An earlier draft of this paper was read by Mr J. Boardman, Dr P. Cartledge, Dr J. K. Davies, Professor W. G. Forrest and Professor G. L. Huxley. I am grateful to them all for their helpful criticism, the more so because the version I showed them was far more long-winded than that which is printed here. Theirs is the credit for what I hope is an improved presentation and for the elimination of many of my mistakes; for all remaining errors the fault is mine alone. A version of what follows was read to the Hibernian Hellenists in November 1975; I am grateful to all those who took part in the subsequent discussion, and especially to Professor H. W. Parke.

Abbreviations Detienne, La Phalange: J.-P. Vernant (Ed.), Problèmes de la guerre en Grèce ancienne 119–42; Forrest, Emergence: W. G. Forrest, The Emergence of Greek Democracy; Greenhalgh: P. A. L. Greenhalgh, Early Greek Warfare; Lorimer: H. L. Lorimer, BSA xlii (1947) 76–138; Sealey: R. Sealey, California Studies in Classical Antiquity ii (1969) 247–69; Snodgrass: A. Snodgrass, JHS lxxxv (1965) 110–122; Snodgrass, EGAW: A. Snodgrass, Early Greek Armour and Weapons.

¹ This paper is primarily intended as a chronological and political study, and I do not therefore consider the question of why the phalanx was introduced. Cartledge argues elsewhere in this Journal (above, 23) that political considerations were part of the reason for the introduction of the phalanx; but I am not persuaded that they were relevant, and am prepared to believe that the new mode of fighting was developed for essentially military reasons. Cartledge views the adoption of the phalanx as a paradox; he rightly points out (following Gomme and others) that the mountainous terrain of Greece is ill-suited to phalanx warfare, and adduces political reasons to explain why this unsuitable method of fighting was invented. The paradox, however, is merely apparent. The phalanx was evidently superior to an aristocratic rabble (which will also have fought on a plain and not on rough ground); the invention can therefore be explained on purely military grounds. The paradox is not the adoption of the phalanx but its continued use; and Cartledge shows that it was social and political considerations which prevented, for a long time, the development of the light-armed forces which could have been an extremely effective answer to the phalanx (see also below, n. 49). I am much indebted to Paul Cartledge for allowing me to see his paper before it appeared in print.

² Cf. e.g. Nilsson, , Klio xxii (1928) 240–49Google Scholar; Andrewes, , Probouleusis 13–15Google Scholar; id., Tyrants esp. 34–8; Forrest, , Emergence 88–97Google Scholar.

³ The chronology for most of the tyrannies in question here is disputed, but this is not the place for a discussion. See below, 92–3 (Pheidon); Oost, , CP lxvii (1972) 16Google Scholar n. 26 with references (Cypselus: c. 657); Leahy, , Historia xvii (1968) 1–23Google Scholar (Orthagoras: c. 655). The date of Theagenes of Megara is not disputed but vague: he must have been tyrant before providing help to his son-in-law Cylon for his unsuccessful attempt on tyranny at Athens in, at the earliest, 636 (Cadoux, , JHS lxviii [1948] 91)Google Scholar.

⁴ Cf. especially Forrest, , Phoenix xvii (1963) 157–79CrossRefGoogle Scholar; see below, p. 93.

⁵ Cf. e.g. Sealey, esp. 249–50; Greenhalgh, 71–4, 150–5; Detienne, La Phalange; Pleket, , Talanta i (1969), 35–6Google Scholar; Zörner, , Kypselos und Pheidon von Argos, 104–7Google Scholar.

⁶ Snodgrass (p. 111) has refuted the old argument that the double grip shield implies (and therefore dates) the phalanx; cf. also EGAW 197. The old view, however, has not yet been abandoned. Greenhalgh 73 argues that the inability of the hoplite shield to protect the rear will have prevented its use until the phalanx was invented specifically in order to prevent attack from the rear; while Cartledge (above p. 20) suggests that the adoption of the new shield implies that ‘a change in tactics in the direction of more organised, hand-to-hand fighting was already in progress’ (his italics; cf. also Cartledge p. 13). There is no denying that the hoplite shield did have disadvantages for soloists; but its advantages over single-grip versions—larger size and greater rigidity, both made possible by the double grip—will have outweighed its drawbacks at least for the more self-confident solo fighters, for it was more especially in flight that it became a liability. Cartledge overestimates the difficulty of manoeuvring the double grip shield for a solo fighter, who could protect his right flank against missiles (or even sword thrusts, though in close combat between right-handed soloists it is very difficult to attack an opponent's right side) not only by moving his shield to the right but also by a body swerve to the left. There can be no doubt that some soloists might have preferred the double-grip shield; that alone makes it impossible to date the phalanx by reference to the date of the shield.

