The battle of Salamis can be dated with a high degree of certainty. Probably about the time of that battle, Cleombrotus was at the Isthmus, constructing the defences there (Hdt. 8. 71. 1). At some point while building the wall, he considered giving chase to the Persian army. When his sacrifice was answered by a solar eclipse, he took this as a bad omen and immediately returned to Lacedaemon (9. 10. 2–3). The eclipse visible to Cleombrotus could only have been that of 2 October 480. Now it is generally supposed that Cleombrotus would not have thought to abandon the construction of the wall and pursue Xerxes unless the latter had just begun his retreat from Athens. Thus, as Herodotus says that a few days () after the battle of Salamis Xerxes withdrew from Attica (8. 113. 1), the battle of Salamis probably occurred before 2 October 480.
1 Ginzel, F. K., Handbuch der mathematischen and technischen Chronologie (Leipzig, 1911-hereafter, Ginzel), ii. 526.
2 C. Hignett, Xerxes' Invasion of Greece, p. 274 (hereafter, Hignett).
3 F. Jacoby, FGrHist 392 F 7 and his commentary, note 62.
4 Georg Busolt, Griechische Geschichte 2 ii. 702, n. 2.
5 Hdt. 8. 97. Xerxes begins building a mole over to Salamis and prepares his fleet; both are feints to permit his escape. On the historicity of the mole, see Hignett, pp. 415–17.
6 The procession began on 19 Boedromion (IG ii2. 1078 and schol. to Aristoph. Vespae, 324), but Busolt (Gr. Gesch.2 ii. 359, n. 2) explains that the ritual lasted through the night into the next day (so Plut. Cam. 19. 6 and Phoc. 28. 1).
7 Busolt, Gr. Gesch.2 ii. 703, n. 3.
8 Karl Julius Beloch, Griechische Geschichte 2 ii:2, 47.
9 Gr. Gesch.2 ii. 702, n. 2; 703, n. 3; cf. Hignett, 452.
10 Ginzel, i. 551.
11 See (most recently) Pritchett, W. Kendrick, Ancient Athenian Calendars on Stone (Berkeley, 1963), pp. 330–45. Moreover, the festival calendar did not necessarily even begin on the new moon: Pritchett, , ‘Calendars of Athens Again’, BCH 81 (1957), 273–4.
12 16 Mounichion, mentioned twice by Plutarch as the day of the battle (Mor. 349F and Lys. 15. 1), is only the date of its commemoration. For its rationale, see Hermann Sorge, ‘Der Mond auf den Miinzen von Athen’, Jahrbüch für Numismatik and Geldgeschichte 2 (1950–1), 7–13. For an entirely new explanation, see now: Badian, E. and Buckler, J., ‘The Wrong Salamis?’, Rh. Mus. 118 (1975), 226–37.
13 e.g. Schol. to Pindar ol. 3. 35 and ol. 3. 90; Bacchylides 7. 2–3.
14 Euripides, Alc. 445 f., but the interpretation of this passage is far from certain: see the commentary by A. M. Dale (Oxford, 1954), pp. 90–1. In any case, the Carnea and Olympia of 480 appear to coincide: Hdt. 7. 206.
15 Cf. Stephen Miller, G., ‘The Date of Olympic Festivals’, Ath. Mitt. 90 (1975), 228. I wish to thank Dr. Miller for kindly allowing me the use of his galley proofs.
16 For the possible dating of the Carnea, see the Appendix.
17 Walter Snyder, F., ‘When was the Alexandrian Calendar Established?’, AJP 64 (1943), 385–98.
18 Ginzel, i. 200. If the scholion was written before the Egyptian calendar was fixed, then the scholiast would have later dates in mind for the Olympia. As the earlier Egyptian calendar lost one day every four years, Mesori and Thoth moved slowly backwards through the autumn, until they reachec the position at which they were fixed; cf. Beloch, Gr. Gesch.2 i:2, 139. For the Julian equivalents of the months before then, see Lundsgaard, E., Aegyptischer Kalendar der Jahre 3000–200 v. Chr. (Copenhagen, 1942).
19 Ginzel, ii 521(at 38° latitude, since it is an Elean chronicle). The ancients have varying dates for the two risings: see Lydus, Joannes, De Ostendis et Calendaria Graeca Omnia (ed. by Kurt Wachsmuth, 2nd edn., 1897), p. 348, s.v., and p. 346, s.v.
20 Nissen, H., ‘Ueber Tempel Orientirung’, Rh.Mus. 40 (1885), 349–63;Mommsen, A., Über die Zeit der Olympien (Leipzig, 1891); Ludwig Weniger, ‘Das Hochfest des Zeus in Olympia’, Klio 5 (1905), 1–22.
