¹ Cf. 101–8, 439–41 and 473–7. For earlier discussions cf. Salmon, P.La Politique Égyptienne d'Athènes (Brussels 1965) 152 n. 6Google Scholar.

² I am indebted to Sir Kenneth Dover for reading a first draft of this piece and making searching and valuable comments which have led me to re-think or re-phrase several crucial points. He should not, of course, be saddled with responsibility for this final version.

³ Ktesias 63–7.

⁴ Notably by Westlake, H. D., Essays on the Greek historians and Creek history (Manchester 1969) 66Google Scholar; cf. also Adcock, F. E., PCPS cxxxiii–v (1926) 4Google Scholar.

⁵ Hdt. viii 60, Thuc. vii 36.4

⁶ Dover observes that any admiral would have to be sure that he would have early warning of any such enemy enterprise before taking his fleet up-river. This is no doubt true, but it is not clear what sort of warning could be expected. If Persia would have to muster a special fleet for this purpose then ample warning could be expected. But if there was a standing force available in Phoenicia and it sailed without an accompanying land force, then news of its departure could hardly reach Memphis in time for precautionary action—in the absence of a system of fire-beacons.

If the Greek fleet at Memphis was small, this would not be too serious since it could rely on the larger fleet which was on the open seas to keep watch for any such development and come to the rescue: but if the main fleet was at Memphis and news failed to reach it in time, this could have been disastrous indeed.

⁷ The main Persian fleet was, of course, on the open sea near Cyprus, not wasting itself senselessly in the Nile.

⁸ Thuc. i 105.2.

⁹ Hdt. viii 46.1.

¹⁰ Hdt. viii 43.

¹¹ Libourel 607 argued that the Peloponnesians would not have dared to confront the Athenians if they and their allies had only left 50–60 ships in Egypt. There seem to be serious flaws in this argument:

(i) He estimated the Athenian fleet at 300 which seems too high at this date, as Meiggs implicitly agrees, so it was not as overwhelming as he suggests.

(ii) More important is that Athens was the aggressor anyway and Aegina had no choice but to resist, whilst Corinth could not accept the drastic change in the balance of seapower if Athens gained control of Aegina. Corinth's willingness to undergo great danger in this cause is shown by her invasion of the Megarid by land shortly afterwards, which led to disaster (Thuc. i 105–6).

¹² The size of the Athenian citizen-body implied here needs justification. The basic source is Thuc. ii 13–6 where hoplites (including metics) are put at 29,000. Busolt (Gr. Ges. iii 2.884 ff.), after allowing for metics and adding the two upper classes, puts the citizens excluding thetes at c. 22,000–25,000. Ed. Meyer preferred c. 35,000 and Beloch c. 20,000, but Busolt showed the objections to both. Subsequently scholars have generally followed Busolt's figures, though they have explained them in varying ways: cf. Gomme, HCT ii ad. loc. and Jones, A. H. M., Athenian democracy (Oxford 1957) 161–77Google Scholar.

All scholars, including Meyer and Beloch, accept a thetic population of c. 20,000 (cf. Ar. Vesp. 709), so the estimates for the total citizen-body only vary between c. 42,000 and 45,000: this is not significant for present purposes. The figure at the time of the Egyptian Expedition was certainly not larger than this—probably a little smaller, since in 500 it is referred to by Herodotus v 97 and viii 65 as 30,000 in c. 500 bc.

¹³ This is well discussed by Amit, M., Athens and the sea (Brussels 1965) 30–49Google Scholar.

¹⁴ Dr N. L. Young of the Imperial War Museum Research Dept. informs me that in the 1914–18 war the best estimate for the percentage of deaths among mobilized troops was: France 16 per cent, Germany 16 per cent, Russia 14 per cent. It must be remembered that these are percentages of mobilized men and not of total adult male population, so the percentage of losses would be much lower in relation to the total adult male population. Only in the Second World War was there a much higher figure—the losses of the Russians (again only mobilized Russians) estimated at between 30 per cent and 37·5 per cent.

¹⁵ Hdt. viii 94 (Corinth—but Herodotus shows doubt), ix 80.3 (Aegina). As the Mendesian branch is so far east, near the stronghold that guarded the eastern approach to Egypt, the relief force would be running great risks: and even if they had succeeded in passing up it they would still not have reached Prosopitis itself, only the Sebennytic branch about 10 miles to the north or south of the island, as the map shows. If the conjectural position of the waterway on the map should be questioned and it be suggested that the island might have reached up to the point where the Mendesian branch diverges from the Sebennytic—the answer is that this seems impossible. The point of divergence is c. 30 miles from the southern point of the island and the gap between the Canopic and Sebennytic branches is very wide at this point (30 miles): if the length of the Canopic bank back to the southern point of the island is added this involves another c. 30 miles making a total circuit of 90 miles which is much too much. Herodotus ii.41.5 puts the circuit at 9 schoinoi which Meiggs assesses at under 60 miles. Nor can one try to reduce the 90 mile circuit by postulating a very lopsided island as shown in the accompanying sketch since no one would make a cutting from branch to branch of the Nile except by the shortest possible route. The shape of the island must therefore have been roughly isosceles.

¹⁶ Westlake (n. 4) 71 n. 35 suggests that the arrival of the relief force at the mouth of the Mendesian branch of the Nile, which is not one of the three main branches, may suggest that it was trying to slip past enemy positions: but this fails to explain why so small a force was sent, if the danger was known, or how it could be expected to get through to Prosopitis if the land was in enemy hands.

¹⁷ Westlake 67–8 is also impressed by the language but attributes it to ignorance and emotionalism (73). He also fails to add allied ships to the 40 Athenian ships in the main force.

¹⁸ Although Thucydides' account of the Sicilian Disaster clearly possesses many aspects of language and thought in common with Greek Tragedy, as has been well shown by Colin Macleod in his Collected Essays (Oxford 1983) 140–58Google Scholar, there is no suggestion by him that Thucydides was seduced by the muse of Tragedy from his loyalty to the muse of History. Macleod notes that the tragic construction of the history and interpretation of events was not at all contrary to Thucydides' aims as a historian. Thus the heightened emotion of the writing should not breed scepticism about his facts and figures: Dover in HCT iv shows no such tendency.

Nor do I accept Westlake's suggestion (73) that Thucydides was quite unable to get reliable information about casualities in 454, since it was so recent and casualty-lists should in any case have been available: nor can I believe that his treatment of this topic was affected by his profound shock as a six-year-old.