¹ I would like to thank the following for the financial assistance which enabled me to conduct the research for this article: the Keith and Dorothy Mackay Travelling Scholarship (University of New England), the Australian Archaeological Institute at Athens, and Mr G. L. Frazer. Associate Professor G. R. Stanton kindly read and commented on an early draft of this paper which also benefited from the referees' comments. The final version was prepared at the University of Queensland in 1989. Unless otherwise specified all dates are BC, all Greek quotations are from the Oxford text, and all translations are my own. The following abbreviations apply:

Anderson, MT = Anderson, J. K., Military theory and practice in the age of Xenophon (Berkeley & Los Angeles 1970).Google Scholar

Hanson = V. D. Hanson, Warfare and agriculture in classical Greece (Pisa 1983).

Holladay = Holladay, A. J., ‘Hoplites and heresies’, JHS cii (1982) 94–103.CrossRefGoogle Scholar

Ober FA = Ober, J., Fortress Attica (Leiden 1985).Google Scholar

Ober ‘Thucydides’ = Ober, J., ‘Thucydides, Pericles, and the strategy of defense’, The craft of the ancient historian: essays in honour of Chester G. Starr ed. Eadie, J. W. and Ober, J. (New York 1985) 171–89.Google Scholar

OPW = de Ste Croix, G. E. M., Origins of the Peloponnesian war (London 1972).Google Scholar

Westlake Essays = Westlake, H. D., Essays on the Greek historians and Greek history (Manchester 1969).Google Scholar

Westlake ‘Seaborne raids’ = Westlake, H. D., ‘Seaborne raids in Periclean strategy’, CQ xxxix (1945) 75–84.CrossRefGoogle Scholar

² Thukydides i 143.4–5.

³ See for example Beloch, K. J., Griechische Geschichte ² ii I (Berlin 1914) 300 and n. I.Google ScholarOPW 208–9 also stresses the inability of Perikies' plan to win the war. However, Knight, D. W., ‘Thucydides and the war strategy of Perikles’, Mnemosyne ser. 4, xxiii (1970) 150–61CrossRefGoogle Scholar goes further and argues that in a prolonged war fought according to Perikles' strategy the cost of maintaining her navy and empire would eventually have proved too heavy for Athens.

⁴ Westlake ‘Seaborne raids’ 75–84.

⁵ Ibid. 84; for the view that Perikles had in fact planned an offensive war see H. T. Wade-Gery, OCD ² 1069.

⁶ Cf. OPW 209 and Kagan, D., The outbreak of the Peloponnesian war (New York 1969) 334–5.Google Scholar

⁷ It was quite some time after this article had been accepted for publication that I discovered the discussion in Ober ‘Thucydides’, luckily in sufficient time to make some acknowledgement of his important contribution to the debate in the footnotes and to add a short appendix covering some points of difference. Although there is inevitably some overlap in content (for example we are in complete agreement that the cavalry was used to defend Attika) Dr Ober and I approached the subject with different aims and from a different direction—we also differ somewhat about the degree of Perikles' forward planning and the extent of the protection intended for Attika. Hanson (104–6) also notes the defensive rôle of the cavalry, although without relating it to the war strategies of Perikles and his successors. He also underestimates (103 n. 1) the value of cavalry against formed bodies of infantry, see below pp. 97–9.

⁸ εἰ γὰρ ἧμεν νησιῶται, τίνες ἂν ἀληπτότεροι ἧσαν; καὶ νῦν χρὴ ὅτι ἐγγύτατα τούτου διανοηθέντας τὴν μὲν γῆν καὶ οἰκίας ἀφεῖναι, τῆς δὲ θαλάσσης καὶ πόλεως φυλακὴν ἔχειν καὶ Πελοποννησίοις ὑπὲρ αὐτῶν ὀργισθέντας πολλῷ πλέοσι μὴ διαμάχεσθαι.

⁹ Westlake ‘Seaborne raids’ 75; it is also unequivocally accepted by Kagan (n. 6) 334.