⁷ Oost, (CP lxvii [1972] 10–30)Google Scholar has argued that ‘Cypselus the Bacchiad’ claimed to be restoring orthodox Bacchiad government after the excesses of the last years of their regime. His case is in my view overstated; but however that may be, it can hardly be argued that a majority of Bacchiad hoplites supported Cypselus in view of the tradition that his coup involved the wholesale slaughter or exile of the Bacchiads, (FGH 90Google Scholar F 57, 8; cf. Hdt. v 92 є 2).

⁸ Detienne suggests an unorthodox chronology for the phalanx, on the ground that the Spartans were already fighting in this manner in the First Messenian War c. 715 (La Phalange 139); but his evidence is the romance that passes for Messenian history in Pausanias iv, and cannot be taken seriously (Pearson, , Historia xi [1962] 412–6, 425)Google Scholar.

⁹ Payne, , Protokorinthische Vasenmalerei 20Google Scholar, with Necrocorinthia 16; modified slightly by Coldstream, , Greek Geometric Pottery 326–30Google Scholar.

¹⁰ Cook, , BSA lxiv (1969) 13–15Google Scholar, esp. 14.

¹¹ For the order of the phalanx vases, cf. Dunbabin, and Robertson, , BSA xlviii (1953) 179Google Scholar: Macmillan Painter nos. 10–12. There seems to be general agreement that the crucial vase, no. 11 (Fig. 1 above), belongs to the time when MPC II was giving way to LPC; that is, on Payne's scheme, c. 650. The magnificent vase from Samos (Walter, , AM lxxiv [1959], 60–3Google Scholar, pl. 54, 102–3, 114.2, fig. 1) gives no help. It does not clearly depict phalanx fighting; and although it belongs to the mid-seventh century (Walter, op. cit. 61) it stands outside the main development of PC and cannot be given a precise date.

¹² BSA lxviii (1953) 179. He is named the Ekphantosmaler by Benson, , Geschichte der korinthischen Vasen 18–9Google Scholar. Dunbabin and Robertson attribute far more vases than Benson to this painter's hand; but all are agreed that he painted the three phalanx vases I discuss here (the Macmillan aryballos: below, 88).

¹³ Cf. Snodgrass, , EGAW 198Google Scholar.

¹⁴ The overlapping of the figures on the Lechaeum aryballos (below p. 89 with n. 20) is inconsistent; see also Lorimer, 101 n. 2, on the Perachora aryballos (below, p. 89 with Fig. 4). An early example of successful overlapping: Johansen, Vases Sicyoniens pl. 34.2 (not much before the Macmillan aryballos, which I argue below is the first recognisable attempt at depicting a phalanx).

¹⁵ Cf, Cartledge's excellent description, above, pp. 15–16.

¹⁶ The C Painter, however, seems to have had as great an interest in the problems of depicting the phalanx as the Macmillan Painter (for whom see below), and to have devised similar methods; cf. Beazley, , Development of Attic Black-Figure 23–4Google Scholar.

¹⁷ Compare the extremely effective Assyrian relief illustrated by Myres, Homer and his Critics pl. 6 d; even here, however, the battle is not yet in progress. On the difficulty of depicting phalanx fighting, cf. Fittschen, , Untersuchungen zum Beginn der Sagendarstellungen bei den Griechen 34Google Scholar.

¹⁸ Cf. Benton, , BSA xlviii (1953) 340Google Scholar n. 546: the blazon of a lion's head with lolling tongue symbolises the death of the owner of the shield.

¹⁹ Dunbabin, in Perachora ii 16Google Scholar; cf., however, Fittschen, op. cit. (above, n. 17), 179.

²⁰ C. W. J., & Eliot, M., Hesperia xxxvii (1968) 348–50Google Scholar.

²¹ Fittschen, loc. cit. (above, n. 19) supposes that this vase depicts contemporary and not mythological warfare; we are agreed, at least, that the flautist figured in contemporary battles.

