¹ There were over 5000 Spartiates of military age at the time of the battle of Plataia in 479 (Herod. 9.28), while in the life-or-death struggle at Mantinea sixty-one years later, there were only about 4200 Lakedaimonian troops present (Thuc. 5.68), of whom at most perhaps half — or 2000–2500 — were Spartiates (Thuc. 4.38 gives us a rough indication that typically less than half of a Lakedaimonian force in this era was Spartiate).

² The major modern chronologies which this paper will make reference to are those of The Athenian Tribute Lists, Vol. 1 by Meritt, B. D., Wade-Gery, H. T., and McGregor, M. F. (Cambridge, Mass., 1939)Google Scholar; A Historical Commentary on Thucydides, Vol. 1 by Gomme, A. W. (Oxford, 1945)Google Scholar; and Deane, P., Thucydides' Dates, 465–431 (Ontario, 1972)Google Scholar. In addition, less comprehensive chronological analyses of the Pentekontaetia have recently been presented in French, A., The Athenian Half-Century (Sydney, 1972)Google Scholar; Milton, M. P., Historia 28 (1979), 257–75Google Scholar; and Schreiner, J. H., Symbolae Osloenses 51 (1976), 19–63CrossRefGoogle Scholar and 52 (1977), 19–38.

³ Diodorus' chronological blunders — arising from his ill-considered (and very lazy) attempt to fit the topical narratives of his sources into the annalistic framework of his own history — are so notorious that we will illustrate them by a single example from this period. The single archon year of 471/0 is given over wholly to the narrative of the latter years of Themistokles, from the origins of his downfall because of association with Pausanias, through his travels around Greece and Molossia and his eventual flight to Persia, to his eventual death (which Plutarch, , Them. 31Google Scholar plausibly places in the late 450s) — a span of perhaps twenty years! (Diod. 11.54–9).

⁴ Thuc. 1.98–102.

⁵ Schreiner (1976) and (1977) is the only dissenter, but his highly unusual (and rather erratic) analysis is based on a fundamental disbelief in Thucydides' veracity (e.g. [1976], 37), which consequently forces him into a heavy reliance on Diodorus and other secondary sources, supplemented by his own guesswork. As might be expected, this leads to the creation of an almost unrecognisable chronology for the events of the Fifty Years.

⁶ A Babylonian legal text is dated to the months of Kislimu of the 21st year of Xerxes (465), which began on December 17, while news of Artaxerxes' accession had reached Egypt by 2 January 464. cf. White, M. E., ‘Some Agiad dates’, JHS 84 (1964), 140–52, especially p. 142 n. 13CrossRefGoogle Scholar.

⁷ Deane (1972), 10–11 convincingly argues the case for a short stay at Ephesos. Of our other chronologies, ATL iii.160 dates Naxos to 470, implying a stay of five years in Ephesos, while Gomme makes no attempt to date Naxos, saying that Artaxerxes' accession date ‘creates great difficulties’. In the most recent discussion of the problem, Milton (1979), 261–2 argues that a delay of three or four months is reasonable (while Themistokles' friends were sending him the cash), and, if interrupted by winter, eight or nine months; this is marginally possible. But when Milton attempts to salvage his own chronology by using this to argue that (Thucydides believed) the siege of Naxos ended as early as Autumn 466, his argument begins to break down completely. Xerxes died in December 465, and even if Themistokles had passed Naxos near the end of the 466 sailing season, this would still imply a stay of at least fifteen or sixteen months at Ephesos, violating Milton's own plausibility argument.

⁸ Paus. 4.24.5; Diod. 11.70.

⁹ First, it is possible that the Spartan political year in which Archidamos came to the throne would have been considered as the last year of the reign of his predecessor Leotychidas; and the slight displacement between the Spartan political year and the Athenian archon year might possibly account for another year of the discrepancy, if the succession took place during the overlap. However, it seems rather more likely that the 469/8 succession date - which is based on several questionable arguments - is in error.

¹⁰ Deane (1972), 11-12 is forced into this conclusion in order to accommodate Eurymedon in his chronology; all of the other main chronologies consider this possibility highly unlikely.

¹¹ Thucydides mentions the battle of the Eurymedon after he describes the surrender of Naxos; by the orthodox ‘strict chronology’ view, this corresponds to absolute chronological sequence, and even such an opponent of the orthodox case as Deane (1972), 12 accepts this argument without question.

¹² Plut. Kimon 12-14; Thuc. 1.100-101.

¹³ Plut. Kimon 14-15.

