The following abbreviations are used:

Arr. = Arrian, Anabasis Alexandri. Atkinson=J. E. Atkinson, A commentary on Q. Curtius Rufus' Historiae Alexandri Magni, Books 3 and 4, i (Amsterdam 1980) Bosworth = A. B. Bosworth, A historical commentary on Arrian's History of Alexander i (Oxford 1980). Brunt L = P. A. Brunt, Arrian i (London 1976), ii (1983) in the Loeb edition. Gomme = A. W. Gomme, A historical commentary on Thucydides (Oxford 1945-56). Hammond A = N . G. L. Hammond, Alexander the Great: king, commander and statesman (New Jersey 1980). Hammond THA = N. G. L. Hammond, Three historians of Alexander the Great: the so-called Vulgate Authors, Diodorus, Justin and Curtius (Cambridge 1983). HM=A history of Macedonia i by N. G. L. Hammond (Oxford 1972), ii by Hammond and G. T. Griffith (1979), iii by Hammond and F. W. Walbank (1988). Pritchett= W. K. Pritchett, The Greek state at war iv (Berkeley 1985).

¹ τοῖς δημοσίᾳ θαπτομένος.

² See Gomme ii 100.

³ Each tribe probably supplied the state official with a list of its dead; see IG i³ 1141 and Bradeen, D. W. in The Athenian Agora xvii (1974). 73Google Scholar. The list was like our regimental roll of honour. The tribal officials had to find replacements for the tribal regiments.

⁴ Despite some scholarly reluctance; but cf. e.g. Gomme iii 656, ‘the figure may be believed.’

⁵ Rounded, it seems, for stylistic reasons. As Gomme ibid. remarks, the figures are ‘startling.’

⁶ Summarised in Pritchett 139 ff., and to be taken ‘seriously’ as evidence (Krentz, P., GRBS xxvi [1985] 13Google Scholar).

⁷ Pritchett 244; Sparta kept her military arrangements as confidential as possible (Thuc. v 68.2).

⁸ Curt, v 4.3.

⁹ See Pritchett 138 for a short summary and full references to the excavation reports.

¹⁰ Brunt L i 68 n. 3 held that Aristobulus' 9 infantrymen in contrast to Ptolemy's 30 infantrymen may have arisen because Aristobulus did not include non-Macedonian infantry. This is implausible, because Alexander deployed very few non-Macedonian infantry (only 1,000 or so Agrianians and Archers as opposed to 12,000 Macedonian infantry). See Brunt i lxx and Hammond, JHS c (1980) 82Google Scholar.

¹¹ The context is important. Arrian reports at iii 15.5 that Alexander rested ‘the cavalrymen with him' (τούς ἀμφ᾽ αὑτὸν ἱππέας) and then continued the pursuit and captured the treasure at Arbela. Arrian then records the effect of the long pursuit in the losses in men and horses ‘of those with Alexander’ (iii 15.6, ἀπέθανον δὲ τῶν ἀμφ᾽ Ἀλέξανδρον). These are obviously the same group of cavalry. Arrian was not giving the total loss in the battle, as Brunt L i 273 n. 3 apparently and Bosworth i 312 emphatically believed.

¹² The manuscript reading is here to be retained; so Atkinson 243, whereas the addition made by Hedicke and accepted in the Loeb text (ed. Rolfe i 136) yielding 4,500 wounded is inconsistent with the next sentence's ‘at so small a cost’.

¹³ Brunt L i 12 n. 2 ‘Pt. is perhaps cited because A felt that such precise figures needed justification’ is more consonant with modern doubts. Arrian gave very many precise figures without citing his source. He did so here in accordance with his stated practice in the Proem, that where Ptolemy and Aristobulus differed he would choose the more trustworthy. It is likely that he intended his reader to deduce that he would be following Ptolemy in other statements of Macedonian casualties.

¹⁴ So too Brunt L i 68 n. 3.

¹⁵ Brunt L i 13 did not translate the αὐτῶν which is emphatic by position: αὐτῶν δἐ Μακεδόνων. The implicit contrast is with the non-Macedonians, i.e. the archers and slingers of i 2.4 and the light cavalry of i 2.6.

¹⁶ No one has yet doubted the existence of a King's Journal of Philip II and of Alexander. The modern discussion is whether the Journal cited by Alexander-historians was a genuine Journal or a forged version. If the Journal they used was a forged one, it is difficult to see why a forger would produce such detailed figures. My arguments for believing that they used a genuine Journal are presented in Historia xxxvi (1987) 1–21Google Scholar.

