta. There seems no good reason, then, why tbe kite-shield ofthe Samnite gladiator should not have becn adopted unchanged from Samnite soldiers. There is, however, no supporting evidence for it.

Small round shields, dished with narrow flat rims like miniature versions ofthe hoplite shield, are common in the Pacstum paintings and also carried by one Samnite in che lost Allifac paintings. Ifthesc wcre common in Lucania but rare in Samnium, this would cxplain why Dionysios distinguishes the Samnites, but not the Lucani or Bruttii, as thureophoroi. As well as the shicld the Paestum warriors carry javelins in the left hand, dose to the horizontal; this argucs that the shield was carricd by a centrai handgrip held horizontally like that of the scutum, as a Greek-sryle double grip would Jead to javelins being hcld vertically. The sourcc for 146 shows plain bronze shields; 146 a and bare blazons from drawings oflost Paescum paintings. The star is popular in most ofltaly. The late lcxicographer Festus says tbc Brunii used a panna, a light round shicld perhaps like this figure 's.

164

147 and 148.

OSCAN ARMOURED INFANTRYMEN

Some infantry wore cuirasses of a distincùve ltalian design, small brcast· and backplates ofleatber·backcd bronze linkcd by hinged straps of similar construction passing over tbc shoulders and under tbc arms. These gavc only partial protection to the torso, though they tended to increase in size. The earliest had merely a single disc front and back; rhis typedates back to tbc bronze age, but 153 shows it was stili used. Commonest bowever is 147's triple-disc sryle. Most are quite plain, but thcre are two magnifice.n t decorated examples (one from Apulia, one from near Carthage, perhaps taken tbere by onc ofHannibal's veterans) wbicb have the lowest disc replaced witb a bust ofa goddess in rriple-crested belmet. 147a shows the African example; breastplate is on tbc left. 148 wears a less common style with square back and breastplates embossed with muscles in imitation ofGreek cuirasses. lt is shown on three paintings, worn by a borseman and footman from Paestum and a horseman from Capua. Severa! cxamples survive, most ofunknown origin; one is Lucanian. A style witb larger plates, almost completely covering the torso, is probably a later development perhaps ofthe late 4tb or 3rd century.

147 is based on a Paestum painting, but the Samnite statuette from Sicily is armoured tbc same, and also bad crest and side featbers (tbey are missing, like bis weapons, but locating boles remain). Tbc shield is a rype common in Lucania.o art, apparently ofwicker, rising to a centrai point: Couissin compares it to a Chinese bat. Tbis assumes it is round, but the profile illustraùons in art do not make rhis clear , and I suspect it is ovai and is tbc scucum of osiers covcred witb bide which Servius assigns to the Lucani. Again, the position of javelins in tbe left hand suggests a scucum·style single borizontal handgrip.

148 has an elaborate panoply found in Lucania with square cuirass, strapped greaves and Attic belmet witb sbcct bronzc wings as well as crest and feathers. He has a Greek·style hoplite shield shown carried by javclinmen in Lucanian and occasionally Campanian art. 148a is the blazon oftbe Paestum shield; centre and rimare polisbed bronze, with a ligbt gr ey wreath on a broad white circle. A Campanian painting shows javelinmen witb yellow (polished bronze?) hoplite shields with centrai red discs.

Notali armoured men wear greaves. Armourcd javelinmen do not feature in Campanian art, perbaps because those infantry who could afford bronze armour used tbe boplite panoply like the next rwo.

165

149 and 150.

CAMPANIAN ROPLITES

41h century Campanian vases show many warriors in Oscan dress and armour but wi1b the Greek boplite's long spear and large round shield, no doubt with Greek swords 100. Coins of the Mamenini, the Campanian ex路 mercenaries ofSyracuse, also show hoplites. Tbese two are from che vases. 149 has the triple-disc cuirass and Attic helmet, in this case with three-disc cheekpieces styled in imitation ofthe cuirass, which were adopted by the Celts (figure 125). His sbield is shown plain white. 150 is unarmoured, and his helmet lacks cheekpieces. His shield路rim is faced with bronze, but not the rest ofthe shield, which has a star blazon.

151 , 152 and 153.

OSCAN C AVALRY

151 and 152 are typical of the 4th century cavalry in the Campanian and Lucanian tomb paintings, and also in thc lost Samnite paintings. Roughly halftbe cavalry in an seem to bave metal armour. One instead wears a wbite Greek linen or leather cuirass. Cloa ks are worn more often than infantry, who would obviously be more hampered by them. 151 has his gathered up over his left arm, perhaps as protection as cavalry in most ofthe paintings lack shields.

He also has a riding路whip dangling from his left wrist. Cavalry rarely wear greaves, often replacing them with bronze anklets. The paintings show them riding bareback, but see note to figure 175. Weapons are javelins; Livy confirms this, mentioning Samnite cavalry throwing javelins. Swords are likcly.

153, from Paesrum, unusually has che coumerweighted tbrusting spear instead of javelins, and wears greaves as well as a bronze disc on cross路straps. He carries a shield, so may be a little later than the last two and date to a period when some or ali Oscan cavalry had adopted shields. A Campanian vase of the end of the 4th century shows a borseman with a ribbed scutum, and this man's shield 100 scems to be an ovai scutum, white inside with a bronze rim.

T his suggests shields were adopted somewhen in the late 4th cenrury, though of course their generai adoption may have been a very slow process. The 3rd or 2nd centu ry evidence discussed under 140 also indica tes shields.

154.

OSCAN STANDARDS

I 54a is a bronze cockerel standard on a tali bronze pedestal, found in Samnium. Otber possible animals for standards are the bull, associated in myth with the Samnites, or the wolf, connected with the Lucani and the Samnite tribe ofthe Hirpini. Many paintings show warriors carrying spears over their shoulders, with bright cloths and yellow streamers hanging from them, and these are often interpreted as flags. However, they are not. T he clearer of the paintings show that the "flag" is a runic, the spear passing through the armholes, while thc "streamer" is the Oscan broad beh; the spearhcad is thrust through the bronze, and the belthooks are often shown, as in I 54b. These represent trophies taken from dead or captured enemy who have been stripped - as was a common practice in both haly and Greece. The custom of so displaying trophies goes back to Greek rule in Campania, and vases ofthe ea rly 5th century show Greek horsemen with tunics hanging from their spears, driving naked prisoners before them. Confirmation that 1hese are trophies, not standards, is provided by the bloods1ains thai often mark che tunics, and che fact that greave.s and even shields are some1imes added 10 1he trophy - 144's shield blazon is from just such a captured shicld.

155, 156 and 157.

APULlAN INFANTRY

5th and 4th cenrury Apulian vase paintings show locai warriors in a costume similar to Oscan sryles, with short tunic and broad belt. However the Apulians seem to have favoured deco ration with vertical stripes or panels, perhaps reflecting the Illyrian heritage (see figures 73-75), wore high laced boots, and some1imes favoured longer hair. Some warriors are shown bareheaded, with or without the headband of 155.

The most popular headgear was the Greek-style pilos (which was also worn, sometimes cresccd, in Campania).

Severa( bronze examples havc been found, bue l 56's tali and somewhat floppy example is clearly non-mecallic, perhaps felt, as may be l 56a. l 56b is a bronze pilos found in Apulia, decorated in Ccltic style with sheet bronze horns, cwo wheels as a crest (compare l 24a and j), and embossed with a face and rwo running dogs over che forchead. l 56c, in Naples museum, also has shcet bronze horns bui with an Italian-style stile crest holder. Severa! ofthe 4th century Ccltic incursions imo Italy reached Apulia, and must be responsible fo r che Celtic influence seen in some Apulian armour. The Etrusco-Corinthian helmet worn by 134 and 162 was al so used, and may ha ve been developed originally, in Apulia.

Infantry are shown unarmoured, with javelins and shield, often the Jarge round style carried by 155. This is similar 10 the Greek Argive shield bue with a far narrower rim. lt is decorai ed with a pattern of dots arranged in concentric circles, which may perhaps be bosses on a bronze shield-face or metal studs on a leather one. l 55a, with a much broader rim, mayactually be an Argive shield. Smaller types like l 55b are also seen. I 55c shows che side view of one shield as it appears on one vase, rising to a poi ne; 1his may be the high point of the centrai rib of a small ovai scutum,

166

153

152

154

167

which would look like I 55d in front vicw. Despite the use ofthe Argive or similar large round shields, and the Greek influence visible in body-armour and helmets, the long spear of thc Greek hoplite does not seem to have been adopted, infantry being shown with cwo or three throwing spears. A short sword or long straight dagger was also carried.

While body armour is relatively frequent in Lucanian and Campanian art, Apulian warriors always seem to be unarmoured, suggesting body armour was limited to officers and nobles. These cenainly did use it, as several cuirasscs survive. 157 is a reconstruction based on a bronze panoply from a 4th century tomb from Conversano in Peucetian territory. He has muscled cuirass, greaves, and a helmet which combines Thracian and Attic fearures. A bronze belt was also found in the tomb but seems impractical to wcar with the cuirass. The helmet has a fine wave decoration cresting the tali skull, which macches the decoration down the sides of che cuirass. lt also has sheet bronze wings with feather holders behind them, and curls ofhair represented in stylised relief above the forehead.