21 e.g. Ziehen, in RE 18.1, col. 2, points out that cannot be a historical present (as it must be if is read ), because of the tense of
22 Beloch, Gr. Gesch.2 1:2, 139–43.
23 Ath. Mitt. 90 (1975), 217, following th caveat of Raphael Sealey, ‘The Olympic Festival of 324 B.C.’, CR 74 (1960), 186.
24 For further statements on the non-use of the octaeteris, see E. J. Bickerman, Chronology of the Ancient World (London, 1968), p.29, and W. Kendrick Pritchett, The Greek State at War (Berkeley, 1974: a reprint of Ancient Greek Military Practices, Berkeley, 1971), i. 117.
25 Fotheringham, J. K., ‘Cleostratus’, JHS 39 (1919), 177, and ‘Cleostratus (III)’, JHS 45 (1925), 83.
26 Originally suggested by Unger, G. F., ‘Der Olympienmonat’, Philologus 33 (1874), 227–48, in order to understand Polybius' dating methods, and followed by Paul Pédech, La Méthode historique de Polybe (Paris, 1964) pp. 449 ff. But see (most recently) Errington, R. M., ‘The Chronology of Polybius’ Histories, Books I and II', JRS 57 (1967), 96–108, especially 99, and (for bibliography) Walbank, F. W., Polybius (Berkeley, 1972), pp. 101 ff.
27 Beloch, Gr. Gesch.2 i:2, 139–43; for the latest defender of Nissen's chronology, see Weniger, Klio 5 (1905), 8–14; Miller, Ath. Mitt. 90 (1975), 227–31. They all in turn dispute the inferences from the evidence for the following Olympic festivals: Ol. 75 (480 B.C.), Ol. 88 (428 B.C.), Ol. 90 (420 B.C.), Ol. 106 (356 B.C.), Ol. 143 (208 B.C.), and Ol. 184 (44 B.C.). Miller, p. 230 produces another date: Ol. 114 (324 B.C.). But this is merely suggested by Sealey CR 74 (1960), 186–7), and tentatively accepted by E. Badian (JHS 81 (1961) 42–3), through the employment of Beloch's Olympic cycle.
28 Evidence for the and , collected and analysed by Giovannini, A., Étude historique sur les origines du catalogue des vaisseaux (Berne, 1969), pp. 55 ff.
29 It was only after this paper was in its final form that I gained access to Alan E. Samuel's Greek and Roman Chronology (Handbuch der Altertumswissenschaft, vii:1 (Munich, 1972)). We appear to be in general agreement on most of the important issues surrounding the dating of Olympic festivals (pp. 191–4), but one difference should be brought out. Samuel, ignoring the argument that the scholiasts were only speaking of approximate alternation between forty-nine-Ind fifty-month periods, believes that the ontradiction between the information given by them and the theoretical models of octaeterides indicates that the scholiasts themselves were confused. Just as likely, lowever, is the possibility that they did not ntend their information to be construed as iescribing perfect octaeterides. Parenthetic-Lily, Samuel treats the information found in TT 1 and 3 as written by one scholiast. But we are here presented with two sources of widence, for the scholiast is quoting another Luthor in T1. Thus we have testimony from an Alexandrian scholiast and an Elean chronicler.
30 Beloch, Gr. Gesch.2 i:2, 140; cf. ii:2, 148–9.
31 Hignett, p. 449.
32 J. E. Powell, A Lexicon to Herodotus 2 (1966), s.v., . The analyses of are also based upon Powell.
33 W. W. How and J. Wells, A Commentary on Herodotus, ii. 238.
34 A. W. Gomme, A Historical Commentary on Thucydides, iii. 702 and 705; and W. K. Pritchett and B. C. van der Waerden, ‘Thucydidean time-reckoning and Euctemon seasonal calendar’, BCH 85 (1961), 17–52. But the controversy concerning the specific dates continues. See (most recently) Pritchet ‘The Thucydidean Summer of 411 B.C.’, CPh 60 (1965), 259–61 and The Choiseul Marble (Berkeley, 1970), 94–5; and for the opposing view: B. D. Merritt, ‘A Persian Date in Thucydides,’ CPh 61 (1966), 182–4, and Andrewes, A. (with A. W. Gomme and K. J. Dover), A Historical Commentary on Thucydides, iv. 18–21.
35 Or by interpreting I9Epoc as understood by Hesiod or Hippocrates: Deman, A., ‘La Date de la bataille des Thermopyles’, RBPh 36 (1958), 100, and Jules Labarbe, ‘Léonidas et l'astre des tempêtes’, RBPh 37 (1959), 84–6.
