² Busolt, , Gr. Gesch. ii.²672Google Scholar, n. 4, ‘glaub lich’; Hauvette, A., Hérodote, 313Google Scholar; Nöldeke, Th., Aufsätze zur persischen Geschichte, 44Google Scholar; Bauer, A. in Jahresh. vol. iv. (1901), p. 94Google Scholar, very emphatic; Dr.Grundy, G. B., The Great Persian War, 219Google Scholar, ‘no solid grounds for doubting it’; Raase, H., Die Schlacht bei Salamis (1904)Google Scholar; to name only the most recent. It is curious to see how Raase's really learned pamphlet ignores Delbrück and Meyer, and still talks of the Greeks not being heavily outnumbered at Salamis, only by some 300 ships! In fact, the authentic fleets of as many as 300 in antiquity can almost be numbered on one hand. [Dr. Macan gives 1,200, divided (arbitrarily) into three squadrons of 400 each, but suspects there may be some exaggeration.]

³ Gesch. d. Kriegskunst, vol. i. p. 70: cf. pp. 76, 78.

⁴ Zur Gesch. d. Perserkriege (Neue Jahrbücher für Philologie und Pädagogik, 145, 1892, p. 158).

⁵ Griech. Gesch. i. 368.

⁶ Hist. of Greece, i.² 287.

⁷ J.H.S. xxii. (1902), pp. 294, 300.

⁸ Gesch. d, Alterthums, iii. § 217.

⁹ I assume that Prof. U. von Wilamowitz-Moellendorff has sufficiently shown that the account of Salamis in the Persae of Timotheos is merely a sea-fight at large of Timotheos' own time, whatever corrections may ultimately be made in interpretation of details.

¹⁰ I do not know what this 207 means. One is familiar in the later Athenian navy with ships reckoned as first class, ἐξαίρετοι but for a fleet in large part newly built, 207 such is a highly improbable number; cf. n. 62.

¹¹ Plut., Cimon, 12Google Scholar.

¹² Several writers—e.g. Busolt, ii.² 694, n. 6; Welzhofer, , Die Seeschlacht bei Salamis (Hist. Taschenbuch, 1892, p. 48)Google Scholar; Meyer, , G. d. A. iii. § 217Google Scholar; Munro, l.c. p. 299 Lehmann-Haupt, C. F., Klio, vol. ii. (1892), p. 338, n. 2Google Scholar; [and Macan on H. 8, 66]—accuse Herodotus of raising his figure for the fleet again after the storm to its original strength by supposing that reinforcements from the islands, etc., balanced the losses. Fortunately, he never said, anything so foolish. What he does say (8, 66) is that Xerxes' men, both those that marched overland and those who came on shipboard, were as numerous at Phalerum as before Thermopylae; for the losses of men in the storms, at Artemisium, and at Thermopylae, were balanced by reinforcements. There is not a word about ships. The Boeotians turned out πανστρατιᾷ except the men of Plataea and Thespiae; if we reckon them at 8,000–10,000, the latter being one half of their total levy at Delium (see Beloch, , Griech. Aufgebote ii. in Klio, vi. 1906, p. 35)Google Scholar, and add another 2,000 for the Malians, Dorians, Locrians, and islanders, then H.'s statement is sobriety itself, provided that (as regards the fleet) he is reckoning the loss in fighting men only and not in rowers, i.e. the loss as it affected the Persian army, of which the Persian marines formed part.

¹³ Aeschylus gives as total 1,000 ships, and later on a division of 250 (Pers. 323); it looks as if we had another allusion here to the four divisions.

¹⁴ If Megabazos' father be the Megabates of H. 5, 32, he was a collateral of the royal house. It does not appear if Prexaspes was related to the well-known Prexaspes of Cambyses' reign.

¹⁵ One of one's difficulties is the constant use of ‘Phoenician’ for a Persian fleet generally. See, e.g., for Herodotus, the proceedings of that fleet after Lade; for Thucydides, 1, 100 (the Eurymedon campaign).

¹⁶ The total of the Ionian and northern fleets is 360, i.e. the 353 of Lade in round figures. Most of the exaggeration falls on the (less known) Asiatic contingents. [Dr. Macan treats H.'s navy-list as substantially correct, but has no new reasons.]

¹⁷ Hauvette, , Hérodote 314Google Scholar, justly points out that the expense of provisioning the army must have precluded the towns of Thrace and Chalcidice from doing much else. They also furnished land troops.