This hair dccoration is charactcristic of a group of similarly elaborate and expensivc bronze ltalian helmets, of which I 57a and bare examples; both ofthese also have other relief decoration. Another helmet with similar curls abovc thc forehead, and with a wave decoration like that of 157 along its skull, which is much taller and ends in a duck's hcad, was recenĂšy found in Cyprus.

e

a

e

168

d

158, 159and160.

APULIA N CAVALRY

As with the Campanians and Lucanians, Apulian cavalry are shown riding bareback, bue again see note to figure 175. Many of the cavalry on the vases are unarmoured and shieldless, like 158, who has a pilos with sharply outturned rim which must be bronze. 159, from a vase ofthe mid-4th century, however has a small round shield. He wears a Thracian helmet with crest and feathers. 160 is a reconstruction of a heavy cavalryman, with a Greek-sryle muscled cuirass and Celtic helmet found together in a late 4th or 3rd century grave at Canosa in Peucetia. The bronze plate cuirass flares widcly at hip leve! likc that of25, indicating ii was made fora cavalryman. The helmet, iron and bronze with red coral mounts, is ofthe Montcfonino typc, embosscd with Celtic sryle spirai decoration, and with ltalian style feather holders as well as a centrai crest-holder.

161 , 162 and 163.

ETRUSCAN HOPLITES

By 600 the Etruscans had adopted the Greek hoplitc panoply, and they kept much ofit, in panicular the large round shield, till they lost their independencc, and beyond. Greek cquipment was supplemented by native ltalian, distinctively Etruscan, and Celtic items. Tunics were usually red, perhaps edged in colours including white, yellow and blue. Other tunics are whicc edgcd in crimson or purple. h is tempting to connect these with the purple-edged white togas ofthe Roman aristocracy, and suggest that the white tunics denote Etruscan nobles and officers.

161 wears perbaps tbc most familiar style ofEtruscan armour, thc Attic helmet and lamellar cuirass shown on a number ofbronze statucttes. Tbc cuirass is of flexible, probably linen, materiai cut in Greek style with preruges, tied-down shoulder-pieces, and a rcctangular flap to cover thc back ofthc neck, but is covercd ali cxccpt the pteruges witb metal lamellae, narrow rectangular overlapping metal plates, laced co che foundation. Scaled cuirasses ofthe same cut were used, and so were otbers witb no metal reinforcement (at least externally). These were white or a pale straw colour, decorated in red, rose and black. 162, bascd on 2nd-ccntury sculpted urns from Clusium and Volterra, wears a quilted variant with metal scales reinforcing the breast and narrow shoulder straps. It could alternatively have quilting reaching funher down the hips, with shorter pteruges, and an additional scaled scction over the abdomen. 163 has a muscled piace cuirass worn, as often in ltaly, without pteruges. Most ofthese were bronze, but some are sbown in an painted grey; this could perhaps indicate tin-plated bronze, but iron piate is perfectly possible by tbe 3rd ceotury (see tbe cxamples memioned under 35). The old h alian disc-breas1plate was stili in occasionai use in the 4th century.

161 's Attic helmet is a sryle long f voured by Etruscans, often shown with the cheekpieces hinged upwards. 162 has an Etrusco-Corinthian helmet ' debased version ofthe Greek Corinthian style; this could only be worn on top of 169

the head, and thc cyeholes and nasal bar wc re purelydecorative. lt could ha ve cheekpieces. The crest could be raised on a stilt, and featber-holders added at Lhe sides. 162a is an Etrusco-Corinthian helmet with transverse crest, shown on an urn from Volterra. This stylc ofcrest was also worn with the Negau hclmet, as sbown in 162b. lt was later used to distinguisb Roman centurions and migbt therefore similarly bave distinguished Etruscan officers. The Negau helmct was used from thc 6th century to the 4th or bcyond. 162c is an unusual 4th-century examplc with large cheekpieces embossed and engraved to produce a pania! mask, with beard, moustache, and pierced mouth nnd nostrils. In the 4th century the Celtic Montefortino helmet was adopted, 163 wearing an example found with his cuirass in a tomb at Orvieto. This bas three-disc cheekpieces; thc scalloped style shown (for instance) by 144d was also used, and in che 3rd century may have displaced tbc three-disc style entirely. H elmet crests were white or red.

Most hoplites wore bronzc greavcs, ofGreek style, except that some lacked modelling at the knees, like 161 's. The tomb which provided 163's hc lmct and cuirass also hcld greaves, but I ha ve shown him in short boots, as sometimes worn by similarly armoured men in art.

Shield blazons resemb led Greek styles. 161 a-e are ali from paintings in the Giglioli tomb at Tarquinii, about 300 BC. Ali ha\•e white dcviccs on red circles; the edge ofthe shield's flat rim is black, the rest ofthe shield-face bronze.

161 a's "A" shape is not a tener, but a surveyor's leveI; botb this and the boar of 161 b occur on Tarquinian coins.

161 d is from a va se of similar date, from Volterra. Insides of shields are shown black, light grey, or light blue.

Swords wcrc mostly ofthe straight Greek type, worn from a baldric. A longer version ofthe curved kopis was also used. The long thrusting spear remained in use, but many men replaced it with pila. While long-shanked javelins ancestral to the pilum werc used long bcforc our era, the eartiest javelin which can defĂŹnitely be called a pilum is 163a, a Stb-ccmury socketed Etruscan weapon from Vulci. lt is unusually long, 48 inches ( l.2m) from socket 10 point. 163 carries a rivc1cd versi on based on a 4th-century examplc from Grosseto, which is matched by pila painted in the Giglioli tomb. The pila from Telamon mentioned under 132 are very similar, and may perhaps be Etruscan rather than Roman.

a

161 b

e

d

170

162

163

164.

ETRUSCAN CAVALRYMAN

From a fune rary urn from Clusium, perhaps 3rd century, where he is shown riding down Gauls. Hc wcars an iron mail cuirass, siroHar to I 62' s in that it has scales on thc breast, narrow shoulder-srraps, and pteruges. Headgear is a T hracian helroet, somet imes shown in Etruscan art from the 4th century on, sometimcs crested or with wave decoration like 157. H e also wears a Hellenistic officer's waisr sash. He is armed with a thrusting spear and a shield, the shape ofwhich is nor entirely clear. It is probably round, and may beone or other ofthe Roman cavalry shield types.

165.

ETRUSCAN U GHT lNFANTRYMAN

A Roman spy ve nturi ng imo Etruria in 3 10 dr essed as a herdsman and carried "rustic weapons", two javelins and afalx. The word means scyrhe or sickle, and need not indicate anything more precise than a curved biade, with rhe cu tt ing edgc on the inside ofthe curve, like a sickle. lt would li t the kopis-like swords, bu e in the rusric context is more likely to mean some large gencral-purposc knifc or falchion. T his is probably the sty le ofweaponry used by Etruscan skirmishers and peasant levies. Some may bave had shields. Archers and slingers are also shown in earlier Etruscan art. T he broad-brimmed hat is an occasionai item ofEt ruscan dress, especially for the lower classcs. Ii is sometimes shown perched neatly atop the head like the Greek petasos (ligure 26a), bur one relief shows a ploughman in a rather floppy hat with dow nrurncd bri m, pe rhaps more rypica l ofpoorer countrymen.

17 1

166.

PERSIAN CA VALRY HORSE

.Persian horses, especially che famous Nisaian breed of Media, were known for their size and mength. In facc reliefs suggesc they were noc especially cali, even by ancient scandards - the horses sculpced ac Perse polis scale up ac about 14·15 hands - bue were heavily and solidly builc, with large heads and heavy necks. Warhorses were generally scallions; some were crained co rcar up and strike with their hooves againsc infantry. Most were black or chestnuc, gene rals somecimes riding whice horses. Ocher parts of the Persian empire produced good horses, Cilicia for instance providing the Great King with 360 white horses as tribute every year. Armenia also provided horses in tribute, Xenophon saying chey were "smaller than the Persian horses, but much more finely bred". Later writers say Armenian horses had Panhian blood, perhaps indicating they had been crossbred with the Nisaian herds now under Parthian ownership, and were fairsized, spirited but clumsy. Kappadokian horses seem to bave been similar.

Typically this horsc has his mane cut short, and both the forelock and the tait are tied up with ribbons, the forelock forming a piume. The bridle could be red or pale brown leather, sometimes bronze studded, or oflinked bronze plates as here. Boars'-tusk-shaped ornaments of stone, bone or bronze could be worn where the straps crossed.