36 Apostolos Dascalakis, Problames historiques autour de la bataille des Thermopyles (Paris, 1962-hereafter, Dascalakis), p. 130.
37 Jules Labarbe, ‘Un Vmoignage capital de Polyen sur la bataille de Thermopyles’, BCH 78 (1954), 1–21;RBPh 37 (1959), 69–91.
38 Deman, RBPh 36 (1958), 96–102; Dascalakis, pp. 111–40; Hignett, pp. 450–1; Burn, A. R., Persia and the Greeks (1962), pp. 403–4; and Evans, J. A. S., ‘Notes on Thermopylae and Artemisium’, Historia 18 (1969), 400, n. 65.
39 ‘Leonidas no. 3’ in RE 12, col. 2018 f.
40 On Dexippus, see (most recently): Millar, Fergus, ‘P. Herennius Dexippus, the Greek World and the Third-Century Invasions’, JRS 59 (1969), 12–29.
41 Eunapius' use of in the active voice makes no sense, but in the passive the verb means precisely the rising of a star (LSJ s.v. (B)). Eunapius' misuse of the verb is not troublesome: even in antiquity, he was known for his idiosyncratic vocabulary (Photius, no. 77, 159 (54a. 13–16)), and LSJ often records his unique usage of standard vocabulary: e.g. , no. 2.
42 Cobet, C. G., ‘Ad Eunapii Fragmenta’, Mnemosyne 10 (1882), 26–7, believes that Eunapius was actually quoting Plutarch's statement in dating the battle of Salamis. But Eunapius' reference to the movement of a star is not found in Plutarch. Furthermore, as Eunapius in this passage is aiming his polemics directly at Dexippus, it must be supposed that these statements too are found in Dexippus. Jacoby (FGrHist. 100 F 1, 6. 25–8) records the sentences as direct quotations from Dexippus.
43 On the value of dating Euripides' birth to the battle, see Jacoby, F., Apollodorus Chronik (Berlin, 1902), pp. 250–60.
44 See the discussions in Hignett, pp. 448–9, and Dascalakis, pp. 140–69.
45 Cf. Hignett, pp. 451–3; most recently Lazenby, J. F., ‘The Strategy of the Greeks it the Opening Campaign of the Persian Wars’, Hermes 92 (1964), 265, and Sealey, Raphael, ‘Again the Siege of the Acropolis’, CSCA 5 (1972), 191.
46 Giulio Giannelli, La spedizione di Serse da Terme a Salamina (pubblicazioni della universitá cattolica del sacro cuore, serie quinta: scienze storiche, vol. v (Milan, 1924)-hereafter, Giannelli), p. 21, dates the battle of Thermopylae to 21 August and the arrival of the Persian fleet at Phalerum to 29 August. Thus he is left with a four-week hiatus in events.
47 Sealey, R., ‘A Note on the Supposed Themistocles-Decree’, Hermes 91 (1963), 376–7, and more fully in CSCA 5 (1972), 183–94. Originally argued by Bury, J. B., ‘Aristides at Salamis’, CR 10 (1896), 414–18; supported by Munro, CAH v. 300, 303–4, Grundy, G. B., The Great Persian War (London, 1901), pp. 357–9, and Dascalakis, p. 147.
48 Nothing is to be gained from the reference in the Themistocles Decree (Meiggs and Lewis, GHI, no. 23), lines 11–12; see Sealey, CSCA 5 (1972), 184–6.
49 Discussed by Sealey, loc. cit., 188–9.
50 Hignett, p. 213. In fact, conservative political elements in Athens were apparently inventing stories about feats of valour performed by hoplites during the battle of Salamis: Charles Fornara, W., ‘The Hoplite Achievement at Psyttaleia’, JHS 86 (1966), 51–4.
51 Hignett, pp. 213–5.
52 Cf. Hignett, p. 206, n. 4.
53 Ernst Obst, Der Feldzug des Xerxes, Klio 12 (1914), Beiheft 164–6; apparently also Alb. Deman, Miltiade et Themistocle, Recherches sur l'histoire politique, diplomatique et militaired 'Athenes, 500–480 ay. J.-Chr., p. 281, cited in RBPh 36 (1958), 96 n. 1, and unavailable to me.