¹⁸ Diodorus has an extraordinary figure here. His total for the first four fleets corresponds with that of Herodotus, though he makes the Ionian fleet 20 larger, the central 20 smaller, than does the latter, But Aeolis and the Hellespont do not correspond; H. gives 160 for the two, Diodorus 120. D. then tacks the surplus on to the islands. I draw no deductions from this: but see § 9. I see, however, little to warrant the conjecture of von Mess, A., Untersuchungen über Ephoros (Rhein. Mus. 1906, vol. 61, pp. 360, 399)Google Scholar, that Ephorus here used, in addition to Herodotus, a (supposed) navy list of Ctesias giving a total of 1,000 ships, and consequently smaller separate contingents. See also n. 117.

¹⁹ No other power in antiquity ever collected a fleet of 600 warships. Octavian may have controlled 500, partly borrowed from Antony, and Organised as two distinct fleets in different seas, at the beginning of the campaign which ended with Naulochos. In that year, 36 B.C., there were about 1,000 ships in commission in the whole Mediterranean. In 480, apart from the Greek and Persian fleets, totalling together almost 1,000, we have those of Corcyra, Carthage, Syracuse, Etruria, Marseilles. If we take Kromayer's view, that in the civil wars the fleets, reckoning in quinqueremes and Liburnians, came out at about the average power of a fleet of triremes of the same total, we must rank the total sea-power of the early part of the fifth century extraordinarily high. It seems possible, however, that the zenith of Mediterranean sea-power would have to be placed about 260–250 B.C.

²⁰ See Prášek, J. V., Gesch. der Meder und Perser, i. 215Google Scholar.

²¹ See Prášek, op. cit. 223, 239. If the Phoenician terms were as I suggest, 120 pentekontors must have been the force contemplated. Doubtless the extension of the meaning of these terms, however worded, so as to apply to triremes, would be one of those measures of reorganisation which earned for Darius his nickname ὁ κἀπηλος We can see that the division between the northern and Ionian fleets must correspond to that between the satrapies of Daskyleion and Sardis, whatever it was.

²² That the Greeks dedicated Phoenician triremes after Salamis is conclusive as to their opinion.

²³ I mean, if he had a military command at all. (Egypt sent no land troops.) I am not expressing an opinion on the controversy whether, in the ordinary way, the satraps had the military command.

²⁴ The Greeks of a later time were much perplexed over the Persian command, and felt it necessary to manufacture a single admiral for the fleet; so Megabates (Diod. 11, 12), perhaps meant for the father of Megabazos; and Plutarch's, Ariamenes (Them. 14)Google Scholar, who appears to be a conflation of Ariabignes and Achaemeues. See on these names Marquart, , Untersuchungen zur Gesch. von Eran (Philol. 54), 499–502Google Scholar. It is hardly worth mentioning that Ctesias has the same error.

²⁵ A fine field for speculation can be opened up if one treats the jealousy as really existing between Phoenicians and Cariane, and going back to the ‘dark ages’ when they may have fought over the relics of Minoan sea-power. We find the Phoenician circumnavigation of Africa matched by that of Western Asia under the Carian Skylax; and now we have another Carian, Heraclides of Mylasa (see § 4), teaching men how to meet the Phoenician diecplus. Naturally, the duel between Phoenicia and Themistocles ended in the latter acquiring a Carian mother (Plut., Them. 1)Google Scholar; and there may be a lot of other material of the sort to be collected. Doubtless the Phoenician version of Salamis dealt very faithfully with the Creto-Carian Artemisia.

²⁶ Mr.Hall, H. R., The Oldest Civilisation of Greece, 88Google Scholar; Prof.Hommel, F., Grundriss d. Geog. u. Gesch. d. alten Orients, i. 57, 58Google Scholar.