Reins are usually red edged white. The bit could be bronze or iron, and some are qui te severe. The large saddlecloth is rypically long on che back and hcavily embroidered, often with scalloped edges. Sometimes several cloths were worn - "they have more covcrings on their horses than on cheir couches". T he cloth is he ld on by a similarly decorated breast strap and probably a girth. The beli shown hanging from a collar is an optional extra occasionally seen in Persian art.

167.

PERSIAN ARMO UREO HORSE

Metal horse armour was first uscd by the Sakic Massagetai and adopted by the end of thc 5th century by the Persians, though only fora minority oftheir cavalry. At this epoch the armour was limited to a poitrail for the chest and a chamfron for the head, the horse's flanks being in addition protected by the parapleuridia of the rider (figure 5). Herodotos says Massagetai borse armour was bronze, and Xenophon says the same ofthe Persians, but neither describes its form in detail. Curtius says the horse's breastplates were ofmetal platcs, lamnac, linked to cach other. This could dcscribe scale armour, but then one wouJd expect the more familiar term squamae, so he probably means lamellar armour. This figure wears a lamellar poitrail from the Pergamon reliefs, which though Hellenistic may derive from such a Pe rsian prototype. The onty son of chamfron known from this period is the Greek style of narrow bronze pi ate worn by figures 49 and 168, and this may indeed ha ve been the srylc uscd by Persians. However I have shown an alternative possibility, a chamfroo ofscales used by later Kushan cavalry, as this derives from the centrai Asian steppe tradition and may therefore resemble styles used by earlier Saka and Persians.

This style of armour would ha ve been used by the 2,000 Massagetai at Gaugamela, the 1,000 Bactrians brigaded with them, and probably thc Armenians and Kappadokians on the other wing. However the Seleucid cataphracts introduced by 200 BC seem to have wom the later style of complete borse armour, illustrated by figure 140 in Armies and Enemies of Imperia/ Rome (a companion volume co this), and to bave adopted it from thc Parthians, indicating tha t full horse armour was in use among the Parthians and their Saka kin by the 3rd century, as it certainly was later.

168.

SCYTHED CHARIOT HORSE

The four horses of the scythed chariot (for which see figure 176) were armoured, with headpiece, poitrail and parapleuridia. Obviously these cannot be the dual-purpose parapleuridia/parameridia of figure 5, and must just protect the sides of tbc horse, like a large saddle·blanket. Xenopbon implies that ali thcse defences were bronze, which would givc an effect mucb like latcr complete cataphract horse·armour, but that seems unduly heavy for this early date. I ha ve therefore sbown chamfron and poirrail as mecal, but parapleuridia as heavy cloth or leather, which could be quilted, and could bave developed from tbe earlier Assyrian caparisons.

The collar (red leather decorated with white ducks), girth and other hamess and decoration come from Darius' chariot-horses in the Issos Mosaic. These are an unarmoured team offour matched blacks.

169.

HELLENISTIC CAV ALRY HORSE

Greek horses were smaller than Persian and regarded as inferior cavalry mounts. Contcmporary art suggcsts tbere was littlc difference in beigbt, but tbc Greeks were lighter, with smallcr heads and fincr legs, gene rally less muscular and powerful. OfGreek horses, Thessalians were the best and fastcst . Coins of Philip II show strongcr, hcavicr horses than in earlier Greek art, so he may have improved the Macedonian breed. Thc traditional colour for Grcck horses was xa11thos, the colour offair human hair, in horses pcrhaps dun but more likcly a light chestnut. By this

172

pcriod howcvcr rcd-bays, dark greys and blacks are also found. Paintings show Macedonìan horses as chestnut or occasionally black. Alexander's famous Boukephalas was black, perhaps taking his name "oxhead" from the shapc of a whitc blazc on his forehead, whilc Alexandcr rides a chestnut on the Issos Mosaic. His father Philip rides a white horse in a hunt scene painted in his tomb. Thracian horses in the Kazanluk paìntings show a similar range of colours.

H orses are often shown bareback, but this may be artistic convention, the better to show the lines ofthe horse's body. Xenophon mentions a man may ride barcback, but assumes saddlecloths are standard, at least for war. Most cloths were simple rectanglcs, but the Grccks of Asia had long used scalloped Persian-stylc cloths. Some caparisons were large cnough to cxtend round the front ofthc horse's cbest as in figures 29 and 64, or whole animai skins could be used as in 31. Cloths would be held on by brcast strap and girth. This figure, based on a coin of Philip II, seems to havc a sccond cloth wom over the first, and perhaps a thick saddle·pad, shown more clearly by a later terracotta.

173

Xenophon recommends a thick quilted saddlccloth, and the pad may derive from this. Bridles were similar lO Persian styles, of red or dark brown leather with bronze finiags, or else ofbronze plaques. Reins could be the sa me but those on 1he lssos Mosaic are white, with two narrow yellow lines running the length ofthe rein. Bronzc or silver harness omamems, sometimes gilded, could be wom on a decorative collar like figure 175, or else (especially

by Thracians) on bridle or breast strap. Thracian omamems were like Scythian, fea1uring real or famastic ani mais; I 69a shows a Thracian bridle. One grave held 20 ornamems, 11 from a bridle, 9 probably from the breast si rap.

Some Thracian riders had che Scythian saddle of 170.

Greeks and Macedonians campaigning in Asia might ride locai horses and use locai hamess. Thus Alexander issued his Companions with Persian horsc furniture after caking Persepolis, and with silver harness ornaments in India; Antigonos One-eye collected 1,000 fully harnessed Median horses before Paraitakene; and a painting from a Ptolemaic colony in Palestine shows a lancer whose horsc has a hcavily decoratcd Pcrsian-style saddlecloth anrl tasselled breast-strap.

170.

STEPPE NOMAD'S HORSE

This figure is takea, like 82, from early Sarmatian gold plaques, but is also typical ofthe mounts ofScythi:ms, Saka and the related Pazyryk people. Most of 1he horses buried in 1he Pazyryk grave mounds were bay or chestnut, varying from dark brown to yellowish, with a few blacks. The majori1y werc of the Przewalski breed which stili survivcs in centrai Asia, small, hcavily built, hardy ponies, distinguished by their naturally short manes and long tails which nearly reach thc ground. Scythian ponies were similar but perhaps interbred with lighter built western stocks, :md are illustratcd wi th shorter tails. However each Pazyryk barrow held one or more larger horses, obviously valuable beasts as they had bccn fcd on grain while the ponics were turned out to graze. Thcse wcrc probably the famous "heavenly horscs" ofFerghana. The armoured horscs of the Saka nobility would be of similar stock.

Mos1 nomad horses had their manes shon, naturally or cropped. This one has its 1ail plaitcd; 01hcrs at Pazyryk had them knotted at halflength. Scythians however usually left their horses' tails loose.

The nomads did noi ye1 use stirrups, but had a simple saddle, in i1s mos1 primitive form merely two cusbions strapped beside each other along 1he horse's spine. A more advanced model bada simple wooden frame providing a low pommel and cantle. Thesc would be faced with decorated wooden arcbes and the whole saddle could bave a highly dccorated felt cover, and be worn over a felt or Chinese silk saddle-cloth. The Pazyryk cloths are larger 1han 1his figure's, and some are 1asselled. Bridle, breast strap and other harness were decorated, sometimes profusely, with a variety of plaques. This figure's harness is qu ite lightly ornamented, and from their similarity to some Pazyryk types the kite-shaped plaques are probably wooden, painted rcd or gilded. Sarmatians also used silver or gold disc phalerae, while the Scy1hians favoured a variety of metal ornamems in tbc dis1inc1ivc "animai style".

171.

INDlAN CAVALRY HORSE

Ancient sculprure suggests Indian horses were kep1 rather on the fat side, and also 1ha1 most warhorses were stallions. As for colours, the Mal1abliara1a mentions dark grey, "pigeon-coloured" (lightcr grey?), ivory-white, piebald and red-bay horses. Arrian records that the horses werc controlled by a leather noseband, studded on the in si de with bronze, iron or ivory spikes. The bit was a simple plain iron mouthpiece, connected both to the reins and the noseband, and the spikcs on the noseband servcd lO rcinforce the action ofthe bit. The bridle, saddlecloth and other trappings were elaborate ly decorated, with the horse 's forelock drawn up through a tube 10 act as a piume.

This figure's harness, with the girth apparcntly right at the front ofthe embroidered saddlecloth, is from a sculpture from Samath ofthe 2nd cemury BC.

172.

NUMIDlAN PONY

Numidian and Moorish ponies were, like their riders, small and lean, "their mounts were withou1 bridles, and even their movement was ungainly as they troned with stifT necks and ou1s1re1ched heads". They were, however, docile, hardy, and quite fast, though their small size was one reason why Numidian cavalry could not deliver an efTective charge. Ano1her was 1hc harness, or ra1ber lack ofi1, merely a rope about che neck by which the rider could exert some comrol on the horsc, partly by exening pressurc on the windpipe. Livy and other writers cali hea\•ier cavalry "bridled" 10 comrast 1hcm with the Numidians. The rider would control the horse by tapping it with a stick between the ears. The later poet Oppian calls Moorish horses dappled. The horses of Kyrene were similar 10 Numidian mounts, but larger.