¹⁰ Cf. Ar. Ach. 1018–36 and 1071–7. The statement at Thukydides ii 18.2 that Oinoe was there to guard the frontier should be interpreted to mean guard against raids or small incursions. The fort was patently unable to deal with anything larger, as the siege of 431 shows: Thuky. ii 18.

¹¹ For the general pattern of hoplite warfare see Gomme HCT i 10–15 and Anderson MT 2–3. For the moral victory see Xen. Hell. vi 5.21 and below, pp. 104–6.

¹² Gomme HCT i 16–19; Lawrence, A. W., Greek aims in fortification (Oxford 1979) 39–42.Google Scholar The siege of Potidaia for example cost two thousand talents, Thuky. ii 70.2.

¹³ For example, Xen. Hell. v 4.56 and vii 2.10, 17 ff.

¹⁴ Plut. Ages, xxii 5; the tactic worked.

¹⁵ Thuky. iv 84.2–88.1, cf. Epameinondas' decision to despatch his cavalry against Mantineia, Xen. Hell. vii 5.14.

¹⁶ I exclude here the passive measure of completely abandoning the chora without any defence at all. The three main types of defence: hoplite, border, and mobile, are readily identifiable to anyone working in the field but the pre-emptive strike is less obvious and was first drawn to my attention by the discussion in Ober FA 70 ff.; he also examines the other types there.

¹⁷ Thukydides iv 92.6. The battle of Sepeia in 494 kept the Argives quiet for a generation, Herodotos vi 83.

¹⁸ Krentz, P. ‘Casualties in hoplite battles’ GRBS xxvi (1985) 14–20Google Scholar, especially the table on 19.

¹⁹ Thuky. iv 91.

²⁰ Cf. Thuky. i 143.5 on this as a problem even if the Athenians were to win a pitched battle against the Peloponnesian League.

²¹ Although, as I argue below (p. 104), the loss of the Boiotian cavalry would have severely hampered League operations in Attika.

²² OPW 190–5. Cartledge, P. A., ‘Hoplites and heroes: Sparta's contribution to the techniques of ancient warfare’, JHS xcvii (1977) 22 ff.Google Scholar

²³ For a list of examples of its use see OPW 192–4. For objections to this strategy in a general Greek context see Holladay 97–9.

²⁴ For a more detailed description of the routes into Attika see Cruickshank, W. W., Topography, movement and supply in the warfare of ancient Greece south of Thessaly and Epirus (Diss. London 1955) 174–278.Google Scholar Ober FA 111–29 also has a useful discussion.

²⁵ Holladay 98. Agesilaos' seizure of the passes into Boiotia in 378 and 377 is a good example of this, Xen. Hell. v 4.36–7, 47–8.

²⁶ Thukydides iv 94.1.

²⁷ Gomme HCT i 14; Anderson MT 5. For the daily ration per man and the number of transport animals required to move supplies see Engels, D. W., Alexander the Great and the logistics of the Macedonian army (Berkeley 1978) 18–19 and 123–6.Google Scholar

²⁸ Thuky. ii 18.1–2.

²⁹ Holladay 99.

³⁰ An exception to this is the Plataian campaign of 479 where the Persian cavalry succeeded in destroying a Greek supply train en route for the army, Herodotos ix 39.

³¹ Cf. Thuky. viii 98.2 for the rôle of the Oinoe garrison in ambushing a Korinthian force returning from duty in Dekeleia.

³² Ober FA passim, especially 191–222.

³³ OPW 190 ff.

³⁴ Although the Megarian decrees of the 430's may well have been designed to force Megara into such an arrangement (Hornblower, S., The Greek world 479–323 [London 1983] 92Google Scholar) this had clearly not been achieved by the time war broke out.

³⁵ Ober FA 192–5.

³⁶ For example, Anderson MT 58, Gomme HCT i 15.

³⁷ ‘. . . τῆς μὲν γῆς ἐκράτουν ὅσα μὴ προϊόντες πολὺ ἐκ τῶν ὅπλων (οἱ γὰρ ἱππῆς τῶν Θεσσαλῶν εΙ̃ργον) . . .’. Cf. Diod. Sic. xv 71.4–5.