¹⁷ See Hammond THA 22–7. If Alexander's Journal accompanied his corpse, it passed into the control of Ptolemy, who kept it to himself, it seems, during his lifetime. It then may have gone into the Library at Alexandria. For a possible fragment of Strattis' Commentary on the Journal of Alexander see my article in GRBS xxviii (1987) 331–47Google Scholar. Ptolemy II kept a record of his forces in the basilikai anagraphai (App. Preface 10).

¹⁸ This applies to any period. In the Second and the Third Macedonian Wars the Senate was evidently informed of the casualties by the commander in the field, so that it could provide replacements and if necessary reinforcements; see Livy xxxii 1.3, in supplementum, 9.1 and 6; 28.10, supplementum; xliii 11.10–12, supplementi at 12.2; 12.9; 44.1, supplementum; xliv 21.8.

¹⁹ G. T. Griffith in HM ii 472 assumed that the 150 wounded and the one man killed had been engaged in a preceding battle, but that is not what the context suggests—namely that they, like Philip, were casualties of the pursuit.

²⁰ Herodotus' ‘about 6,400’ at Marathon is well discussed by How and Wells in their Commentary, ii 114.

²¹ Hdt. ix 62.2–3 and 63.2, ἄνοπλοι, γυμνῆτες.

²² Well illustrated by Livy's remark at xxxi 39.10: ‘the Macedonian phalanx with its specially long pike places a defensive barrier (vallum) in front of the shields’.

²³ As the Alexander–mosaic and the Alexander Sarcophagus show (see Hammond A 282–3).

²⁴ Especially the single great wedge of infantry and cavalry at Gaugamela in Arr. iii 14.2; (not a set of little wedges as in Devine, A. M., Anc. World xiii [1986] 114Google Scholar Fig. 2).

²⁵ See my article in CQ xxviii (1978) 136–40Google Scholar.

²⁶ Following Ephorus, a reliable contemporary historian, as argued in CQ xxxi (1937) 79 fGoogle Scholar.

²⁷ Diod. xvi 35.4–5, probably following Demophilus (CQ xxxi 84 f).

²⁸ Brunt L i 163 translated ‘deep gully’ but the Greek says simply ‘gully’ (φάραγξ). Bosworth i 217 holds that Ptolemy was romancing. However, it must be allowed that Ptolemy and Arrian knew more of cavalry warfare and of panic in flight than we do, and it is to be noted that Callisthenes—not a combatant like Ptolemy but an observer—repeated the report that ‘the most of the Persians in their flight were destroyed in such hollows (κοιλώμασι)’, i.e. the gullies (Callisthenes in Plb. xii 20.4; cf. Diod. xvii 34.9)

²⁹ See my account in JHS xciv (1974) 66 ffGoogle Scholar. The account of Bosworth, published in Ancient Macedonian Studies in honor of Charles F. Edson (Thessaloniki 1981) and in Macedonia and Greece in late classical and early hellenistic times, edd. Barr-Sharrar, B. and Borza, E. N. (Washington 1982)Google Scholar, is not compatible with the terrain, of which Bosworth lacked personal knowledge. See HM iii 43 note.

³⁰ For the identification of the battlefield see Hammond A 106; gullies of the kind mentioned by Ptolemy and Callisthenes (n. 28 above) are shown on p. 276 f.

³¹ The figures given by Diodorus at xvii 14.1 for the action at Thebes in 335 BC were comparable: 500 ‘Macedonians’ (i.e of them and their allies) and over. 6,000 Thebans. The source of Diodorus was probably Cleitarchus (Hammond THA 15 ff.), a Greek contemporary writing for Greeks of the time, who may have obtained a reasonable estimate. Bosworth i 312 dismisses the 400 dead at Tyre as ‘fatuous’. He does not mention the figure at Thebes or discuss the source of Arrian.

³² The source was probably Cleitarchus (Hammond THA 96), who was interested in the contrast with Greek systems of enlistment.

³³ See n. 18 above. P. A. Brunt did not take these matters into consideration in Appendix 28 of his Italian manpower.

³⁴ See Bosworth in CQ xxiii (1973) 245 ffGoogle Scholar. for pezetairoi and asthetairoi, and for the view in the text Hammond A 26 f.