²⁷ [As Dr. Macan thinks there were native epibatae throughout the fleet, I must give my reasons for this statement. The navy-list (7, 96) says that all the marines were Persians, Medes and Sacae. Persian epibatae on a Sidonian ship (7,181 compared with 8, 92). This is again borne out by 8,130; see p. 226 post. But 7, 184 (the chapter of the great exaggerations) refers to native as well as Persian, etc. epibatae. One might discard this as an obvious means of working up a large figure; but we hear of Egyptian epibatae (9, 32), heavy-armed troops (7, 89). To my mind, two sets of epibatae on one ship are impossible; the ships of this epoch did not carry, probably could not carry, many epibatae. I can only conclude that four fleets carried Persians, etc., and the Egyptian fleet natives. I do not say that the four fleets carried no native epibatae; but if they did, these were few and unimportant. On the contrary, the Egyptian marines were a substantial body, or Mardonius would hardly have landed them: ergo, there can have been little or no room for Persian marines in the Egyptian fleet. It will be seen, I hope, that this fits the story extremely well.] Now thirty epibatae to each trireme is too high. Meyer properly cuts down the rowers to 150, and twenty is ample for the epibatae; the Greek ships, if we like to follow Plutarch, carried eighteen, but the regular Athenian number later was ten. Four hundred and eighty ships at twenty epibatae each = 9,600 men, or with officers say a round 10,000. I cannot help suspecting that the total Persian army on mobilisation was not 360,000 in six corps of 60,000, but 60,000 in six corps of 10,000, one complete corps being assigned to the fleet. [Dr. Macan does not see why H. should give the armament of each of the nations that contributed to the fleet unless they sent epibatae. But on the analogy of any other fleet, e.g. the Roman, the rowers must have had their arms with them; and this is expressly stated of the Samians, 9, 99.]

²⁸ H. 8, 10. The Greek ships were heavy by comparison, 8, 60. Plutarch, (Them. 14)Google Scholar says the Persian ships were tall, with lofty poops, compared with the Greek ships, which were much lower in the water. It is a pity that theories have been built on this, for it is mere moralising, like his similar statement about Aotium; the just cause must have the smaller ships. The galleys on the fourth-century coins of Sidon and Aradus are not in the least like Plutarch's description; and his reference to Ariamenes fighting ὤσπερ ἀπὀ τείχους shows that what he has in his mind is not the fifth century at all, but the τειχομαχία of the first century.

²⁹ [Macan reads τῶν πεντηκοντέρων καὶ τριηρέων but this last word is merely an emendation. It is not very material.]

³⁰ Thuc. 6, 42; 134 triremes and two pentekontors to 131 supply and service ships; many volunteer merchantmen also accompanied the fleet for the sake of trading. This last may be true of Xerxes' fleet also.

³¹ If we like to assign eighty to each fleet, we get, not only Aeschylus' 1,000, but the 200 ships per squadron so common in H. and later writers.

³² B.M.C. Caria, Introduction.

³³ Of the ancient writers, Strabo 9, 443 is best, though he complains that he could not get information. The modern authorities are given by Mr.Wace, A. J. B. in J.H.S. 26 (1906), p. 143Google Scholar, The Topography of Pelion and Magnesia; and I am much indebted to him for further information as to this coast-line, and some references, which he most kindly sent me in reply to some questions. The accompanying map has been drawn by Mr. F. Anderson from Admiralty chart no. 1,085, reduced to ¼ scale, with some alterations in the way of names for which I am responsible. It has not been possible on the reduced scale to indicate the little beaches in the manner done in the chart itself.

³⁴ Researches in the Highlands of Turkey, ii. 104.

³⁵ Θεσσαλία first edition (1880), pp. 213, 218. I regret that I have been unable to see the second edition, so my quotations must stand subject to correction.

³⁶ Κασθαναίας κώμης ὑπὸ τῷ Πηλίῳ κειμένης

³⁷ Mr. Tozer states that the learned men of Zagora claimed that that place was Casthanaea, and supported their claim ‘by the abundance of chestnut trees in that neighbourhood, while there are none near Keramidhi.’ According to Georgiades, Zagora is the most important place in the neighbourhood.

³⁸ J.H.S. 26, 146. If C. Sepias had been Kato Georgi, why did not the Persians put to sea and run round the corner, out of the wind? I fancy that with a gale blowing on shore this would be easier said than done with galleys; however, I hope this paper will answer the question; the fleets were strung out in detachments at least as far north as Thanátu (Meliboea). This leaves only a passage from Apollonius Rhodius, an unsatisfactory passage (see Georgiades) in an unsatisfactory geographer, and it is only a deduction at that. The natural view is certainly that of Bursian, , Geog. von Griechenland i. 99Google Scholar; C. Pori is Strabo's Ipni, τόπον τραχὺν τῶν περὶ τὸ Πήλιον If we make Pori, Sepias, and Ipni, Venéto (Georgiades), then the heel of Magnesia is left nameless both by H. and Strabo, which seems unlikely. Mr. Wace proposes Myrae; but surely Mézières' identification of Myrae with Mouresi is, in the absence of inscriptions, sufficiently probable.