³⁸ Greece: Thuky. i 111.1, ii 100.5; Herodotos v 63.3–4; Diod. Sic. xv 71.4–5; Xen. Hell. v 3.3–5. Other examples of cavalry used against ravagers and foragers (although not necessarily as the sole means of defence) include Xen. Hell. v 3.1–2, vii 1.20–2 and 2.4, 10 (helped by eptlektoi). Sicily: Thuky. vii 4.6; Plut. Mc. xix 6; Diod. Sic. xi 21.2, xiii 44.3–4 and 88.1. Although the Syracusans also used their infantry on occasion it was their cavalry which often played the most effective part. For Persia see Xen. Hell. iv 8.18–19 and p. 102 below.

³⁹ The defeat of Anchimolios' army in Attika in 511 is one of the best examples of this, Herodotos v 63.

⁴⁰ See Xen. Hell. vii 1.21 (quoted below p. 99) and 2.10.

⁴¹ See below pp. 98–100 and also Xen. Hell. iv 8.18–19 (Persian cavalry), vii 2.4 and 21–2 (backed up by hoplites in the second example) and Thuky. ii 79.6, v 10.9 (supported by psiloi in both cases).

⁴² Although this has been challenged by Cawkwell, G. L., Philip of Macedon (London 1978) 150–3Google Scholar, Holladay 947–7 adequately disposes of his arguments. The close-ordering of the ranks was emphasised as late as Onasander xxvii.

⁴³ Harrison, A. R. W.The law of Athens: procedure (Oxford 1971) 32.Google Scholar See also Aristophanes' jibes at Kleonymos and others: Pax 446, 674–8, 1185–6, 1302–4; Av. 290, 1473–81; Nub. 353–4; Vesp. 15–23, 592, 822–3. m Lysias x (c. Theomn.) 6 ff it is associated in discussion with charges of murder, parent-beating, assault, and abduction.

⁴⁴ Plut. Moralia 220A.

⁴⁵ Snodgrass, A. M., Arms and armour of the Greeks (London 1967) 109 fGoogle Scholar; Anderson MT 13 ff. and 40–2; Anderson, J. K., Ancient Greek horsemanship (Berkeley 1961) 141–2.Google Scholar

⁴⁶ I do not wish to develop my theory of cavalry warfare further here, as I believe that it deserves fuller treatment than can be provided in this article. It is enough for the moment to state that I consider that the Greek mounted arm prior to the develop ments under Philip and Alexander has been under valued by most historians. Although incapable of charging into the front of an intact phalanx (but see n. 50 below), such cavalry could be effective against its rear or flanks, could use missile fire to destroy its cohesion, or could take advantage of any disruption caused by obstacles or by enemy action. While some examples of the efficacy of cavalry against even large hoplite forces and the theory of their use are given below, I shall be elaborating my views more fully in a book to be published by Oxford University Press.

⁴⁷ Lorimer, H. L., ‘The hoplite phalanx’, BSA xlii (1947) 76–7Google Scholar; cf. Holladay 95.

⁴⁸ Xen. Hell. vii 5.24. Sokrates' actions during the retreat at Delion are an exception and described as such, Plato Symp. 220E–221B, cf. Plut. Alc. vii 4.

⁴⁹ This receives further confirmation from Euripides HF 190–4 where Amphitryon, arguing that the bow is superior to the spear states that:

A hoplite is a slave to his weapons,

And from the lack of bravery of his formation-fellows

Himself perishes, through the cowardice of his neighbours;

Having broken his spear, he, who has only one defence,

Is unable to ward off death from his body,

⁵⁰ On the type of shock employed by cavalry see Keegan, J., The face of battle (Harmondsworth 1986) 95–6 and 154 ff.Google Scholar; but I have recently been convinced by Dr M. M. Markle that there is nothing in the nature of the horse per se which precludes its riding into a body of troops. He is intending to publish a paper on this area in the near future. However, like Keegan, I believe that cavalry normally achieved its effect by moral rather than physical shock and this is particularly true of the classical Greek arm. The front of a phalanx could be breached, using the tactics described by Arrian Tact, xvi 6–9, but in my opinion only by a force equipped with spears longer than those carried by the infantry (cf. Markle, M. M., ‘The Macedonian sarissa, spear, and related armor’, AJA lxxxi (1977) 339Google Scholar). Any penetration of a hoplite phalanx which stood firm almost certainly occurred at a pace much less than a gallop and was possibly achieved at little more than a walk.