³⁵ Assuming that the 1,500 cavalry left with Antipater (Diod. xvii 17.5) were heavy and light in the same proportion as those taken by Alexander to Asia.

³⁶ The population figures of 1961 are given in HM i 16 f. and Hammond A 29 f.

³⁷ Diod. xvi 3.1 (weapons) and Curt, vii 1.32–4 (horses). Each cavalryman had one groom at least.

³⁸ Although first mentioned at Gaugamela in 331 BC, the Macedonian archers were presumably among the ‘archers and slingers’ of the Balkan campaign of 335 BC (Arr. i 1.12 and i 2.4).

³⁹ Allowing the total field army of 26,800 to serve for twenty-five years on average; see Hammond A 152 f.

⁴⁰ As they balanced the Cretan Archers in the order of battle at Gaugamela.

⁴¹ Defined as ‘Macedones’, distinct from Thracians and others, in both passages. Recruited by Amyntas (Diod. xvii 49.1) they were the men mentioned by Arrian, who omitted to give the numbers at iii 16.10–11. The source of Diodorus and Curtius here was probably Diyllus, an accurate if dull Hellenistic historian (Hammond THA 54 f. and 129 f.).

⁴² Curtius v 1.40 called the new arrivals an ‘incrementum’, ‘reinforcement’; for the further brigade see Milns, R. D. in GRBS vii (1961) 160Google Scholar, preferable to Bosworth i 320.

⁴³ The source was probably Aristobulus (Hammond THA 141).

⁴⁴ See Hammond A 240 and CQ xxviii (1978) 133Google Scholar with n. 21 for ‘the sons of the Hypaspists’.

⁴⁵ See Hammond THA 240 for the source being probably Diyllus.

⁴⁶ Bosworth, in JHS cvi (1986) n. 51Google Scholar calls the word ‘Macedonian’ here ‘a blanket designation’; but the next sentence shows that Diodorus was writing of ‘citizen soldiers’ (xviii 12.2) who alone were ‘Macedones’. The source was Hieronymus, a contemporary historian familiar with Macedonian affairs and trustworthy.

⁴⁷ This is clearly so in Diod. xvii 109.1 ἀπέλυσε τῆς στρατείας and xviii 4.1 τῶν ἀπολυθέντων τῆς στρατείας, since ἡ στρατεία is ‘the campaign’ and in particular a campaign abroad (L-S-J⁹ s.v.). Arrian used στρατιά, which sometimes has the meaning of στρατεία.

⁴⁸ The history of these two groups is given in my article in GRBS xxv (1984) 54 ffGoogle Scholar.

⁴⁹ This last passage has been misunderstood by Brunt L ii 489. It runs thus: (Craterus having come to Macedonia to help Antipater, besieged in Lamia) ἧγε.. πεʒοὺς μὲν τῶν εἰς Ἀσίαν Ἀλεξάνδρῳ συνδιαβεβηκότων ἑξακισχιλίους, τῶν δ' ἐν παρόδῳ προσειλημμένων τετρακισχιλίους, Πέρσας δέ…… The word προσειλημμένων is the standard term for ‘taking on’, i.e. enlisting, troops (used at xviii 14.5 of Leonnatus, προσελάβετο), and ἐν παρόδῳ means ‘on the way’, ‘en route’, whether Craterus came through Thrace where Lysimachus was in trouble (xviii 14) or more probably by sea from Cilicia to Amphipolis (see HM iii 113). The passage was correctly translated by R. M. Geer in the Loeb edition as ‘four thousand from those who had been enlisted on the march.’ Brunt stated that the ‘6,000 were survivors of the original expeditionary force, and 4,000 of the men who had joined Al. later’. This is not correct; as the sentence is written Craterus has to be the agent with προσειλημμένων. Bosworth, in JHS cvi (1986) 8Google Scholar followed Brunt's interpretation of the passage. For the 6,000 veterans with Craterus and the 4,000 veterans left by him in Asia see my article in GRBS xxv (1984) 55 fGoogle Scholar.

⁵⁰ The extent of the Asian recruitment has sometimes been belittled; see my comments in JHS ciii (1983) 139 ffGoogle Scholar.

⁵¹ For example in JHS cvi 9 ‘few of the men Alexander took with him (to Asia) ever returned’ and ‘the reserves of Macedonian manpower in 323 were less than a half, probably nearer a third, of what they had been in 334’.