174

173.

SPANISH CAVALRY HORSE

Spanish horses were fast, swift enough to catch thc spccdy Numidian cavalry, and Strabo says they were hardier than most. Oppian however claims that, though very fast, they lacked staying power, but he is wriling much later and the breed may have degenerated. Other writers say Spanish horses had fine hcads carried high, but oarrow waists and poor quarters, and 1hat they we re uoruly, their hard mouths making them difficult to conirol even wi1h a severe bi1. Agaio, this may only bave been true later than our era, as most Spanish bits were fairly mild snaffies, and Strabo says the horses were well trained, even kneeling obediently for their riders 10 dismount.

The tack shown is fairly 1ypical, wi1h a blanket, here a double 1hickness, held in piace by a surcingle and breas1 strap. Furs or skins could be used over or instead ofa blanke1; some blankets are shown 1hick enough 10 suggesc 1hey might be padded. One 3rd-2nd century vase paincing shows che horns ofa saddle, probably 1he Celtic type ofthe nex1 figure, but chis was not usual. The casse! shown at thc chroa1 and the forehead piume are not always present, or 1he piume could be worn on a noseband.

175

174.

CELTIC CAVALRY HORSE

What littlc is known oftbc borscs ofGaul sugges1s 1hey were lightly built, and probably quite well bred since, as Caesar says, the Gauls were inordinatcly fond of good horses, and would pay grea1 sums for 1hem. No doub1 expensive importcd stock improved che native bloodlines.

The tack is interesting, as the Gauls seem to ha ve developed the 1ype of saddle used la1cr by 1bc Roman Empire. A fallen Gallic borse on 1he Julii monument from St Remy in France bas such a saddlc witb its cbaracteristic four horns or handgrips at the comers, and the handgrips are also visible on che Gundes1rup cauldron. Roman saddles ofthis type were offairly tbin tanned goa1skin, with bronze stiffeners forche horns. This would give very litùe support so would bave to be worn over a pad or cushion. Tbe Julii monumcn1seems10 show such a cushion, and also a small saddle·blanket beneath it. Thc bridlc is fairly simple, usually with a snaffie bit thougb some qui te severe curb bi1s are known. Breasc and crupper sm1ps are ornamen1ed with round silver plialerae, usually embossed with a ring ofsevered hurnan beads. A real severed head migbt be carried from one of1hc saddle homs as a trophy; Livy 1clls how che Rornans learaed ofthe defeat ofonc of1heir forces by the Senones only when they saw Gallic cavalry with Roman heads bung from 1bcir saddles or brandished on their spears.

175.

OSCAN CAVALRY HORSE

ltalian cavalry equipment was generally like Greek. For inscance 1he Roman Mcnius Curtius relief shows a cavalryman with a simple rec1angular saddlecloth and broad breast strap, plus a narrower strap round tbe rump.

Oscan equipment in che Capua and Paescum painiings bowever has some dis1inctive features. The horses are sbown bareback, but his may, as wi1h 1he Greeks, be an artistic convention. Livy mentions Samnites witb gold· embroidered saddle clotbs, bui unfortunately in a generally unreliable passage, so che reference may be useless.

Some horses wore armour, a chamfron and poitrail ofbronze piace lined with linen or lea1ber and embossed wi1h considera bi e artist ry. Such armour was used by the Italiote Greeks in the Sth cen1ury and adopted from chem by the Oscans. From tbe quality ofthe work i1 musi bave been very expensive, and hence rare. The cwo pieces of armour are not always worn cogethcr. Whcn thc chamfroa is absent, 1he nose and forehead straps ofche bridle can have bronze ornamems. Tbe upright fcathcrs, probably in holders attached to 1he chamfron, are not common. Collars wi1b bronze plialeraeare fairly common, sometimes onlyoae being worn, somctimes two as bere. 1t seems likely that tbe lower collar helped suppor! the poi1rail, 1bough il is also found on unarmoured borses wbere it can only be decorative.

The Roman poet Lucilius in 1he 2nd ccntury BC says Campanian horscs were fiery and menlesome, though lacking 176

176

177

the endurance ofSpanish horses. Lucanian horses are not distinguishable from Campanian in the paintings, both being high-stepping and proud, but later writers cali Lucanian steeds small-bodied and ugly in appearance and colour, though hard workers. The horses in the paintings are mostly chestnuts with light-coloured manes and tails, less often blacks, in eithcr case usually having white legs, feet and faccs. A Frentanian cavalry officer who auacked Pyrrhos rode a black borse with white feet; Silius ltalicus mentions a Roman consul on a white steed.

176.

SCYTHED CHARIOT

These were used by the Achaemenid Persians, aod copied from them by the Seleucids and later by Pontos. The besi descriptioo is in Xenophon's Kyropaidia, other sources giving different arrangements of scythes but no overall description ofthe chariot. This drawing is bascd on Xenophon plus the unscythed chariot on the Issos Mosaic and a small gold model from the Oxus ueasure. lt has heavy wheels with studded rims, a fairly broad wheelbase for stability, and two poles. The yoke on the Oxus model is a simple straight bar, carrying four rein-guide rings.

Xenophon refers to a door in an otherwise closed back, and a seat; the Oxus model is open at the rear but confirms a simple bench seat. The purpose ofthe straps attached to the front edges ofthe car is unclear; they do not appear to be reins, nor traces for the outer horses, though they may be part ofthe harness ofthe inner pair.

Xenopbon describes 3 feet (90cm) scythes on the a:<ie-ends, and others along the underside ofthe car. These would havc been dangerously vulnerable to the slightest irregularity of the ground, andare not heard of again. Diodoros says the chariots at Gaugamela had 3-span (27 inches/68cm) blades at the ends ofthe yoke, and longer, broader blades on che axle-ends, curved forwards at the tips. The arrangement shown is that Livy describes for tbc Seleucids at Magnesia; 3 feet (90cm) blades projecting "like horns" one each side ofeach pole, rwo blades at each end ofthc yoke, one horizontal, the otber pointing downwards, and two smaller blades at thc axle-ends. Decoration shown is that ofDarius' unscythed vehicle; the car is grey-brown with bands of dark red ornamentation, edged and bearing rows ofsmall winged ani mais (gryphons?) in white. 176a is alternative decoration from the front ofthe Oxus model.

Scythed chariots were drawn by four armoured horses (figure 168) and manned by a single armoured driver (figure I O). lt was important for the vehicles to build up speed, so they were vulnerable to swift pre-emptive attacks by light troops. Once thc cba riot was galloping 1owards its target, the driver would often bai! out for his own safety, which unfortunacely made it likely that the horscs would miss or shy away from the target. Ifthey did make comacc, the chariots would cause immense damage but would more or less destroy themselvcs by the collision. Generally formed infantry bodics could stand firm after such contaci unless the cbariots were closely supported- which they all too often were not. Cavalry might have been casicr to break than solid infantry bodics, but scythed chariots do not sccm to bave beeo used against them, perhaps because cavalry could have avoided their charge.

177.

INDIAN C HARIOT

Tbe earliest Indian chariots were light two-horse vehicles with a driver and one warrior, and che epic tradition, despite numerous later accretions, still has its heroes fĂŹghting in this way. There are occasiona} references to chariots with three or four horseseven in very early sources, and both 2- and 4-horse vehicles were in use throughout our period. The form ofthe chariot seems to have becn the same regardlcss ofthe number ofhorses, though details do vary. Size also varied, chiefly to accommodate different numbers of crewme n. The Arrhasasrra records seven sizes of chariot. The largest is I Op11rusas (7ft 6" /2.28m) high by 12 p11rusas (9 feet/2. 74m) long. The different sizes then diminish by one purusa (9" /23cm) in length till the smallest is 6 p11rusas (4ft. 6" / l.37m) long. How many of these sizcs re present war chariots is not certain, since the Ar1hasas1ra also gives thc superintendent of chariots the responsibiliry for various ceremonial and travelling vehicles. War chariots in art are about 4ft 6" -5ft 6# ( l .3- l .6m) high at the front rail, with the body about rhe same length or a little shorter, fitting the smaller end of the Arrhasastra's range.

The chariot body was oflight construction, probably lcather on a woodcn framework, and was decornted, especially round che edges. lt could be adorned with bells and tigerskin coverings, and the floor could be carpetcd. There was usually a flagstaffat tbc rear ofthe body, and some chariots had a canopy for protcction from rhe sun, though this may not have been used in battle. Thc wheels were small, with many spokes- from 16 possibly up to 32 - and iron tyres.

Thcrc was only one pole and yoke, even for four-horse .::hariots, where only the inner pair was yoked, the outer horscs attached by traces. Two-hor$e teams also had side traces but it is not clear from represcntations exactly how or wbcrc rhcy were attached; they are somctimes obscured by the horses' tails, which seem to be tied or rucked into tbc traces. Otherwise chariot horses were harnessed by breast straps, ginh and a bridle similar to cavalry horses.