⁵¹ The more lightly equipped hoplites of c. 431 onwards (Anderson MT 41) were of course even more vulnerable to missiles than their more fullyarmoured predecessors. There are problems in determining cavalry weaponry at Athens as the hippeis could be armed in a variety of ways: with two javelins, with a thrusting spear, or with javelins and a thrusting spear. A sword was sometimes carried in addition to any of these combinations. Although Attic vase paintings suggest that most cavalrymen were equipped with two javelins it is sometimes difficult to know how far to trust this evidence. I intend to include a detailed study of this question in my forthcoming book.

⁵² Thuky. iv 125–7; Xen Hell. iv 5.14–17 (used against peltasts but demonstrating the same principles), Anab. iii 3.15. In the first example Brasidas placed his psiloi inside the hoplite square.

⁵³ Diod. Sic. xv 68.1. These figures applied at the start of the expedition and as Xenophon's account of the engagement contains no mention of the Theban cavalry it might quite possibly have been occupied elsewhere at the time.

⁵⁴

⁵⁵ Disorganised troops: Xen. Hell. v 4.39 and 44–5; organised troops: Arrian Anab. i 2.5–6 (in conjunction with infantry), cf. the similar (but less successful tactics) in Thuky. ii 100.4–5. The Getai fled before contact, although Alexander's unexpectedly easy crossing of the Ister apparently contributed to the shock of the attack, Arrian Anab. i 4–3.

⁵⁶ For this equipment see Best, J. G. P., Thracian peltasts and their influence on Greek warfare (Groningen 1969) 139–42.Google Scholar

⁵⁷ Thuky. vii 30.2. He only mentions one item of their equipment, the machaira (vii 27.1), but because of their ability to fight in formation they must also have been armed with a shield and a spear as well. They lost 250 men out of 1300, mainly stragglers or those who drowned trying to swim out to the boats.

⁵⁸ The distance from Mykalessos to the nearest bay is only some 5–7 kilometres so, allowing for the time it took the Theban cavalry to arrive, the pursuit was probably fairly short.

⁵⁹ Thukydides ii 9.3 states that the cavalry was supplied by Boiotia, Phokis and Lokris. Hell. Oxy. xi 3–4 tells us that the cavalry of the Boiotian League in the early fourth century was 1100 strong.

⁶⁰ Although this seems strange today, presumably one of the main problems involved in detailing troops to protect foragers or ravagers was that they were probably unwilling to stand by and watch others have the pick of the available food and/or booty. See also nn. 73 and 75 below.

⁶¹ Hardy, W. G., ‘Hellenica Oxyrhynchia and the devastation of Athens’, CPh xxi (1926) 348 ff.Google Scholar

⁶² Supplies: Thuky. iii 1.3; 425 invasion: Thuky. iv 6.1.

⁶³ For example, Thuky. ii 23.1 and 3.

⁶⁴ Jackson, A. H., ‘The original purpose of the Delian league’, Historia xviii (1969) 12–13Google Scholar; cf. the discussion in Hanson 14–20 (which makes rather more of the use of fire).

⁶⁵ W. G. Hardy (n. 61) 348 n. 4; for house construction see Jones, , Graham, , and Sackett, , ‘An Attic country house below the cave of Pan at Vari’, BSA lxviii (1973) 355–452.Google Scholar

⁶⁶ For the burning of trees and the distance between olive trees see Theophrastos CP ii 3.3 and 5.6. On the difficulty of igniting trees with thick bark see Kayll, A. J., A technique for studying the fire tolerance of living tree trunks, Dept. of Forestry, Canada (Ottawa 1963) 19.Google Scholar The problems involved in burning olive groves have now been remedied by modern ordnance, cf. Robert Fisk, The Times (11th August 1983) 4.