¹⁴ Aristotle, , Ath. Const. 25Google Scholar dates them to the archonship of Konon, namely 462/1, and since we would expect accurate dates for constitutional reforms from a constitutional history, this is almost universally accepted, even though the account involves Themistokles (long since ostracised by 462/1) as a leading participant.

¹⁵ I will temporarily beg the chicken-and-egg question of whether it was Ephialtes' reforms which forced Kimon to return from Sparta or Kimon's humiliating return which created the political climate leading to the reforms. Cf. Cole, J. R., GRBS 15 (1974), 369-85Google Scholar for a good recent (though partisan) discussion of the issue.

¹⁶ Even to place all of these events within a single year - as Deane is forced to do in his chronology - seems very improbable.

¹⁷ So Gomme (1945), i.390, seconded by Milton (1979), 258.

¹⁸ Sch. to Aesch. 2.31; Diod. 32.3. Actually, since Diodorus' account of the foundation of Amphipolis is contained in a single short sentence unconnected to the other events included in the year, it is probably based on his good chronological source and correct. Cf. Meiggs (1972), 453.

¹⁹ The scholiast lists a disaster which befell an Athenian attempt at colonisation (undoubtedly Drabeskos) in the Archonship of Lysikrates; this is apparently an error either for Lysistratos (467/6), Lysanias (466/5), or Lysitheos (465/4).

²⁰ Thuc. 1.100.

²¹ Milton (1979) attempts to evade this trap by arguing that (Thucydides believed) Themistokles passed Naxos in Autumn 466, but as discussed above in n. 7, this would imply a sojourn of fifteen or twenty months in Ephesos, which violates Milton's own standards of plausibility.

²² Thuc. 1.98–9.

²³ Under this interpretation, the μєτ⋯ τα⋯τα introducing the Eurymedon campaign would (ambiguously) be in reference to any of the Athenian actions described in 1.97 — perhaps the subjugation of Karystos — rather than the last events described in 1.97–8, namely the surrender of Naxos and the later subjugation of all Athens' remaining allies.

²⁴ Thuc. 1.100 claims that the Athenians captured or destroyed the entire Phoenician fleet of 200 ships. Ephorus somewhat more plausibly puts the total Persian fleet at 350 and the ships lost at 200, and along with several other secondary sources contains a much more detailed account of the campaign than Thucydides' single sentence; for these reasons, Thucydides' figures should not automatically be preferred, These accounts of the Eurymedon are contained in Plutarch, , Kimon 12–13Google Scholar and Diod. 11.60–62, though the latter conflates the Eurymedon campaign with Kimon's much later campaign in Cyprus, just prior to the Peace of Kallias.

²⁵ Diod. 11.69.

²⁶ Just as happened in 397 (Xen. Hell. 3.4.1).

²⁷ During the first year of the Egyptian expedition, there were 200 League ships in Egypt, and probably at least 100 or more deployed against Aigina (the Aiginetan fleet, renowned for its naval excellence, was heavily defeated in a great sea-battle, and over seventy Aiginetan vessels were captured): Thuc. 1.104–5. During the Samos revolt a decade or two later, Athens, Samos, Chios, and Lesbos deployed a total of 265 warships in the area (Thuc. 1.115–17).

²⁸ Cf. Deane (1972), 22–30 for a long discussion of the arguments involved, which strongly affirms the second possibility; Gomme (1945), i.401–11 and ATL (1949), iii.162–8Google Scholar present the contrary, more orthodox view.

²⁹ For example, French (1971), 46 n. 63 mentions the suggestion that the helot uprising began several years before the Thasian revolt and was merely intensified by the earthquake at Sparta. Gomme (1945), 407–8 lists several other versions contradicting Thucydides.

³⁰ Gomme (1945), i.404 argues for six years, ATL (1949), iii.163–4Google Scholar supports four years, and Lewis, D. M., Historia 2 (1954), 415–16Google Scholar holds out for five. Some of these suggestions as to how a corruption might have occurred are ingenious, but the very number of different suggestions weakens the strength of the argument, and the thorough refutation of Deane (1972), 22–30 is very convincing.

³¹ Diod. 11.64.4. According to him, the rebellion lasted ten years, while Thucydides says it lasted into its tenth year; but this is a very natural mistake for a third-hand source such as Diodorus to make, even if his number does ultimately derive from Thucydides.

³² Gomme (1945), i.403 very properly grasps this nettle, admitting that ‘all our other authorities, beginning with Ephorus, say that the war lasted nine years (or ten, as a round number)’.