54 Eugène Cavaignac, Histoire de l'antiquité (Paris, 1919), 1:2, 383–95, presents a some what different view. Herodotus states that an eclipse occurred at Salamis, in 480, when Xerxes began his march (7. 37); but no eclipse is known to have been visible there that spring. Cavaignac believes it is actually the eclipse of 2 October and Xerxes was starting out, not from Sardis, but from Therma. This gives Cavaignac dates of 16–18 October for the battle of Thermopylae and 26 October for the battle of Salamis, with the Olympia culminating on 17 September. Furthermore, Cleombrotus saw that same eclipse, not when he was considering giving chase to Xerxes (9. 10. 3), but when he was bringing aid to Leonidas. If for no other reasons, Cavaignac's thesis fails, because Herodotus explains that the Peloponnesians did not have time to reinforce Leonidas (7. 206) and if that was what Cleombrotus was intending, why does Herodotus say that he was at the Isthmus building the wall (8. 71. 1; 9. 10. 2–3)?
55 Others have recognized that Herodotus' internal chronology is completely consistent, but their Julian dates appear to be in error: Obst and Deman (loc. cit.) and Cavaignac (loc. cit.); Macan, R. W., Herodotus, the Seventh, Eighth, and Ninth Books (London, 1908-hereafter, Macan), ii. 293, generally holds to Herodotus' chronology without supplying the Julian dates. Hignett, p. 211, believes that Giannelli, pp. 37–49, has provec Herodotus inconsistent in this matter; but this is not so, nor does Giannelli claim to hav done such. He merely argues that Herodotus‘ chronology is a priori impossible on military grounds: the Persian army and fleet would need more time to take their positions (pp. 40–1). Giannelli's analysis could correspond to the presently proposed scheme: his day Y1 = 27 September; Y2 and X1 = 28 September; and X2 = 29 September.
56 Ginzel, ii. 557.
57 The battles most likely were concluded on the same day: Hdt. 8. 15. 1; cf Hignett, pp. 379 ff. In any case, as all alternative chronologies have the battle of Artemisium occurring before that at Thermopylae, the possibility of the two engagements not being simultaneous does not affect the present discussion.
58 Cf. Macan, i:2, 390. The Persians did not take Histiaea until well after midday (8. 23. 2). Because the visit to Thermopylae took an entire day (8. 25. 3), it must have occurred on the day after the capture of Histiaea.
59 e.g. Hignett, p. 193, Macan, ii. 292, and Giannelli, p. 38.
60 Xen. Anab. 1. 2. 5–4. 11; cf. Obst (loc. cit.) 132 f. See Hignett, pp. 195 f., for the standard objections.
61 Cf. Giannelli, note on pp. 42–4.
62 Especially by Hignett, p. 206, and Giannelli, pp. 37–49.
63 On the length of the campaigning season as a consideration, see Evans, , Historia 18 (1969), 388–406.
64 Contra, Giannelli, p. 47. Similarly, Xerxes gave his main army two days' rest at Thermopylae before beginning the forced march.
65 Cf. Hignett, p. 211.
66 See Hignett, p. 448.
67 BCH 78 (1954), 19, n. 2. For the moods of , which, in this construction, must have the same force, see Smyth, Greek Grammar (1966), § 2615; cf. 1862 a, 1863 a and b.
68 See pp. 232–3. The present chronology has the main Persian army arriving in Athens two days before the battle and possib the vanguard up to two days before that. As the anecdote concerning the Iacchus procession does indicate that the Persians were in Athens by the time of that celebration, 19/20 Boedromion must have occurred no more than four days before the battle of Salamis.
69 Pritchett, The Greek State at War, i. 117; Gomme, Andrewes, and Dover, A His torical Commentary on Thucydides, iv. 74; and especially Bickerman, Chronology of the Ancient World, p. 32.
70 Mardonius was still in winter quarters (8. 131. 1). After consulting the oracles (8. 133–5), he dispatched Alexander to Athens (8. 136 f.; cf. 9. 7a. 1); at which point the Peloponnesians probably began work again.
71 e.g. Macan, i.2, 603, and Hignett, p. 282.
* Many friends lent aid in this project: Arthur Eckstein, Pericles Georges, Michael Maas, John Pollini, and in particular Professor Raphael Sealey-in this as in so much else. Though our differences of opinion are manifest, he has been a much needed source of advice and encouragement.
The original version of this paper was presented orally in the Herodotus Seminar at the University of California, Berkeley, offered in winter 1972, by Professor W. Kendrick Pritchett and benefited much by his comments. As I also acquired my reading knowledge of Herodotus in his undergraduati course of the previous year, it is perhaps especially fitting that I dedicate this article to him on the occasion of his formal retirement. It is hoped that Professor Pritchett wil enjoy continued years of good health and of interest in all things Greek.
Email your librarian or administrator to recommend adding this journal to your organisation's collection.
* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.
Usage data cannot currently be displayed