⁶⁷ Watson, D. J., ‘Inflammability of cereal crops in relation to water content’ Empire Journal of Experimental Agriculture, vol. xviii no. 71 (1950) 150–7.Google Scholar The experiments relate to U.K. conditions, and crops in Greece, because of the drier conditions there, would have been vulnerable for a longer period than Watson's results show. However, his overall conclusions, that it is not as easy to fire grain as most think, is still valid. Cf. the discussion at Hanson 42–6 which basically agrees with my views; although he tends to concentrate more on the success of fire most of his examples refer to dry crops (particularly stored grain or harvested sheafs in the field) and he too emphasises the relatively short period of vulnerability.

⁶⁸ D. J. Watson (n. 67) 157. However, he also points out that with large crops such delays are not infrequent because of the difficulty of harvesting the entire crop as soon as it ripens.

⁶⁹ See also Westlake Essays 93 n. 27.

⁷⁰ Anderson MT 3.

⁷¹ For example, Arrian Anab. i 4.1

⁷² Hanson 21–5.

⁷³ Foragers were certainly unprotected (see n. 75) and it seems that unless moving under orders as organised bodies (which reduced the amount of damage possible), ravagers also normally conducted their activities without much protection. Although we might expect psiloi and cavalry to have performed this task it seems that they normally either concentrated on supporting the main body of hoplites or were off ravaging/foraging on their own behalf (see for example Thuky. iii 1.2 for the apparent lack of protection afforded to those operating away from the main force).

⁷⁴ Pritchett, , The Greek state at war i (Los Angeles 1971) 30–52.Google Scholar

⁷⁵ Thuky. vii 4.6 and 13.2 (the Athenians at Syracuse). See Onasander x 7–8 for the defencelessness of those out looking for supplies—he advocated both banning unauthorised foraging and the protection of authorised foragers. Despite this, it was not usual Greek practice to protect foragers and the first examples of this concern Alexander, Arrian Anab. i 5.9 and iii 20.4(?). In later periods such protection may have been provided as a matter of course, Onasander x 8 (cf. Livy xxxi 2 for Roman practice). For the equipment carried by foragers see Xen. Anab. vi 4.23.

⁷⁶ See above p. 97.

⁷⁷ Thuky. ii 10.2, 47.2, and iii 15.1. He does not specify the numbers but Plut. Per. xxxiii 5 states it was sixty thousand; cf. Gomme, , HCT ii 13Google Scholar, who estimates it at not more than about thirty thousand at most.

⁷⁸ Thuky. ii 19.2.

⁷⁹ Thuky. ii 22.2.

⁸⁰ Traill, J. S., The political organization of Attika, Hesperia Supplement xiv (Princeton 1975) map 1.Google Scholar

⁸¹ τὸν πλεῖστον ὅμιλον τῶν ψιλῶν . . . τὸ μὴ προεξιόντας τῶν ὅπλων τὰ ἐγγὺς τῆς πόλεως κακουργεῖν.

⁸² I am grateful to Mr A. French for pointing this out to me.

⁸³ For example, Thukydides' statement (ii 55.1) that in 430 the Peloponnesians ravaged the plain (ἔτεμον τὸ πεδίον) before moving on to the coastnear Laureion and his remarks at ii 57.2 suggest that a larger area was devastated that year. However, I do not believe that ‘the property protection scheme had apparently broken down completely’, Ober ‘Thucydides’ 179. In my opinion it was never intended to provide total protection, merely to reduce the damage as far as possible; in 430, for the reasons stated in Ober ‘Thucydides’ 179–80, this happened to be more difficult than at other times.

⁸⁴ Presumably because of the Athenian threats to kill the captives, Thuky. iv 41.1.

⁸⁵ OPW 194.

⁸⁶ See especially iv 13–20 and vii 5–14.