³⁹ Mr. Wace states (l.c. 147) that north of Kato Georgi at least as far as Zagora there is no beach at all to accommodate a fleet, and uses this as an argument for Sepias being C. Pori. But, whereas there are some little beaches south of C. Pori, there is absolutely nothing between C. Pori and Keramidhi (see Bursian, l.c. i. 99); so the argument is at least double-edged. It will be seen that Mr. Waee's premises, which I fully accept, seem to me to necessitate a very different conclusion.

⁴⁰ I did not know when I came to this conclusion that Georgiades (l.c. p. 213) had said the same thing twenty-eight years ago. He thought that the Persian fleet was strung out at all the little harbours below Zagora, Kissos, etc. It is strange that no one has followed up this very just conclusion. [Dr. Macan says that the αἰγιαλός is defined in H. 7, 188, 2 as ‘extending from Kasthanaia to Sepias.’ Can μεταξύ bear this meaning? Anyhow the αἰγιαλός is conceived as small, 7, 188, 5 and 15.]

⁴¹ Aristarchus ad Il. Ξ 34 explains this as κλιμακηφὸν νενεωλκημέναι ὤστε θεατροειδὲς φαίνεσθαι which Dr. Leaf explains as en échelon, each projecting somewhat beyond the other, like the steps of a staircase. I take this to mean that, in Aristarchus' opinion, the sterns of row two would be between the prows of row one, and so on, to save as much space as possible. Homer is certainly describing some method of getting more ships ashore than the shore would hold in the ordinary way, as the context shows. This too seems what Hesychius means by ἐπάλληλοι Stein, however (H. 7, 188), explains πρόκροσσαι as parallel files of ships, eight deep, each file perpendicular to the line of coast. I prefer Aristarchus myself, as Stein's explanation would hardly increase the number of ships ashore; but if I am right in what follows, it is not very material.

⁴² This follows from their dispositions at Therme. But even the first Athenian expedition to Syracuse, 136 warships and about as many supply ships, sailed in three separate divisions.

⁴³ Herod. 7, 188, ἐξ αἰθρίης τϵ καὶ νηνϵμίης τῆς θαλάσσης ζϵσάσης: Medit. Pilot, vol. 4, 1900Google Scholar, under ‘winds’; the north wind blows with much force, even in summer. Summer gales are almost always preceded by calms with a dark appearance round the horizon.

⁴⁴ Great Pers. War. p. 327, n. We have little real evidence of the pace of triremes: and even so, single ship voyages are no evidence for a fleet, tied to its slowest member, and moving at an economical rate, i.e. using its rowers in relays of one-third at a time. Bauer has frequently and justly pointed this out. We rarely know the conditions of any recorded voyage, or even if the sails were being used. A lot of such evidence as exists is given by Droysen, in Hermann's, Lehrbuch, ii.³2, 302Google Scholar; the best is Xen., Hell. i. 1, 13Google Scholar (on which Bauer relies in his account of Salamis), Alcibiades with eighty-six ships, going fifty kilom., takes all night in late autumn and up to ἄριστον some eighteen hours. Xenophon was at least a practical man, who knew what a trireme meant. In allowing for twelve hours' rowing, we must remember that much time would be lost over launching the fleet, dinner, anchoring, or drawing ashore again.

⁴⁵ Stein justly remarks, ‘Die ganze Stelle ist unter dem Vorbilde von II. ξ 33 ff. geschrieben,’ but unfortunately goes on to say that H. interprets Homer.

⁴⁶ Welzhofer, , Neue Jahrb. f. Phil. und Päd., 145, p. 660Google Scholar, rightly discredits this ἔρκος Is it perhaps a real reminiscence of using wreckage to make a breakwater?

⁴⁷ Themistocles' explicit appeal to the Ionians and Carians (8, 19 and 22) quite precludes the idea that any other large body of Greeks was still with the fleet. Neither is it possible that the northern fleet never sailed at all, but remained at the Hellespont; the story presupposes that the bridges were not guarded, and it does not appear (as it would have to) either at Mycale (where the number of Persian στρατηγοί is conclusive: see post) or after. Neither can it be hidden under the term ‘Ionians’; for elsewhere H. is precise: 4, 89, the Scythian expedition, τὸ ναυτικὸν ἧγον ᾿Ιωνές τε καὶ Αίολέες καί ῾Ελλησπόντιοι 6, 98, Datis to Eretria ἀγόμενος καὶ ᾿Ιωνας καὶ Αἰολέας

⁴⁸ It is certain that the Persians, after elaborately organising their fleet, would not proceed to disorganise it by picking out the ships to go round Euboea ‘from all the ships’ (8, 7). A definite squadron, accustomed to work together, was sent. It meant something, I suppose, even to bring 120 ships to anchor without collisions: see Thuc. 6, 42 on the anchor drill of the Athenians before sailing for Syracuse, ξὐνταχιν ὤσπερ ἔμελλον ὁρμιεῖσθαι . . .οί στρατηγοὶ ἐποιήσαντο

⁴⁹ B.S.A. ii. 83. In his history, Prof. Bury sends these ships off from Aphetae. Has he abandoned his earlier view [which Dr. Macan has adopted]?