⁵² Brunt L ii 532 f. wrote that the speech ‘is an epideictic display by Arrian’. The only approximation to a disease was the eating of an explosive wheat—one of many miraculous events in distant India—as recounted by Theophr. HP viii 4.5 (accepted by Fox, R. Lane, Alexander the Great [London 1973] 364Google Scholar).

⁵³ The authority for the 1,500 horse is not given by Brunt; it probably stems from his misinterpretation, as I see it, of Diod. xviii 16.4 (see n. 49 above).

⁵⁴ For this phalanx see Hammond A 240; Bosworth, JHS cvi 4Google Scholar put at the same number the Macedonian element in what he saw as ‘a bizarre amalgam.’

⁵⁵ This assumption is rendered more unlikely by the fact that reinforcements from Europe after 330 BC were reported but did not in any case specify Macedonians (for instance, Diod. xvii 95.4 allied and mercenary troops, Curt, ix 3.21 Thracians and others, Arr. vii 23.1 cavalry of unstated nationality).

⁵⁶ Fatigue, even with miserable food, did not cause fatalities in my experience in guerrilla warfare in Greece in 1943—4.

⁵⁷ Bosworth, in JHS cvi 6Google Scholar argues from Polybius xii 19, citing the acount of Callisthenes FGrH 124 F 35, that the 5,000 infantry and the 800 cavalry ‘from Macedonia’ (ἐκ Μακεδονίας) represent ‘massive reinforcements of the Macedonian phalanx’, i.e. that they were Macedonians. As was pointed out in Hammond A 152, a work to which Bosworth does not refer, Macedonia was the mustering point for reinforcements raised in Europe, and this body of troops need not have included any Macedonians. The context of Callisthenes shows that he was dealing with the total force under Alexander's command and not just the Macedonian element.

⁵⁸ I am here in agreement with Bosworth, in JHS cvi 3Google Scholar, whose estimate was ‘3,000 at a maximum’, i.e. the equivalent of two phalanx brigades. It is interesting that Callisthenes' figure for those on other duties and so not available for the Battle of Issus was 3,000 infantry and 300 cavalry (Plb. xii 19.3); most of those 3,000 were probably mercenaries, but Antigonus is likely to have had some Macedonians too.

⁵⁹ When Alexander reorganised his forces after 331 BC, the Macedonian archers, the Lancers, the Paconian cavalry and the Thracian cavalry disappeared from the record. Suitable men among them may well have been promoted to serve in the Companion Cavalry units and in the phalanx and so have obtained citizen status as ‘Macedones’.

⁶⁰ Diod. xix 28.1 and my comments in CQ xxviii (1978) 133Google Scholar with n. 21.

⁶¹ The first group of fifty boys came out in the winter of 331–330 (Curt, v 1.42). It makes sense that a similar group came out each year, since we find the Pages guarding the corpse of Alexander in 323 (Curt, × 7.16).

⁶² I am here following Curtius × 2.8 (see CQ xxx [1980] 469 f.Google Scholar), of which the source was probably the factual and dependable Diyllus (Hammond, THA 158). Bosworth, in JHS cvi (n. 22)Google Scholar claimed that Curtius' figure for the Asian army [i.e. the 15,000 to be retained] ‘cannot comprise Macedonians alone’. This claim is inconsistent with the context of Curtius. For Curtius is giving the background to the mutiny, which was one of Macedonian soldiers only, both in Curtius and in Diod. xvii 109.2–3, Plut. Alex. 71.1–5 and Arr. vii 8.1. The contrast between Macedonians to be released from the campaign and Macedonians to stay was clearly made by Curtius and by Diodorus xvii 109.1–2 and Arrian vii 8.1 (keeping in Arrian's text the reading μένουσι, as I suggested in CQ xxx 470 and as adopted by Brunt L ii 224).

⁶³ Including the Pages who served as cavalry during and after their schooling.

⁶⁴ But the probabilities are strong enough to rule out the need to depart from the literary evidence and to postulate either the sending of large reinforcements of Macedonian citizen troops to Asia between 330 and 323 or a spectacular upgrading of such reinforcements before 330 and enormous casualties of citizen troops in Asia. In as far as there was a decline in the power of Macedonia, it was due not to the losses incurred by Alexander but to the civil war instigated by Antigonus in 321 and continuing even beyond his death in 301. For some of its effects see HM iii 187–92.