⁸⁷

⁸⁸ Thukydides vii 28.1. ἥ τε τῶν ἐπιτηδείων παρακομιδὴ ἐκ τῆς Εὐβοίας, πρότερον ἐκ τοὔ Ὠρωποῦ κατὰ γῆν διὰ τῆς Δεκελείας θάσσων οὔσα, περὶ Σούνιον κατὰ θάλασσαν πολυτελὴς ἐγίγυετο τῶν τε πάντων ὁμοίως ἐπακτῶν ἐδεῖτο ἡ πόλις, καὶ ἀντὶ τοῦ πόλις εἶναι φρούριον κατέστη Cf. Plato Crito 43? 2–5 for the practice of passengers disembarking at Sounion and continuing overland, apparently because of the common problem of adverse winds around the cape itself.

⁸⁹ Although, as Hanson points out, 142–3, while damage to agriculture was increased, the main losses were of property.

⁹⁰ Thuky. ii 23.2–31, 56. If de Ste Croix's interpretation of the Megarian decrees (OPW 225–289) is correct then the invasions of the Megarid may also have been influenced by religious considerations.

⁹¹ There are strong links between the hoplite and the concept of the agathos. It was the hoplite who defended the city and his family, the key function of the agathos (Adkins, A. W. H., Merit and responsibility [Oxford 1960] 236–7Google Scholar). Other hoplite virtues such as steadfastness and cooperation also figured in contemporary or near-contemporary ideas about the ideal citizen; see for example Plato Laches 182A, Sophokles Antigone 666 ff., Ar. Ran. 1009–17, cf. Adkins 165.

⁹² Garlan, Y., War in the ancient world (trans. Lloyd, J.) (London 1975) 60.Google Scholar Cf. Thuky. i 140.5, ii 21.2–3 ar,d Diod. Sic. xii 61.2.

⁹³ This was expressed as late as 355 in Isokrates viii 77. See also Xen. Hell, vi 5.20–1, Plut. Nic xx 4–8, and Thuky. viii 27. Although the last two of these examples involve naval engagements, the strategoi involved were of hoplite class. Cf. Plut. Per. xxxiii–xxxiv.

⁹⁴ For similar sentiments see also Hermippos fr. 47 Kassel—Austin = Plut. Per. xxxiii 7. Gomme HCT ii 75–76 cites two fragments of comedy which may also refer to this incident.

⁹⁵ ὡς δυνάμενος στρατηγεῖν καὶ τοῖς Λακεδαιμονίοις διαπολεμεῖν (Budé).

⁹⁶ Thukydides ii 22.2.

⁹⁷ Westlake is correct to state that the morale factor was not the sole reason for Perikles' programme of raids and invasions, Essays 91 ff., ‘Seaborne raids’ 79–80 and 84, but I believe it was an additional, and not unimportant, reason for staging them.

⁹⁸ Although the activities of the garrisons here were almost certainly not directed in detail from Athens but presumably relied upon the initiative of the local commander. It may well be that the function of these forts was to guard against raiders and that their use here was accepted simply as part of the natural scheme of things rather than representing any innovative, or even consciously formulated, part of Perikles' strategy. This is perhaps supported by Thukydides' remarks at ii 18.2 on the rôle of Oinoe in wartime.

⁹⁹ See the discussion in the appendix.

¹⁰⁰ This passage is quoted above p. 104.

¹⁰¹ Ober ‘Thucydides’ 181 and 182.

¹⁰² For its survival as late as 341 cf. Dem. ix 49.

¹⁰³ On the date see Bugh, G. R., The horsemen of Athens (Princeton 1988) 76.Google Scholar

¹⁰⁴ The evacuation of all their property, including the woodwork from their houses (Thuky. ii 14.1), suggests that the Athenians placed little faith in the possibility that their homes could be protected.

¹⁰⁵ Thuky. ii 22.1.

¹⁰⁶ On this see Taylor's, A. J. P. analysis in The origins of the second world war (Harmondsworth 1973)Google Scholarpassim, but especially 98.

¹⁰⁷ See my opening remarks.