Thcre are occasionai references to chariot horses being armoured, once in " leather robes".

178

Just before the start of our period there is refcrence to a chariot with four men, but ancient commentators interpret this as driver, one crewman, and two escorting infantry, so the two·horse two-man vehicle may stili have been in use. Curtius describes Poros' chariots in 326 as four-horse vehicles with the unusually large nurnber of6 crew two archers, two drivers armed with javelins, and two shield·bearers to protect the others. This secms an impracticably large crew even fora big four-horse chariot; and while Indian art often shows tbree or four men in a four-horse chariot, I know of no confìrmation for six. Perbaps Curtius or his sources have misinterpreted thc two escorting infantry of carlier sources (who could wcll be described as protecting the crew) as shield-bearing crewmen.

In any case, ifthe six-man chariot was used it seems to have dropped fairly quickly out ofuse, since Mcgasthenes describes Chandragupta Maurya's chariots as having one driver and two crcwmcn. He does not mention the number ofhorses, but four is most likely, being commonest in an ofthe 2nd cenrury BC up to the lst or 2nd AD; a relieffrom Bodh Gaya, lst cenrury BC, shows a 4-horse chariot with driver and two archers, which may be like Cbandragupta's. Although four horses were standard, two is stili occasionally seen; both two and four-horse chariots are depicted at Sanchi.

The bow was the commonest weapon used in the chariot, and provision ofquivcrs and arrows was one ofthe tasks of the superintendent of chariots in the Arthasasrra. Special large quivers were used. The Mahabharata also occasionally mentions javelins carried in chariots.

On the march the chariots were pulled by oxen, to spare the chariot horses, which were led in halters.

178.

LIBY AN CHARIOT

The Libyan tribes, the Greeks ofKyrene and the Carthaginians used, until the end ofthe 4th century or the carly 3rd, four-horse chariots; no doubt ali ofthem used the same basic design, Herodotos saying that the Greeks had actually learned the use offour-horse chariots from the Libyans. Only tbe nearest horse has been shown, for clarity, in this drawing. It is based largely on a Persian relieffrom Persepolis. Some crude Libyan rock-carvings, showing more or lessa pian view, confìrm that the chariot had rwo po Ics, running along the outside of the body; one carving shows two yokes, another shows one long yoke-bar attached to both poles and stretching across the shoulders of ali four horses. The Persian reliefs show what seem to be rcin-guidance rings attachcd to the yokes, and strengthening bars anaching the yoke-poles to the front ofthe chariot body. The L ibyan graffiti show the chariot's body about twice as wide as it is long; they represent it as a simple rectangle with a crossbar halfway along the long sides. This may represem the front view of the body, with a rcinforcing bar connecting the top and bottom frames. The signifìcance ofthe diagonal patterning on the body ofthe Perse polis chariot is not certain, unless it indicates that the body was lightly built of criss-cross diagonal leather straps on a wooden framework. The Persepolis sculptures shows heavy 12-spoked wheels with studded rims, but the Libyan graffid have lighter 4- or 6-spoked wheels. Since Persian chariots had similar heavy wheels, I suspect the Persian sculptor may have becn dcpicting the sort of chariot-wheel he is used to, rather than accurately copying, and so I have followed the Libyan evidence.

179

The Libyan graffiti show only one or two crewmen, with spcars, but Diodoros suggests that K yrcnean chariots may have had a crew of three, as Ophcllas' army in 308 had "a hundr ed chariots, and more than three hundred charioteers and men to fight beside them".

179.

CELTIC CHARIOT

This illustration is based chiefly on a grave stele from Padua. Coins show cbariots oftbe same generai shapc, various components have been excavated, and a Galatian chariot yoke is depicted on the Pergamene reliefs. Tbc Celtic chariot was a light two-horsed vehicle (again only onc has beea sbown for claricy) crewed by one noble warrior armed with javelins :md a retainer who dròve.

The chariot wheels were about 3 feet (90cm) in diameter, with iron tyres shrunk on. An example from the lake at La Tene has ten spokes; the Padua stele shows eight. The body was about three feet (90cm) across, and thc Padua stele suggests it was oaly about the sa me length. Some may ha ve been longer, as dead warriors were laid out at full length in their chariots in some burials. The driver sat or squatted at tbe front oftbe vebicle, with a whip. The warrior is shown standing behind him on one coin, beside him on the Padua stele. No doubt he would move freely about thc 180

vehicle depending where he wantcd to throw his javclins. Thc chariot sides were a doublc woodcn arch, panly covered in with lcathcr. Thc vehicle was panly dccoratcd with openwork bronze plaques. The horses were hamessed with a wooden yokc (the polc was reinforced with iron where the yoke was attached), girth and breast straps, and apparently with sidc traccs probably attached via long iron rods to the axle.

lt should be notcd that Lucian's reference to the Galatians using scythed chariots in 273, althougb very probably correct, does not mcan that thc normai Celtric chariot was scythed. Hc distinguishes between two-horsed and scythed chariots, implying that thc lauer do not have two horscs, but are the conventional four-horsed scythed chariots offigure 176. These must have been captured, or perhaps copied, from the Seleucids.

180.

ROMAN ANTI-ELEPHANT WAGON, 279 BC

300 of these curious devices were, according to Dionysios of Halikarnassos, used by Rome against Pyrrhos' elephants at Asculum. They were four-wheeled ox-drawn wagons {the reconstruction is based on a relief of a solid¡ whceled ltalian farm cart) with waulc screens to protect the crew. They were fincd witb upright poles, to which wcrc anacbed mobile borizontal bcams whkb could be swung in any dircction. Tbc beams werc fitted with "tridcnts or swordlikc spikcs or scythcs ali of iron ", wbile some bad grapnels wrapped in pitcb-daubed tow. Thesc would be set on firc and swung at thc clcphants' rrunks and faces. In addition, the wagons were manned by archers, slingers "shooting iron caltrops" and stone-throwers.

Thcrc are rwo accounts oftbcir performance in baule. Dionysios says they initially stopped tbc elepbants' charge, but were then shot at witb javelins by thc elcphant crews and overwbelmed by supporting ligbt infantry, wbo cut through the wattle screens to get at the crew and hamstrung the oxen; the wagon crews then took fligbt. Zonaras however says the wagons were nevcr cngagcd, bccause the elepbants auackcd at tbc othcr end ofthc banlcficld thc wagons no doubt being too clumsy to rcdeploy. As Asculum was a two-day baule, both stories may be correct, cach rcfcrring to one day of tbc baule. Thc wagons' very cxistence has been doubted, as the invention of later Roman annalists; but they are described as such a dismal failure that they hardly reflect well on Rome, and serve no Roman propaganda purpose. They probably represent a real and ingcnious, though ultimately unsuccessful, attempt to dcal with a military problcm new to Rome.

181

181.

INDIAN WAR-E LEPHANT

Elephancs were an essenrial part oflndian armi es by thc 6th cenrury BC. 60-ycar-old elcphants wcre most prized for battle because oftheir experience, but must have been rare as the normai lifespan was 60-80.

Elephants had cloth caparisons, hauhatthara, often heavily embroidered, or padded coverings, parisrara, which gave a more secure seat for the crew, secured with straps or ropes. Riders sometimes had the bene lit of ropes like the next figure's to help keep them on. Bells were hung on the neckrope and clscwhere. Harness could be decorated witb gold. One of thc late books of the Mahabharara mentions elephant armour of oxhide and serpcntskin, while a cbronicle mentions a king ofCeylon in the 2nd cemury BC wbose clephant had sevenfold armour ofbuffalo hide.

Tbc elephams thcmsclves scem somctimes to have been paimed, with leaves and Oowers, but this may not have been common.

Crew was a driver and up to three warriors. Thc driver could be armed; Poros' driver, on rhe coin showing Poros and Alexander, has javelins. Poros seems to have been alone on his mount, except for rhe driver, and other

sculprures show solitary warriors. When a second warrior appears in an be is often a parasol- or standard-bearer; these were probably armed. Tbis elephant from Sanchi, with archer and standard, is rypical. However, Megasthenes mentions Mauryan elephants with driver and three archers, while Scullard reports a day model of Mau ryan date with four riders. Occasionally Indian sources menĂšoo more than four riders; the Mahabharaca once mentions seven riders, two with hooks, rwo with bows, rwo wilh swords, one with spear and banner. This, if not mere rhetorical fancy, may derive from confusing cscorting infantry with the crew. Certainly escorting infantry were used; one Buddhist source says an elephant has 12 men, which an ancient commenmor explains as four riders and eight infantry "two looking after each foot". The weapons of escort infantry are described as bows, spears, javelins, axes, maces, clubs and swords.

T he weapon most often used from elephant-back was the bow, though javelins wcre also used - Poros fought with javelins at the Hydaspes, while the Mahabharata has one king sbowering both arrows and javelins on lhe enemy.