⁵⁰ Bury in B.S.A. ii. and Munro, l.c. p. 310. Note that in 8, 66 H. knows only of ‘the storm’; he must have had two versions at least before him. Mülder, D., Klio, vol. 7 (1907), 29CrossRefGoogle Scholar, treats the whole storm-incident. as a duplicate of the storm that destroyed Mardonius' ships at Athos in 492. If I am right about the fleets, this is impossible. I note that the Mediterranean Pilot, in its Athens table (the nearest), gives an average of three days' gale for August, more than for any month but January and February. [Dr. Macan treats the two storms as certainly one, lasting for three days.]

⁵¹ H. 8, 4: ἐπεὶ αὐτοῖσι παρὰ ὀόξαν τὰ πρήγματα τῶν βαρβἁρων ἀπέβαινε ἤ ὡς αὐτοὶ κατεδόκεον

⁵² H. 8, 22: εἰ . . . ὑπ᾿ ἀναγκαίης μέζονος κατέζευχθε ἤ ὤστε ἀπίστασθαι

⁵³ Hermes 41 (1906), p. 103.

⁵⁴ Philol. 61, p. 352.

⁵⁵ Hermes 42 (1907), p. 512. But for the name Artemisium, it would fit in well enough with the battle off Cyprus in H. 5, 112, in which the Ionians defeated the Phoenicians, for there must be something behind H.’s statement that that day the Ionians were ‘at the top of their form,’ ἄκροι γενόμενοι Having learnt how to meet the diecplus, they then, before Lade, try to practise it themselves.—But though there were many Artemisiums and Dianiums all about the Mediterranean, I cannot find one in these particular waters, or nearer than the one in Caria which Ruehl gives.

⁵⁶ Welzhofer (l.c.), in his excellent study of Artemisium, came to much the same conclusion: the Greeks overwhelmed a portion of the Persian fleet before the rest came up. Ephorus perhaps liad the same idea, but Piodorus does not actually say so, though he comes rather near it: 11, 12, τῶν δὲ βαρβάρων ἐκ πολλῶν λιμένων ἀναγομένων (before we have ἐκ πολλῶν καὶ διεστηκότων λιμένων), τὸ μὲν πρῶτον οἱ περὶ τὸν Θεμιστοκλέα διεσπαρμένοις τοῖς Πέρσαις αυμπλεκόμενοι πολλὰς μὲν ναῦς κατέδυσαν κ.τ.λ.

⁵⁷ This now seems a fixed point; Lenschau, Th., Jahresb. über gr. Gesch. 1904, p. 195Google Scholar. [Macan ii. 261 and 270.]

⁵⁸ By no means the same as the Corinthian tactics against Phormio in the gulf of Corinth. The line would probably become an arc, as they would be overlapped.

⁵⁹ [Dr. Macan's view is, that when the Persians rounded C. Sepias the Greeks were holding the Oreos channel, in case the enemy should try to force it; the Greeks did not attack the main Persian fleet as it made for Aphetae, but managed to cut off the rear-guard under Sandeces, capturing according to the Asianic version fifteen ships, according to the Greek thirty; this was the first day of Artemisium. This is a wide departure from the tradition; nor do I see how ships of Paphos and of Caria could really be in one squadron. But I have already dealt with the Sandoces story, and cannot think that it has anything to do with the first day of the battle of Artemisium.]

⁶⁰ I.e. that on both days the Sidonians did best. See § 9.

⁶¹ Bury, in B.S.A. ii. 83Google Scholar.

⁶² A consideration quite neglected by those writers who seem to look on every number as suspect unless it be a surd. Given a town with a large fleet, this was bound, when at paper strength, to be an easily subdivided or round number. How far subdivision went we do not know: but there is an interesting story in Polyaenus iii. 4, 2 of Phormio manœuvring a fleet in small squadrons of five ships each (πεντανατα) as units; which shows (whether true of Phormio or not) that at a later time the writers of the ordinary books on naval tactics were familiar with the idea of handling a fleet in small sub-squadrons.