It is important to realise that although accidental corruptions, numerical or otherwise, could occur at any time in history, they are much less likely to have occurred in classical times. This is because the chances of a textual corruption surviving uncorrected and propagating are inversely related to the number of independent copies of the manuscript in existence and circulating. Soon after Thucydides' history was ‘published’ (i.e. copied out in large numbers and generally circulated), every significant public or private library would have had its own copy of such an important and influential work. After this (unless the original scribe had made an error in all his copies), textual corruptions could not have widely entered the tradition again until the Middle Ages or later, when the destruction of the great library at Alexandria and most other major repositories of classical manuscripts had drastically reduced the number of surviving copies. For Ephorus' figure to be based on a corruption in Thucydides would compound improbabilities: we must both assume that Ephorus chanced to use a corrupt text and that this same text also became the basis for all our surviving manuscripts of Thucydides. This is extremely unlikely.

³³ The orthodox view, that Thucydides maintained strict chronological sequencing throughout the Pentekontaetia, is stated canonically (and provided with some justification) by Gomme (1945), 391–2, one of its strongest advocates: ‘we must see how far our other evidence…is reconcilable with Thucydides' narrative, keeping as our guide one fact, namely that he gives his events in strict chronological sequence as he conceives it’. Yet in the footnote to this very sentence, Gomme admits that ‘On the other hand, the fact (or the probability) that the excursus is an unfinished essay, allows a possible disruption of the chronological order, which Thucydides would later have put right. That is, we may admit a greater likelihood of error here than in his main narrative’. It is difficult to understand why Gomme's belief in an increased chance of chronological error is matched by his absolute disbelief in the possibility of chronological ambiguity (which, after all, one would naturally expect in an unfinished draft of an essay describing an interwoven pattern of events). Later, when Thucydides' account of the Five Years' Truce seems to violate Gomme's (sound) historical common sense, Gomme chooses again to postulate a fortuitous corruption of the text, rather than simple chronological ambiguity (p. 325).

³⁴ Diod. 11.84; Thuc. 1.108. Gomme (1945), 304 argues that since Thucydides' account of the periplous of Tolmides fails to mention Naupaktos, Diodorus' account ‘deserves no credence’. But Thucydides' many very significant omissions in his unfinished Pentekontaetia — which Gomme (1945), 365–70 himself lists at length and discusses — should make such an argument from silence a very weak one, especially about a town whose capture Thucydides had already referred to earlier. Furthermore, a scholiast to Pausanias 1.27.5 and Aeschines 2.75 adds the names of other towns captured, and Gomme apparently believes them (p. 320).

³⁵ For this and the following, cf. Plut. Kimon 15–18, Per. 9.3–10.5.

³⁶ The losses at Tanagra were heavy on both sides, and afterwards the outnumbered Peloponnesians retreated home. Sixty-two days later, the Athenians invaded Boiotia, were victorious at Oinophyta, and conquered Boiotia, Phokis, and Lokris; soon afterwards Aigina surrendered to Athens, and Tolmides raided the Peloponnese, burning the Spartan dockyards: Thuc. 1.108.

³⁷ Scholiast to Aristides 46.158.13.

³⁸ Aristides 46.158.13.

³⁹ Nepos, , Cimon 3.3Google Scholar.

⁴⁰ Andocides 3.3.

⁴¹ Diod. 11.86.1.

⁴² Cf. Gomme's admission, quoted in n. 33 above.

⁴³ Thuc. 1.108, 111.

⁴⁴ Ctesias, 36.

⁴⁵ French (1971), 60 n. 107 believes Ctesias' account on this point, and views Thucydides' five-year digression as an ‘understandable lapse from strict chronological reporting’, but he seems not to realise the extreme significance of this conclusion: a single demonstrated exception to an alleged absolute principle proves that the principle is not absolute, and greatly increases the likelihood that other exceptions may be found. Yet French equivocates on the Mount Ithome crux, weighing the unlikely possibility that Thucydides' dekatoi (confirmed by Diodorus) is a corruption against the ‘difficult assumption’ that Thucydides violated an ‘established…principle of narration’ by a chronological digression; and French does so without even referring to the execution of Inaros, which is almost an exact parallel, even in length of time.

Although Gomme (1945), i.321–2 discusses the discrepancies between Thucydides' account of the disaster in Egypt and that of Ctesias (which in many ways he prefers), he (uncharacteristically) never makes explicit the chronological divergences, and the strong implication of the Thucydidean digression.

⁴⁶ Thucydides' criticism of Hellanicus' earlier history of the Pentekontaetia (1.97.2) as being ‘not accurate’ (οὺκ ⋯κριβ⋯ς) — in contrast to his own — is often used to argue (even cited as unargued proof) that the Pentekontaetia will follow absolutely strict chronology (e.g. cf. Gomme [1945], i.361–2), but this is merely an implication; and even if it were true, Thucydides would presumably be contrasting the excursus as he intended it to be (when complete and polished) with Hellanicus' poor existing version.