The archer illusuated has a shorter bow than usua l, perhaps a composite weapon. (His headgear is al so notewonhy; this must be what Nearchos mcans when he describes "an uppcr garment which they throw panly over their

183

shoulders and partly twisc in folds around the head" .) A slightly enigmatic wcapon ascribed to elephanr crcws is che tomara. Basham regards chis as a long spear, long enough co allow an elephant ride r 10 reach his enemy; but the Archasasrra defines it as an arrow, with a variant thrown by hand, while a commenta tor on thc Mahabharara, where the term also appears, calls ic a javelin. This seems the likely explanation, the confusion wilh a long spear arising because it was used co reach the enemy at a distance - by being thrown.

182.

MACEDONlAN WAR ELEPHANT

This is the equipment Alexander introduced for his elephant corps; Diodoros describes the elephanrs shown on Alexander's funeral car, each with an lndian driver and a fully¡armed M acedonian seated appareotly on the beast's back. T he rider is shown supponed by a rope as sometimes seen in lndian art, and has left offhis greaves and sandals 184

fora better grip. This cquipmcnt probably stayed in use till the introduction oftowers carly in the 3rd ccntury, though it is not impossible that clephams occasionally weot into battle ridden only by thcir drivcrs.

183 and 184.

ELEPHANTS WITH TOWERS

Plutarch's L1fe o/ Eumenes rcfcrs to towcrs at the battle of Paraitakene, but H. H. Scullard has suggested this is embroidery by a latcr writer. Certainly whcn Eumenes' elephants were ambushed by Antigonos a fcw months later they wcre unable to reply to cnemy missiles, which suggests a lack of missilc-armed crew. The first rcliable referenccs to towcrs are to Pyrrhos' campaign in ltaly, and he may have invented them. Early towers held two javelinmen. These are shown by a Campanian piace depicting one of Pyrrhos' elephants, by a Graeco-Bactrian silver disc on which figure 183 is based, and confirmed for Pyrrhos by D ionysios of Halikarnassos' account of Asculum. Towcrs were wooden, the Campanian plate suggesting light slats presumably over a heavier frame, and showing also a large round sh.ield hung outside for added protection. Cloth caparisons were often red, and this figure's seems to have the centrai portion quihed to cushion the tower. The beli round the neck is common, and Arrian says the tusks were reinforced with iron.

184 shows a later Seleucid elephant with armour and a larger tower. lt is based on a terracotta statuette of one of Antiochos l 's bcasts at the "Elephant Victory", a broken bronze of an armoured elephant, and literary references.

The towers must have been larger as they now held three (Ailian) or even four (L ivy, describing Magnesia) men.

Polybios mentions sarissa-armed crew at Raphia, Josephos mentions archers. Again the tower has shields hung outside, two each side, which may indicate a crew of four. Supporting straps are not visible on tbc terracotta, and 185

185

I

/j

~

),

may thus have passed under the cloth hangings. The hoops round thc legs, shown on the terracotta, are probably to prevent hamstringing by enemy infaocry, and may be leacher. The bronze statuette shows the piace headpiece and scale body-armour. Livy also mentions headpieces, while I Maccabees mentions "cuirasses" for Seleucid elephants.

Livy al so adds crests; no illustration shows them, and the style shown here is hypothetical, based on chac of figure 49's horse.

185.

AFRICAN ELEPHANT

Unlike che previous illustrations, al! ofthe lndian spccics of elcphant, this is ofthe small African "forest" species, Loxodonta Africana cyclotis. This, or perhaps a now extinct race of similar size, was hunted in modem E ricrea and Somalia and in Numidia, and used in war first by che Pcolemies, chcn by Canhage and the Numidians. A Pcolemaic inscript ion differemiates becween Echiopian and Trogodycic elephants, but the distinction may refer only to che area wherc they wcrc humed, not 10 any difference of race. The Afric:m forest elephant is 7-8 feet (2. l 5-2.45m) tal! at thc shoulder, shorter than the Indian elephant at up to IO fcet (3. lm), and much smaller than the great African bush elephant, not used in war, which can be up to 13 fcct (4m). It was an ancient commonplace chat African elephants were smaller and weaker chan Indians, and could not stand up to them in batcle. Differences berween the species include the African's more scrongly segmcnted or ridged trunk, ending in rwo "fingers" rather tban one, and che line ofics back is concave or flat, while the Indian's is convex. The small forest elepbanc also bas rounded ears and small, straight tusks.

186

186

Although smaller, the African elephant was strong enough to bear the weight of towers, crcw and armour.

Ptolemaic beasts at Raphia carried sarissa-armcd men in rowers, while king Juba's Numidian elephants werc both turreted and armoured. H. H . Scullard has suggested that Carthaginian elephants did not use towers, as there are no reliable 1extual references to them, while the Ptolemies a1 Raphia nceded crewmen specifically to protcc1 the elephant and driver from Seleucid elephant crewmen, but the Canhaginians had no such special need. Bui agains1 1his, severa! works of art do show Hannibal's elephants towered, including 1hc s1arucne on which this figure is based; while the Numidians, who similarly had no specific pressing need for towers or crcwmen, did use them, and must have learned 1heir elephant techniques from Carthage. They are likcly to have becn following Cartbaginian practice in giving their elephants towcrs, just as tbc Carthaginians lcarned to use elcphants from Pyrrhos' 1urreted examples. Some la1er wrinen sources also mention Carthaginian towers, though this may be mere literary commonplace. True, occasional coins show Punic elepbants witb driver only, but thcsc clephants are not shown in baule, andare perhaps on the marcb. On ba lance, the plentiful positive artistic evidence seems to ourweigh thc lack of reliablc litcrary confirmation, and it seems to me highly probable that Carthaginian elepbants did in fact use towers, perhaps containing two javelinmen, as Pyrrhos' had done.

The figure shows a Negro mahout in Numidian¡style cunic. Hasdrubal's mahouts at the Metaurus carried a mallct and chisel, with which they could kill che elephant, by swift blow to the base of thc skuU, if it went amok and auacked its own men. This was a relatively recent innovation, introduced by Hasdrubal himselfto counter one of the chief dangers of using elephants. The 1ower may be ofhides stitched over a wooden framework, and is hcld on by chains. The centrai one seems 10 pass through 1he lower saddleclo1h. One carved gem shows an open railed pla1form on which two javelinmen stand, pcrhaps a lighter alternative to the towcr.

186.

EARLY CATAPULT

Greek anillery was dcvcloped from 1he gastraphetes, "belly-bow", a crossbow-like wcapon which uscd a largc composite bow too powcrful to be drawn by hand, and was invented by the engincers ofDionysios I ofSyracuse about 399 BC. As larger versions of the wcapon were buil1, they were mounted on s1ands and provided with a windlass to draw 1he bow. Such catapuhs (from the Greek karapelres, "shield-piercer", used as a generic term forali artillery) were used in the 4th cencury to shoot both bolts and stone shot. The bolt- or dart-shooting engine was termed oxybeles, 1hc s1onc-thrower lithobolos or petrobolos, or in Latin ballista, which term was at this time only used for stone-throwers. This illustration is ofa bolt-shoo1ing "mountaingastraphetes" (stili so called, though no longer cocked by pushing with the gunner's stomach) designed by one Zopyros of Taras in the mid-4th century. The drawing (at !cm to I foot, or about 1:30) is based on E. W. Marsden's invaluable reconstruction.

The power is stili provided by a large composite bow, in this case 7 feet (2.1 m) long. The bolt is placed in the groove of a sliding beam fitted into the stock. Once the ca1apult has been fired, the slider is pushed forward till the claw of the uigger mcchanism, which is mounted on the rear of the slider, engages the bowstring. The slider is then winched back, thus drawing tbe bow ready for the next shot. We are not told what sized bolts this engine used, but 187

187

a

sa

comparison with another gastrapheces designed by Zopyros, which had a 9 foot (2. 7m) bow and shot rwo 6 foot ( l .8m) bolts, suggests this one may have used bolts about 3 cubits (54 inches/ l .37m) long. The stock ofthe machine is attached to the base by a universal joint, which allowed it to be elevated, depressed and craversed.

Stone-throwing engines of this type were generally larger than che bolt-shooters. One designed by Charon of Magnesia had a 9 foot (2. 7m) bow and probably shot stones of5-6 mina (perhaps 5-6 lb or 2.3-2. 7kg); a larger engine built by Isidoros ofAbydos, with a 15 foot (4.6m) bow, may have shot stones ofabout 40 lb(18kg). lt was early stonethrowers like this, no doubt relatively small and mobile models like Charon's catapult, which Onomarchos the Phokian used in the field in 353.

187.