⁶³ The glamour of Thucydides must not blind us to the fact that those tactics of manœuvre which we associate with Phormio and the fleets of Periclean Athens were always a failure in the long run. The power that adopted more robust methods of fighting, refusing to consider the sea as the monopoly of established skill and sea-power, invariably won. So the Athens of 480 beat the Persians; so Syracuse beat the Athens of 413; so Rome beat Carthage.

⁶⁴ Much of the criticism of these figures is rather perverse. Beloch's condemnation of them as round numbers, 180 Ath., 200 the rest, has been sufficiently met by Hauvette, (Hérodote, 391–3)Google Scholar, who pointed out, first that H.'s figure is not 380 but 378 plus two deserters (really 374 + 6 deserters, i.e. four Naxians included), and secondly that we cannot neglect the pentekontois. I hope I have said enough already about round figures (n. 62); and no doubt Themistocles' aim was a fleet roughly equal in power to the rest of Greece. More elaborate is the criticism of R. Adam, de Herodoti ralione historica, which I cite because Delbrück, seemed to think there was something in it (G. d. Kriegskunst, i. 12)Google Scholar. By omitting the twenty ships lent to the Chalcidians—or rather manned by Athenian kleruchs—Adam makes Athens furnish half the fleet, the other states half, including the deserters; next by omitting two of the deserters, he makes the Peloponnese furnish half of the latter half; and so on, ending in complete incoherence. This is supposed to prove that H. invented his figures on a scheme. We can all prove anything with any set of figures if we may juggle with them like this. I regret I have not been able to see Laird, Studies in Herodotus, who, I believe, holds that many of H.’s figures are mere calculations. If any reader will for a year or two keep count of the curious coincidences met with in the figures that he conies across in daily life, he will become very shy of rejecting figurea as ‘duplicates’ or ‘schemes.’

⁶⁵ [Dr. Macan conjectures for Aegina 42 + 18 on guard at home = 60, which one would like to believe.]

⁶⁶ 7, 144; see Kolbe, W., de Ath. re navali (Philol. 58, 1899), p. 509Google Scholar, etc. I may add that 200 would be four times the number (50) furnished by the naucraries (with the Paralos and Salaminia); this squadron of fifty appears in H. 6, 89. If Prof.Bury, be right about Aristides being στρατηγός at this time, with the command ashore (Cl. Rev. x. 414)Google Scholar, it is tempting to suppose that at Artemisium each of the other nine στρατηγοί commanded twenty ships, the remaining vessels, which should have been Aristides' command, going to Ghalcis.

⁶⁷ G. d. A. iii. 358; Forschungen ii. 183.

⁶⁸ Xen., Hell. i. 6, 24Google Scholar.

⁶⁹ Ap. Plut., Them. 8Google Scholar = de gloria Ath. § 7 = de Herod, malig. 34. Cf. H. 8, 18.

⁷⁰ I look oh the 110 of Ctesias, which Beloch adopted, as absolutely worthless. It occurs, moreover, in a context where Ctesias is trying to belittle Athens.

⁷¹ E.g. in Plutarch, , Them. 15Google Scholar: τοῖς βαρβάροις ἐξισούμενοι τὸ πλῆθος

⁷² References since Meyer: Raase, op. cit., with. fall bibliography; Cauer, F. reviewing Raase in Woch. für klass. Phil. 1905Google Scholar, no. 36 (a substantive contribution); Prof.Goodwin, W. W., Battle of Salamis (Harvard Studies in Class. Philol. vol. 17, 1906), p. 75Google Scholar, very full and giving a new explanation, after Lieut. Rhediades of the Greek navy, of the locics desperatus τὸ πρὸς ᾿Ελευσῖνός τε καὶ ἐσπέρης κάρας which Cauer thinks cannot be made sense of on any view.

⁷³ Aeschylus' reference to the main Persian battle as ἐν στοίχοις τρισίν imports that three of the fleets were there; στοῖχοι not ‘lines,’ but ‘divisions’, στοῖχοιas Prof.Bury, (Hist. i.²301Google Scholar) has taken it.

⁷⁴ See under Mycale, post.

⁷⁵ If Aeschylus bears on the question at all (see Goodwin, l.c., p. 93) he only proves that the Egyptians were in action somewhere. Mardonius' speech (H. 8, 100) proves nothing at all; if it did, it would prove that the Ionian fleet was not in action. At best it is mere rhetoric.