⁴⁷ I follow Meiggs (1972), 103–8 and others in believing Thucydides when he repeatedly implies that virtually the entire Delian League expeditionary force of 200 ships remained in Egypt until the end, and was lost. If 200 or more ships — some Athenian and some allied — were lost, then it is likely that most ordinary Athenians, upon learning of the disaster, expected any week to see an enormous and victorious Persian fleet approach the Piraeus, just as they were later to feel after Sicily (Thuc. 8.1). These fears were unwarranted: the Athenian forces had been defeated by Persian engineering and land attack in one case and by surprise in the other; but the ‘paper tiger’ nature of Persian naval power would not have become apparent for some time — the bare fact was that a huge fraction of the standing Delian League fleet had been annihilated at a stroke.

The above reasoning is merely weakened, not destroyed, if losses in Egypt came to no more than 100 ships (as many scholars believe). Moreover, the foundations of this more orthodox reconstruction are rather implausible. Under it, we must assume that the Athenians kept no more than fifty or sixty ships (or less) in Egypt after the first year or so of the expedition (e.g. Gomme [1945], i.321–3). Greek naval excellence was high, and the Athenians were notoriously reckless, but even they would have balked at the notion of leaving such a small fleet to maintain continual control of the seas — an absolute necessity — so far from home and so close to Phoenicia, Persia's principal naval centre. Callisthenes (Plut. Kimon 13.4–5) viewed Perikles' short sail to the south-western coast of Asia Minor with a mere fifty ships as an act of notable daring; how would he have viewed the idea of keeping such a small expeditionary force within a few days' sail of Phoenicia for five or six years, winter and summer?

⁴⁸ The eight years between the apparent date of the reforms of Ephialtes (462/1) and the date of the Five Years' Truce (454/3) do present some difficulty if we are to believe that Kimon returned in the fifth year of his ostracism; however, they can be accommodated. If Kimon had departed for Sparta in early Spring 461, and his absence (along with several thousand members of the more conservative Athenian hoplite class) allowed Ephialtes quickly to gain political ascendancy, then the reforms could have occurred near the end of the 462/1 Athenian political year. Kimon may have remained in Sparta for seven or eight months, hoping to achieve a victory and return in triumph to Athens (with a consequently greater chance of reversing Ephialtes' constitutional changes); the accounts of our sources are vague enough to leave open the possibility that he remained in Sparta over the first winter. Therefore, his ‘dismissal’ (whatever the circumstances) and return to Athens may well have come too late for the first round of the ostracism vote of 461/0, coming in the sixth prytany (some time in December or January), leaving him to spend the next twelve months carrying the burden of failure at Ithome and unsuccessfully attempting to overturn Ephialtes' reforms (Plut. Kimon 15). His ostracism might have come the next year, with the preliminary vote in the sixth prytany and the final decision (Kimon or Ephialtes) in the eighth (Arist. Ath. Const. 43.5; Philochorus, , FGrHist 328 F 30Google Scholar). The second vote would have forced Kimon to leave Attica within ten days, putting his departure within a few weeks of the end of the 460/59 Athenian political year; for this reason, 459/8 might have been traditionally recorded as the first (full) year of his ostracism. This would place his recall in the political year 455/4, the fifth of his exile; if he had returned in early summer 454, it might have taken him a month or two to arrange the Five Years' Truce with Sparta, placing it in the next political year 454/3 as required. A number of assumptions are required to produce the above chain of reasoning, but all are rather plausible. This reconstruction would also explain how Kimon returned from Mount Ithome apparently some time before the beginning of the Egyptian expedition, yet returned to Athens in the fifth year of his exile but some time after the end of that six-year expedition.

⁴⁹ A cynic might argue that Sparta's leaders would never have granted Athens such a respite, despite their friendship with Kimon and their public commitments to pan-Hellenic solidarity against the barbarian; and such suspicion is very warranted. But in the above chronology, the long helot rebellion would have come to an end in its tenth year — 455/4 — perhaps only five or six months before, and exhausted Sparta could herself use a breathing space free from major military threat. It is also possible that Sparta hoped to use the five years to deal a crushing blow to Argive power, and this may have resulted in the thirty-year peace treaty which Argos signed in 451/0 (see text below).

⁵⁰ Cf. n. 47 above.

⁵¹ Thuc. 1.111.