BOLT-SHOOTING TORSION CATAPULT

The next major step forward in artillery construction, probably the work of Philip ofMacedon's engineers in t he 340's, was to replace the giant composjte bows by using the torsion oftwo large venical springs made ofstrands of sinew-rope (or sometimes ropes ofhair). These were set in wooden frames and tightened by short iron levers at top and bottom, while the arms ofthe catapult were thrust through the springs. As the arms were winched back the sinew twisted, springing sharply back into their originai position when the trigger was released. In the mid-3rd century, Ptolemaic engineers at Alexandria developed a set of calibrating formulae which set out the ideaJ proportions ofthe various parts ofthe catapult, and the most elfective dimensions fora catapult shooting a given length of bolt or weight of shot.

The illustration shows a light bolt-shooting torsion catapult or "scorpion ", according to the formulae set out in the manual of Philon ofByuntium, about 200 BC. It di!Ters slightly from Marsden's reconstructions, and Philon's specifications, by including the iron framework of a late Hellenistic scorpion found at Ampurias (Emporiai) in Spain. lt aJso has curved arms, which allowed the sinew to be twisted through a somewhat greater are, thus developing more torsion and more power. The first torsion catapults, and indeed those of Philon's time, had straight arms, but the curved variety were in use by the early 2nd cenrury as they are shown on the Pergamon reliefs.

This figure is to the same scale as the last, but represents a three-span engine, one ofthe most popular types by the end ofthe 4th century. lts bolts are three spans (27n /69cm) long, halfthe size ofthe last figure's. In generai, smaJler engines were now much more popular than before. 2 1h-span engines were aJso used, as were 2-cubit (36" /90cm) types, and 3-cubit catapults with bolts the same length as figure 186's. These would have all their dimensions, except probably those ofthe stand, twice those ofthis figure. The bolts (187a) had a three-finned bronze head and were usually flighted.

188

/

188

189

Torsion catapuhs wcre more complicated to buih and maintain than the early giant bow-powered types, but had more power and range. The grea test range claimed for an ancient catapult was 800 yards/730m with a 4-cubit engine, whilc a three·span shot 700 yards/640m. These were probably the absolute maximum ranges of unusually powerful engines, and normai effective range was probably about 400 yards/365 m, the popular 3-span catapuh baving the best range ofthe common types. lt is clear that the catapults were ve ry accurate, and Mardsen, from experience of working a full-sized replica, suggested aimed shots could be fired at a rate ofat least one a minute. The bohs would pierce most sbields and armour, though in 304 a Cypriot armourer produced two catapult-proof iron cuirasses.

Experimental and largely unsuccessful modifications to thc basic design includcd a bronze·spring-powered cnginc, another worked by comprcssed air, and a repeating catapuh. This was built in Rhodes (perhaps for Demetrios' siege in 304) by Dionysios of Alexandria. This had a hopper-type magazine ofshon (25 daktyls, about 19" /48cm) bolts which dropped one by one into piace; the gunner turned a windlass which, by a chain mechanism, successively pushed forward the slider, pulled back slider and string, and releascd thc string. lt was similar to the much later Chinese rcpcating crossbow. The repeater was not generally adopted because it was coo consistent, wasting a wholc magazine of bolts on one target (with such precision that a l 9th-century replica split one of its own bohs witb a subsequent shot) and had a restricted range, only abou t 200 yards/180m.

188.

STO NE-THROWlNG TORSION CAT APULT

Not long after the development of the torsion principle, probably in the 330's, it was applied to larger stonethrowing engines. To make tbem sufficiently powerful, the two sinew springs were set at an angle to cach other .

This increascd the are through which thc arms wcrc drawn, increasing the power. This illustration shows the popular 30-mina size, shootinga stone ball 30 mina (28lb 13oz/13kg) in weigbt. lt is at a scale of5mm to tbe foot or about I :60; as can be scen, stone-throwing engines tended 10 be very mucb bigger than bolt-shooters. Tbe smallest popular size was tbc 15-mina, whicb would bave ali its dimensions 24% smaller than tbis figure (tbougb sbot as small as 5 mina bave been found). Tbe 30-mùra was popular for use against personnel, enemy cngines or otber siege macbinery, or palisades. Tbe best size for use against stone walls was the 1-talent cngine, wbose sho1 weigbed 571b I Ooz/26kg. This bad an effcctive range of about 150 yards/ 135m agains1 walls, or 400 yards/365m against ligbter targets. lts dimensions, except probably tbosc ofchc stand, wcrc 25% greater than tbe 30-mina. Tbe largcst engine in generai use was che 3-talent, wbich would be 75% larger than the 30-mina.

The large engines could double as bolt-shoo1ers, if the special bowstring, fitted with a sling for 1bc s1one, was replaced by a conve ntional string. A 3-talent engine designed by Arcbimedes fora Syracusan sbip could also shoot a 12-cubit bolt - 18 feet/5.5m long, tbe size of a sarissa! Alternative ammunition includcd earthenwarc pois of poisonous snakes; the exiled Hannibal, serving tbc king ofBitbynia as an admiral, used these against enemy sbips; the pot would sbatter on hiuing thc deck, :md thc co ntents proceed to distract tbe crew.

The one-armed stone-tbrower later known to tbe Romans as tbc onager is mentioned in passing by Philon about 200 BC, but did not come into use on any scale till tbc 3rd century AD. lndian sources not infrequently mention anillcry, ascribing it to armies ofour era and well beforc. Sucb engines are usually calledyamras, an unfortunately imprecise cerm whicb it seems can also be used to mean weapons in generai.

Yamras whicb are cenainly artillery are usually used in tbc defence offonifications; tbey burl stones or arrows, and make a noise Like tbe twang of a bowstring, suggcsting thai tbey are bow·like in form. It is noi impossible 1bat the Indians could have invented artillery independently - after ali the Greeks and Chinesc bo1b sccm 10 bave invented 1hc crossbow sepa rately from each other, and dcvcloped artillery from it - bui it does noi seem likely. h is hard to be cenain, but I 1hink i1 mosl likely 1ba1 many of1he references to artillery in lndian tex1s are anachronistic, and 01hers mere li1erary commonplace; 1be lack of de1ailed, ratber than rbetorical, dcscrip1ions does not inspire confidence. What anillery the lndians did have was probably copied from tbc Greeks, and may in many cases bave becn built and operated by Grcek anisans.

SELECT BIBLIOGRAPHY Primary sources h bas only been possible 10 list 1he most important works. (P) indicates a translation is available in Penguin Classics, (L) tbat parallel tex1 and translation is available in Loeb Classical Library. Sources on artillery will be found in Marsden ( 1971 ).

190

In Greek:

Appian, Ro111a11 History (L) Arrian, Anabasis o/ Alexa11der(P, L); lndike (L); Tacrics Asklepiodotos, Tactics (L, bound with 1he works of Aineias and Onasander) Dionysios of Halikarnassos, Ro111a11 Amiquities (L) Diodoros Siculus, Hisrory(L) I and Il Maccabees Pausanias, Guide w Greece (P, 2 volumes) Polyainos, Stratagems Plutarch, Lives (some P; all L) S1rabo, Geography (L) Xenophon, A11abasis (P, L); Hellenica (P, L); Ky ropaidia (L); 011 '1orse111a11S!iip and Tlie cavalry commander (L, 1oge1her)

In Lario:

Quimus Curtius, History o/ Alexa11der (L) Livy, History o/ Rome /rom its /01111da1io11 (L, P) Silius halicus, Punica

01hers:

Tlie Dead Sea scrolls in E11glis'1, G. Vermes, Penguin, (Harmondswonh, I 962) has 1be Rute o/ War Kawilya Artliasasrra, trans. E. P . Kangle (Bombay, 1972) Maliabharata, trans. R. Dun (London 1972)

Modero works Once again only 1he mosc valuable works have been lis1cd; 1he literacure on che subjecc is vast. Abbreviacions for periodical 1illcs are: AJA - American Journal of Archacology; BCH - Bulle1in de Corrcspondencc Hellenique; ]DAI - jahrbucb des deu1sches archaologiscbes Insci1u1; JHS - journal of Hellenic Studies; PBSR Proceedings ofthe Bricish School a1 Rome; PCPS - Proceedings ofthe Cambridge Philological Society; RA Revue Arcbeologiquc; SCI - Scripta Classica Israelica.