²² Perachora ii, pl. 79, no. 2269; the date is uncertain, but is probably after the end of PC and therefore no earlier than c. 640 (Dunbabin, in Perachora ii 236)Google Scholar.

²³ Snodgrass, , EGAW 198Google Scholar. It is, however, impossible that both spears would have been thrown, as seems to be implied by the fact that a loop is attached to each of the spears (one longer than the other) fixed in the ground beside the man arming to the left of the scene. For another possible case of hoplites with two spears (this time from Sparta), see below, n. 40.

²⁴ The obvious conclusion (despite the mistakes on the Chigi vase), already drawn by Snodgrass, , EGAW 198–9Google Scholar.

²⁵ This view is not beyond question; but fr. 2 West (the spear as the poet's most important possession) and fr. 5 West (the shield thrown away) almost clinch the case.

²⁶ Forrest, , Historia vi (1957) 163–4Google Scholar; Donlan, , TAPA ci (1970) 131–42Google Scholar, esp. n. 22; Greenhalgh 73, 90–3.

²⁷ Snodgrass, , EGAW 179–80Google Scholar.

²⁸ This conclusion will stand even if Tyrtaeus encourages his listeners to remain foot to foot, shield to shield etc. with their enemies, and not (as seems to me far more likely) with their comrades; ‘bite your lip and stand your ground’ is exhortation addressed to a hoplite, not to a soloist.

²⁹ Cf. Donlan, , TAPA ci (1970) 138–9Google Scholar, n. 22. Cartledge (above pp. 25–6) suggests, perhaps too cautiously, that Tyrtaeus should not be interpreted so precisely as I have done here.

³⁰ On the oracle, see below, p. 93. For linen corselets, see Snodgrass, , EGAW 183Google Scholar with references in n. 54; add Törnquist, , Opusc. Rom. vii (1967/9) 81–2Google Scholar.

³¹ c. 750: especially Huxley, , BCH lxxxii (1958) 588–97CrossRefGoogle Scholar. c. 675: especially Andrewes, , CQ xliii (1949) 74–7Google Scholar.

³² Ephorus, (FGH 70Google Scholar F 115) wrote that Pheidon ‘restored the lot of Temenus’. Whatever that may mean in territorial terms, and however much imagination Ephorus exercised in his reconstruction of Pheidon's career, it is unreasonable to carry scepticism so far as to doubt the general view of Pheidon as a conqueror.

³³ See especially, on the attribution to Caria, Snodgrass, , JHS lxxiv (1964) 107–18CrossRefGoogle Scholar.

³⁴ For references, see Snodgrass, , EGAW 262Google Scholar n. 65.

³⁵ Snodgrass, (EGAW 64)Google Scholar suggests that the reason was that Argives won a near monopoly in the manufacture of the hoplite shield; but the production of such a vital piece of military equipment cannot have been monopolised by any one city after its general adoption. See in general Kunze, , Olympische Forschungen ii 215–30Google Scholar; when Kunze wrote, all identifiable inscriptions on shield bands were in the Argive script (op. cit. 212–14), but Corinthian lettering has turned up since then (AD xvii (1961/2) B, 120).

³⁶ Cf. Andrewes, , Tyrants 39–40Google Scholar. The view I have adopted in the text depends on the existence of two chronological layers in the oracle. As Professor Parke points out to me, it is possible that the Delphic preference for riddles might have been responsible for the lack of logic in the reply; if so, the reign of Pheidon will remain the most likely context for the oracle, but the connection with the phalanx will be lost (cf. Parke and Wormell, The Delphic Oracle i 82–3). Riddles, however, are appropriate to predictions, but not to the type of utterance in question here (for the full evidence, see Parke and Wormell, op. cit. ii 1–2); only the assumption of two separate stages in composition can explain the pointlessly illogical expression.

³⁷ Cf. Wade-Gery, , CQ xliii (1949) 79–81CrossRefGoogle Scholar. I am not persuaded by Kelly, (AJP xci [1970] 31–42)Google Scholar to doubt that the Argives (whoever led them) defeated the Spartans at Argive Hysiae in 669. On the general reliability of the early Olympic victor lists, see Kiechle, , Messenische Studien, 10–3Google Scholar; any one date may of course be questioned even if the list as a whole is reliable, but I see no special reasons for scepticism in this case.