⁷⁶ Jahresh. 4 (1901), p. 101. Repeated Berl. Phil. Woch. 1905, p. 158.

⁷⁷ Already commented on, n. 44.

⁷⁸ Favourably received: Lenschau, l.c.; Kallenberg, H., Herodot, in Jahresb. d. Philol. Vereins in Berlin, 1904, p. 248Google Scholar.

⁷⁹ No doubt the point reached by the Corinthians was the temple of Athene Skiras; but we do not know where it stood. Raase, l.c., p. 33, has a useful list of the writers who think that the ‘Egyptians’ must have gone past Trupika to Leros.

⁸⁰ Demaratus' advice (H, 7, 236), given, be it noted, after Thermopylae, must belong here, i.e. after Artemisium. I take Artemisia's speech at the council (H. 8, 68) to mean the same thing. Parts of this speech must be genuine, (so Welzhofer and Meyer); or, if not Artemisia's own, must at least represent the opinion of Halicarnassus. One sign of accuracy is the belittling of the central and Egyptian fleets, but not of that of the traditional enemy of the Asiatic Greeks, the Phoenicians; for a contemporary would have seen the absurdity of running down the Phoenicians, however hated. Another is the amazing ‘quotation’ from Aeschylus: ; στρατὸς κακωθείς τὸν πεζὸν προσδηλήσηται = Pers. 728, ναυτικὸς στρατὸς κακωθεὶς πεζὸν ὤλεσε στρατόν (I have not seen this ‘quotation’ noticed [not even by Dr. Macan], though Plut., de malig. H. 38Google Scholar has some curious observations.) As H. was not really likely to make his heroine quote the best known, and least true, line of the Persae, we must suppose that Aeschylus himself was quoting a well-known saying; and as no one can have coined a phrase so remote from facts after the battle of Plataea, it may well have been a prophecy, traditionally attributed to Artemisia, though reflecting little credit on her judgment. It is true that the Scholiast, on Pers. 728Google Scholar interprets πεζὸν στρατόν as the troops on Psyttaleia; but the contexts are quite char to show that neither Aesch. nor H. meant this for a moment.

⁸¹ Du Sein, , Histoire de la marine, i. 110Google Scholar, suggested that the Persian action at Salamis must have been the result of a compromise.

⁸² The principal argument used by Delbrück and Meyer to show that the Persians were not stronger, or appreciably stronger, than the Greeks at Salamis, is that, if so, they must have divided their fleet and sent part to the Argolid. But suppose they did?

⁸³ I need not recapitulate the shifts to which different writers have been put to account for the Persians drawing out their fleet the day before the battle. Of course Aeschylus does not mention it; but he is writing drama, not a diary.

⁸⁴ Munro, p. 331.

⁸⁵ So Raase. The arguments seem irresistible. It explains why the Tenian deserter, which of course came the other way, was required to confirm truthful Aristides.

⁸⁶ Pers. 392, γνώμης ἀποσφαλεῖσιν

⁸⁷ See Bury, , Hist. i.²302Google Scholar. [If the Persians were roughly on the line Aigaleos-Psyttaleia or Aigaleos-Cynosura (see n. 92), this would bring the Aeginetans across their line of retreat, and account for the story in H. 8, 91.]

⁸⁸ See p. 216.

⁸⁹ Mardonius' speech is no evidence, as I have pointed out above. All Herodotus' details refer to two fleets only, the Ionian and Phoenician; and the fact that after the battle the Greeks, who seem never to have left the straits, expected Xerxes to attack again τῇσι περιεοὑσῃσι νηυσί shows that part of the Persian fleet had not been engaged, as he could not attack again merely with the squadrons that had just been badly defeated. It is possible that the central fleet helped to embarrass the fugitives, 8, 89; but by that time the real battle was over. Even if we reckon in the central fleet, the Persian total, which cannot have exceeded 280, would be barely superior to the Greek total at the best, and may well have been very considerably inferior to it.

⁹⁰ In spite of his words in H. 8, 109 (spoken for a purpose), we might once well have doubted whether he himself did not consider a live Themistocles more useful than any number of dead ἤρωες Yet we have lived to see the merit of another Salamis ascribed no less to the dead than to the living: rescript of the Emperor of Japan after Tsu-shima, ‘The result is due in a large measure to the benign spirits of our ancestors as well as,’ etc.—ἤρωσι συμμἀ χοισι

⁹¹ I am assuming that the Persian land forces were strictly limited in number.