Anderson, J. K . A11ciem Greek Horsemamliip (Los Angeles, I 961) - Military tlieory a11d practice i11 tlie age o/ Xe11op'1011 (Los Angeles, 1970) Andronikos, Manolis. "Sarissa'', BCH 94, 1970, p9 I Arribas, An1onio. The lberians (London, 1963) Bar-Kochva, B. " T bc sources and chronology of Antiochos l 's bacĂše agafost the Galatians", PCPS NS 19 (1973) pi - " Hellenistic warfare in Jonatbon's campaigns near Azocos" SCI II, 1975, p83 - Tlie Seleucid Army (Cambridge I 976) Basham, A. L. Tlie Wonder t!iat was India (London, revised ed. 1967) Beli, M . J. V. " Taccical reform in the Roman republican army'', Historia 1965, p404 Bengtson, Herman, et al. T/ie Greeks and 1/1e Persians (London, 1969) Berthicr, A. and Charlier, R. l e sa11c1uaire punique d'el Hofra a Co11s1a111i11e (Paris, I955) Besc, J. G. P. Tliracian peltasts a11d tlieir influence 011 Greek warfare (Groningcn, 1969) Brown, F. C. "A recently discovered composite bow", Seminari1m1 Ko11dakovia1111m 9, I 937, p I Brun de HofTmayer, Ada. " Arms and armour in Spaio - a shon survey; voi. I", Gladius, tomo especi al, 1971 Brunt, P. A. "Alexander's Macedonian cavalry", JHS 83 ( 1963), p27 Cambridge History o/ India: Volume I, Anciem India (Cambridge, 1922) Cary, M . A History o/ tlie Greek world, 323 to 146 BC (London, revised ed. 195 I) Caven, Brian. The Pu11ic W'ars (London, I 980) Cawkwcll, George. Philip o/ Macedon (London 1970) Connolly, Pecer. Greece a11d Rome at War (London 1981) Couissin, Paul. Les armes ro111ai11es (Paris, I 926) - Les i11stitutio11s militaires et navales (Paris 1932) Criscofani, Mario. Tlie Etrusca/IS - a 11ew i11vestigatio11 (London, I 979) Ellis, j. R. Pliilip Il a11d Macedo11ia11 lmperialism (London, 1976) Ferguson, William Scoct. Hel/enistic Athens (New York, I 969) Feyel, M . "Un nouveau fragment de reglemem mili1aire rrouve a Amphipolis", RA (I 935) p29 - Polybe et l'histoire de la Beotie au liieme siede (Paris, 1942) Forresc, \V. G. A history o/ Sparta, 950-192 BC(London, I 968) F rederiksen, M artin \VI. "Campanian cavalry: a quescion of origins", Dialoghi di arcl1aeo/ogia, 2, 1968, p3 Galili, El'azar. " Raphia, 217 BCE, revisiced", SCI, III, 1976, p52 Garoufalias, Petros. Pyrrhus, king o/ Epirus (English ed. London, 1979) 191

von Graeve, Volkmaar. Der Alexandersarkophag 1md seine werksratt (Berlin, 1970) Griffith, G. T. The mercenaries o/ rhe He/le11is1ic World (Cambridge, 1935) - "Makedonika: notes on the Macedonians ofPhilip and Alexander", PCPS, NS.4 (1956) p3 - "A note on che hipparchies of Alexander'', JHS 83 (1963) p69 Hammond, N. G. L. Problems in Greek hisrory(Oxford, 1973) - A history o/ Greece 10 322 BC (Oxford, 1959) - "The banle ofche Granicus river",JHS 100 (1980) p73 Hatzopoulos, M ., and Loukopoulos, L. (eds). Philip o/ Macedon (Athens, 1980) H odinotl, R. F. Bulgaria in antiquiry (London, 1975) - The Thracians (London, 1981) Lane Fox, Robin. Alexander rhe Grea1 (London, 1973) Lazenby,J. F. Ha1111iba/'s War(Warminstcr, 1978) Lcsquier, Jeao. Les insrirurions milĂŹ1aires de /'Egypre sous /es lagides (Paris, 1911) Mahaffy, J. P. "The army of Ptolemy IV at Raphia", Hermarhena XXIV (1898) pl40 Marklc, M . M. "Tbe Macedonian sarissa, spear, aod related armor", AJA 81, 1977, p323 - "Use oftbe sarissa by Philip and Alcxander of Macedon", AJA 82, 1978, p483 Marsden, F. W. The campaign o/ Gaugame/a (Liverpool, 1964) - Greek and Roman ar1illery, two volumes, His1orical developmem and Technical 1rea1ises (Oxford, 1969 & 1971) Mellink, M. J. " Excavations at Karatas-Semayuk aod Elmali, Lycia", AJA 76 (1972) p257 Merker, I. L. "The ancient kingdom of Paionia", Balka11 Srudies VI, (1965) p35 Milns, R. D. "Alexander's Macedonian cavalry and DS XVU.14.4", JHS 86 (1966) pl66 - " Philip II and the bypaspists", Hisroria 1967, p509 - "The hypaspists of Alexander lll", His1oria 197 I, p 186 Narain, A. K. The Indo-Greeks (Oxford, I 957) Nicolini, Gerard. The anciem Spaniards (London, 1974; originally Les Iberes, Paris, 1973) Olmstcad, A. T. Hisrory o/ rhe Persian Empire (Chicago, 1948) Pallottino, Massimo. The E1ruscans (Loodon, 1974) Parke, H. W. Greek mercenary soldiers /rom rhe earliesr rimes 10 the battle o/ Ipsos (Oxford, 1933) Pcricot Garcia, L. The Balearic Islands (London, 1972) Philips, E. D. The royal hordes (London, 1965) Picard, G. Carthage (London, 1964) Powell, T. G. E. Tlre Celts (London, revised ed. 1980) Rawson, Elizabeth. " The literary evidence for the pre-Marian army", PBSR 39(1971)p13 Robertson, David M. Excavariom al Olynthus, voi. X (Baltimore, 1941) Robinson, H. Russell. The armour o/ Imperia/ Rome (London, 1975) Rostovtzeff, Michael. Iranians and Greeks in sourh Russia (Oxford, 1922) - "Dura and tbc problcm of Parthian art", Ya/e Classica/ S111dies V (1935) p 157 - Socia/ and Economie hisrory of 1he Htllenisric World (Oxford, 1941) Rumpf, Andrcas. Kranos Boio1ourges (Berlio, 1943) Salmon, E. T. Samnium and 1he Sam11i1es (Cambridge, 1967) Sandars, Horace. "The weapons of the Iberians", Archaeologia XIV ( 1913) p205 Schmidt, Ericb F. Persepolis (3 vols, Chicago, 1953, 1957, 1970) Schroeder, B. " T hrakische Helme", ]DAI XXVII (1912) p3 l 7 Scullard, H. H. A hisrory o/ the Roma11 worldfrom 753 ro 146 BC (Loodon, 1935) - The elepham in the Greek and Roman world (London, 1974) Singh, Sarva Darman. Anciem lndian warfare wirh special reference to the Vedic period (Leiden, 1965) Snodgrass, A. M. Anns and armour o/ che Greeks (London, 1967) Stipccvic, Aleksandar. The lllyriam - hisrory and culrure (Park Ridge, NJ, 1974) Sulimirski, T. The Sarmatians (London, 1970) Talbot Rice, Tamara. The Scyrhiam (London, 1957) Tarn, W. W.A11rigonosG011a1as(Oxford, 1913) - Hellenisric naval and milirary developmenrs (Cambridge, 1930) - The Greeks in Bacrria and India (Cambridge, 1938) Trendall, A. D. The red-figured vases o/ Lucania, Campania and Sicily (Oxford, 2 vols, 1971) - The red-/igured vases o/ Apulia (Oxford, 1978) Walbaok, F. W. Philip V o/ Macedo11 (Cambridge, 1940) Warry, John. Warfare in rhe Classica/ World (London, 1980) Weege, F. "Oskische Grabmalerei" aod "Bewaffnung und tracht der Osker",JDAJXXIV, (1909) p99 and pl41 Woodcock, George. The Greeks in India (London, 1966) Zhivkova, Lyudrnila. The Kazanluk Tomb ( 1975) 192

MISCELLANEOUS INFORMATION Many readers will be collectors, modellers, or wargamers. There are too many fìrms :md individuals producing models and fìgures in ali sizes for us to list herc, and we recommend studying the advertisements in Milicary Modelling to gct a good idea ofthe ranges available.

This magazine is momhly, on generai sale, or direct from:

Milicary Modelling, P.0. Box 35, Bridge Street, Hemel Hempstead, Herts. HP I I EE. The Society of Ancients An international Society forali interested in Ancient and Medieval warfare, weapons and warriors. Produces a bi· monthly magazine, Slingslzot, for members, containing articles on military history, arms and equipment, wargame reports and book and figure reviews. The Society sponsors national meetings for members and runs an annua! wargames championship. Whatever else you do, if you are interested in funhering your knowledge and enjoyment, join the Society.

For details write to the Treasurer: W. Thurlow, 'Courtney', 15 LongOeet Road, Poole, Dorset.

Wargames Resear ch Gr oup This is the latest in our Armies and E11emies series covcring from che earliest ùmes up to 1500 A.D. Funher ciel es are in preparation, and cvenrually we hope to cover up 10 the present day.

Fora full lise ofthese and Wargamcs Rules in most periods send a stamped addressed cnvclope or Intcrnaùonal Reply Coupons to: WARGAMES RESEARCH GROUP, 75 ARDINGLY DRIVE, GORING-BY-SEA, SUSSEX BN 12 4TN. U.S.A. and Aumalia: The following bave scocks ofour producùons:

The Courier, Box 1878, Brockton, MA. 02401, U.S.A. Baulefield, 50 Clissold Parade, Campsie, N .S.W. 2194, Australia.