⁵² Meiggs (1972), 110–11 and n. 1 understandably finds these adventuristic enterprises highly peculiar in the aftermath of Egypt and he even ventures to speculate that they ‘may even have come a little earlier than the capitulation on Prosopitis’, adding in a footnote that ‘Thucydides seems not to have been rigorously chronological in his account of the Egyptian expedition’ (!); but this extremely important idea is never carried to its logical conclusion or extended to other chronological cruxes.

⁵³ The only previous time that Argos had considered such a lengthy peace treaty with Sparta was after her manpower had been massacred by Kleomenes (Herod. 6.79–83, 7.149–50).

⁵⁴ Arist. Ath. Const. 26.3 dates the decree to the archonship of Antidotos, namely 451/0.

⁵⁵ For discussions of the dating and nature of the Strasbourg Papyrus (which contains a scholiast's commentary on Demosthenes 22.13f., mentioning the expenditures), cf. ATL (1949), iii.281Google Scholar and Meiggs (1972), 515–18, both of which support the 450/49 dating, and Gomme (1956), ii.28–33, who questions these arguments. The main restorations of the decree are conveniently collected and translated in Fornara, C. W., Archaic Times to the End of the Peloponnesian War (Cambridge, 1983), 95–7CrossRefGoogle Scholar.

⁵⁶ ATL (1949), iii.281Google Scholar places the beginning of Kimon's campaign in Spring 450, and the Peace df Kallias a year later in the first half of 449; Deane (1972), 61–3 and French (1971), 64 n. 119 and 65 n. 123 both support this, the former discussing the various other ideas proposed. Gomme (1945), i.409–10 displaces both dates by a year, putting the Cyprus campaign in Summer 449 and the Peace in Spring 448 (as is noted in the preceding footnote, Gomme refuses to credit the usual date of the Strasbourg Papyrus). Meiggs (1972), 124–8 recognises the ‘crowding’ which such chronologies would require, and much more consistently places the Cyprus campaign in 451 and the Peace of Kallias in 449.

⁵⁷ So Meiggs (1972), 124–8.

⁵⁸ Under my reconstruction, the missing tribute quota list of 449/8 should not be explained by any (rather improbable) remission of tribute in the year following the Peace, but as a consequence of the Periklean building programme begun a couple of years later and paid for by the allied tribute. Since all incoming tribute of that year was directly used to finance building construction, there was no surplus phoros to be stored with the Goddess, and hence no tribute quota was paid. For my arguments on the ‘surplus’ nature of the tribute, cf. R. K. Unz, ‘The surplus of the Athenian phoros’ (to appear in GRBS).

⁵⁹ Plut. Kimon 4.1. The context makes it clear that the claim is not simply a standard political slander.

⁶⁰ Plut. Per. 37.

⁶¹ Schreiner (1976), 26 and n. 17.

⁶² Diodorus' evidence appeals to Gomme (1945), i.325–6, and he considers suggesting another textual corruption in Thucydides to reconcile it, but eventually decides to follow the bulk of the scholarship and reject the date. Meiggs (1972), 453 points to this date of Diodorus as one of a number of such dates which Diodorus probably obtained from a chronological handbook; in each case, the dating is brief — usually just a short sentence — and stands outside the flow (or rather the whirlpool) of Diodorus' main narrative (though naturally it is sometimes difficult to judge whether some borderline cases fall into this category or not). Meiggs says that dates such as these ‘are more likely to be right than wrong’ (which is something of an understatement, since each of the three or four clear-cut cases which we can independently check turns out to be exactly correct); however, Meiggs asserts that we should flatly reject the 454/3 date for the Five Years' Truce (where the signs of Diodorus' having used a chronological source are perhaps strongest of all). Meiggs' obvious reason for this inconsistency is to safeguard his own chronological framework.

⁶³ Cf. Meiggs (1972), 129–51 for a lengthy and thorough discussion of the likelihood and terms of the Peace of Kallias, which firmly concludes that the Peace did in fact exist. Thucydides' silence is seen as strange and embarrassing, and one of the strongest arguments in the hands of doubters of the Peace. Meiggs bravely admits that ‘no convincing explanation has been given’ for Thucydides' omission of the Peace and that ‘the best that can be offered…is a palliative’ (p.140): the unfinished nature of the Pentekontaetia (and indeed the entire history) is the principal one cited. Gomme (1945), 331–5, 364–70 had roughly the same opinion.

I wish to acknowledge the helpful comments of E. Badian, John Barron, and an anonymous referee on an earlier draft of this paper, though naturally none of them should be held responsible for such errors as remain, nor for the views presented. I also wish to thank Harvard University, the Winston Churchill Foundation, and the National Science Foundation for their financial support during the preparation of this paper.