³⁸ Especially Hammond, , Studies in Greek History 46–103Google Scholar, esp. 85–90; Kiechle, , Lakonien und Sparta 255Google Scholar.

³⁹ Inevitably many of the arguments against the early dating—which is supported by the agreement, more or less broad, of the ancient evidence—are general (e.g. Forrest, , A History of Sparta 55–6)Google Scholar; but they are none the less persuasive for that. If the rhetra was a written document; as I believe (contra, Sealey 253–6), it cannot be dated before the arrival of writing in Greece c. 750 (Jeffrey, , Local Scripts of Archaic Greece 12–21)Google Scholar, and is most unlikely to have been as early as that.

⁴⁰ Snodgrass, 116. He is followed, less cautiously, by Toynbee, , Some Problems of Greek History 225–6Google Scholar; he argues that the ‘objective evidence’ for the chronology of the phalanx is enough to refute the ‘acute trains of reasoning’ by which Forrest attempts to establish c. 675 as the date of the reform. For a possible depiction of a Spartan phalanx before 650, see now Cartledge, above, p. 27 with his Fig. 1. If this seal does depict a phalanx (which is not certain), it provides a further case of the use of two spears by members of a phalanx, for the hoplites to left and right both seem to be carrying two spears (see above, p. 90).

⁴¹ For a discussion and bibliography, see Staveley, , Historia v (1956) 75–84Google Scholar. Add (e.g.) Momigliano, , JRS liii (1963) 117–21Google Scholar; Drummond, , JRS lx (1970) 201–2Google Scholar with references.

⁴² Thus Mossé, (La tyrannie dans la Grèce antique 8)Google Scholar cannot be right in referring to hoplites as ‘en proie à une grave crise agraire, lourde de menaces pour sa liberté et son indépendance’ (cf. also op. cit. 29); see Zörner, , Kypselos und Pheidon von Argos 106–7Google Scholar.

⁴³ The reluctance is no more than possible, for Cartledge (above p. 21–2) makes the at least equally plausible suggestion that since warfare in this period, like all hoplite warfare, was largely a matter of defending (or threatening) crops, wealthy non-aristocrats would have been keen to enlist; they would, after all, have been defending their own substantial plots.

⁴⁴ Followed by Greenhalgh, 150–5.

⁴⁵ Cf. Cartledge, above, p. 22; despite disagreement over details, we are in accord on this central issue.

⁴⁶ This passage is closely related to others in Aristotle, in which the same connection between military and political influence is made: Pol. 1321a5–14, cf. 1274a12–5; Ath. Pol. 27.1 (the ναυτικὸς ὄχλος at Athens). See also Ps.-Xen., , Ath. Pol. 1.2Google Scholar.

⁴⁷ Andrewes, (Tyrants 34–5)Google Scholar and Cartledge (above, pp. 18–19 60) are both rightly sceptical about Aristotle's phase of cavalry supremacy; see, however, Alfoldi, , Festschrift Schefold 27–8Google Scholar.

⁴⁸ Drews, (Historia xxi (1973) 129–44)Google Scholar has emphasised the personal ambition of the early tyrants, and scorned explanations which ‘meet the specification of the social sciences, and duly present the tyrants as the necessary consequence of external, objective conditions’. In so far as he attempts to demonstrate more than the self-evident fact that tyrants were ambitious, I am out of sympathy with his conclusions; society in the tyrants' cities had reached a stage at which conditions could be exploited—though in Athens they had not, as the equally ambitious Cylon found to his cost.

⁴⁹ Snodgrass 121 suggests on the evidence of Paus, viii 50.1 that Achaea did not adopt the new methods until the third century (cf. EGAW 184); this would cast grave doubt on the view I have taken in the text, but the conclusion seems most unlikely. Pausanias refers to Achaea not as a geographical entity but as the Achaean League; the member cities of that organisation (including by now, among others, Corinth and Argos) cannot have waited until the time of Philopoemen to become true hoplite armies (though the League itself might have given up using hoplites in favour of light-armed troops in the circumstances of the third century). Pausanias must refer either to the readoption of phalanx fighting, or to the adoption of a particular style of it (presumably the Macedonian: Anderson, BSA xlix (1954) 85 n. 119); he cannot be taken to imply that the states of the Achaean League had not employed true hoplites in the seventh century. See in general Anderson, loc. cit.; Errington, , Philopoemen 63–4Google Scholar, however, is much less sceptical. I have taken it that Pheidon's neighbours quickly developed their own phalanxes; but Cartledge (above, p. 18) justly points out that it was not inevitable that they should do so, since the new mode of fighting was ill adapted to the mountainous Greek terrain. Some means of defence, however, had to be found, for the phalanx could put paid to any aristocratic rabble; and it is hardly surprising that (as future development shows) the states under threat failed to devise light-armed forces but adopted the new methods themselves instead.

⁵⁰ For a different view, cf. Sealey, esp. 262–9.

⁵¹ Andrewes, , Tyrants 46Google Scholar, is sceptical about the value of this evidence; contra, Oost, , CP lxvii (1972) 18–9Google Scholar with n. 38. It is most unlikely to be pure invention, for no imaginative reconstruction would have given civil functions to a magistrate with a military title.

⁵² FHG 105 F 2; ascribed to Ephorus by Jacoby ad loc. I have far less faith in the worth of this evidence than in that of Nic. Dam. on Cypselus (above).

⁵³ Ar. Pol. 1315b27–8; FGH 90 F 57, 8.

⁵⁴ Cf. Forrest, , Emergence 112Google Scholar.

⁵⁵ Andrewes, , Tyrants 34Google Scholar; especially Forrest, , Emergence 94–7Google Scholar.

⁵⁶ Full references cannot be given here. Already Hesiod attacked the of Boeotia (above, p. 95); cf. the of Sparta after the reforms, referred to in hexameters attributed to Terpander (fr. 4 Diehl; cf., however, Page, Poetae Melici Graeci 363). Solon broke the Eupatrid domination in Attica; and cf. Ar. Pol. 1311b26–30 on seventh-century Lesbos.

⁵⁷ There is some evidence that Theagenes gained support in Megara from the poor, since he ‘slaughtered the flocks of the wealthy’ (Ar. Pol. 1305a24–6); but I do not trust this information, for it seems more than likely that Aristotle has interpreted a tradition concerning Theagenes' attacks on Megarian aristocrats in the light of his own experience of fourth-century struggles between rich and poor. Cf. Oost, , CP lxviii (1973), 188–90Google Scholar.

⁵⁸ On δίκη in this period, see now Gagarin, , CP lxix (1974) 186–97Google Scholar; the order, and the existence of rules, however, which δίκη denotes should not be restricted to the narrowly legal sphere (cf. Cartledge, above, p. 22 n. 85).

⁵⁹ It is uncertain whether the Corinthian oligarchy was narrow or moderate; but effective power was probably in the hands of few men. The only direct evidence is Dam., Nic.FGH 90Google Scholar F 60, as restored by Will (Korinthiaka 609–15): there was a council of 80, 8 of whom were probouloi. The small council indicates strongly that the oligarchy was not widely based.

⁶⁰ The bibliography is endless; for a useful selection, cf. Sealey, 250–1 n. 7.

⁶¹ Cf. Ar. Pol‥ 1306b37–1307a2, where the authority of Tyrtaeus is claimed for Aristotle's view that land redistribution was demanded as a result of Forrest, (Phoenix xvii (1963) 171)Google Scholar identifies the Messenian War as the First; this is the most plausible, though not the only possible, explanation, and it would mean that land redistribution was an issue at the time of the rhetra (cf. Cartledge, above, p. 27).

⁶² Cf. Hdt. v 66.2:

⁶³ Tyrtaeus fr. 4 West; the text as we have it does not name the kings, but cf. Forrest, , Phoenix xvii (1963) 158–60Google Scholar.

⁶⁴ Forrest, , Phoenix xvii (1963) 170–1Google Scholar.

⁶⁵ Perhaps this is the explanation of the apparent curiosity noted by Momigliano, , JRS liii (1963) 119Google Scholar: it was only when other factors were added to the hoplite reform that political changes ensued, and those other factors did not exist to a significant extent in Etruria.

⁶⁶ A similar point is made about the far more developed society of England in the 1920s by Taylor, A. J. P., English History 1914–1945 334Google Scholar: ‘Other things being equal, those who rule go on ruling, and those who are ruled acquiesce.’

⁶⁷ On all this, see especially Préaux, , Chronique d' Égypte xi (1936) 522–52CrossRefGoogle Scholar.