⁹² [Dr. Macan thinks that H. only meant that the majority of the marines were Persians and Medes, and that an allusion to the original Medo-Persian epibatae ‘would be far-fetched.’ Why? It would be a natural enough allusion for any source which regarded the fleet as an organised force and not as a mob.]

⁹³ [Dr. Macan's theory of Salamis is, very briefly, as follows: The Persians, on the day before the battle, decide to blockade the Greeks in the bay of Salamis; they therefore send the Egyptians round to the Megara channel, the main fleet to the Psyttaleia end (this avoids the time difficulty for the Egyptians, and also accounts for the Peloponnesians wanting to go home, 8, 74, when they heard of the Egyptians passing, though Dr. Macan does not notice either point; it also accounts for the Persian fleet drawing out the day before the battle). On receipt of Themistocles' message they alter their first plan and sail in not expecting any battle (it will be seen that I agree with both these points). On the morning the Persians sail in in column of three lines ἐν στοίχοις τρισίν) between Psyttaleia and the mainland; the Athenians take the head of the column in flank and break it, deciding the action. The Persians on Psyttaleia were either landed during the action, or else belong to the first (abandoned) plan and were meant to invade Salamis.—While there is much to be said for this, I adhere to what I have written above, on the few points where I differ. (1) Dr. Macan admits that the Persians, if they meant to fight (first plan), were bound to try to get the Greeks into open water; why then blockade them? A blockade would have given Themistocles just what he wanted: the Persians could not have avoided close quarters. (2) Even if Themistocles' message reached, not Xerxes (Aesch.), but the admirals (H.), it is clear that the latter could not change the whole plan without consulting their commander-in-chief, as the army and fleet were co-operating; the fleet then must have been back at Phalerum when the message arrived in the early part of the night, and put out (afresh) that night, as Aesch. says. Consequently, the movement of the fleet on the day before was a demonstration only; and what becomes of the blockade? (3) Dr. Macan has to treat the objective of the army as the Megara channel, to co-operate with the Egyptians. But, after all, H. says the Isthmus; let us keep what of tradition we can. (4) The battle must, I think, have been fought in line; Dr. Macan (ii. 315–6) cannot explain the Aeginetan ἀριστεῖα No doubt the Persians entered in column, either one column or two; but (supposing now with Dr. Macan that it was one column) they could never have been caught in column by a fleet coming across from Salamis, when a mere half-turn by each ship would have brought them into line abreast facing the enemy; and we cannot press Aeschylus' ῥεῦμα to prove the contrary. Two hundred triremes in column of two lines, 100 in each line, would cover about a mile from end to end; the whole column would be in the bay in six to seven minutes, or even less (Fincati's trireme did nine miles an hour, and the Phoenicians might do better than that for a short distance); by the time the Greeks had got under way, hesitated, backed water, and finally attacked, the enemy might have formed line abreast, roughly on the line Aigaleos-Psyttaleia. No doubt, however, there was some confusion. (5) Psyttaleia. We might suppose that the object of the ‘blockade’ was to throw a corps, behind and under shelter of the main Persian fleet, across into Salamis, capture the Greek base from the land side, and leave the Greek fleet in the air. But the tradition contains no hint of anything so exciting; and, if this were the plan, why land the troops on Psyttaleia?]

⁹⁴ This follows from the fact that its admiral Achaemenes, who was not superseded, was not at Samos (H. 8, 130), or at Mycale, or with Mardonius.

⁶⁵ H. gives 300 ships. This figure is of no use; like Mardonius' loss at Athos, it is so obviously one half of the whole.

⁹⁶ H. 9, 96. It has been pointed out by von Domaszewski, A., Beiträge zur Gesch. d, Perserkriege (Neue Heidelberger Jahrbücher, 1891), p. 187Google Scholar, that H. does not expressly say that the Phoenicians went home, and he has an attractive theory that the bulk of the Persian fleet, after Salamis, returned to the North Aegean to guard Mardonius' communications. I am afraid that the presence of three admirals at Mycale disposes of this view; no fleet could keep the sea without its marines. Moreover, Leotychides could not possibly have sailed for Samos with a strong Persian fleet, unopposed, on his flank and rear; and we can hardly suppose that the Greeks had a second fleet at sea, plus the army at Plataea.

⁹⁷ This follows, as to the Phoenician fleet anyhow, from the στρατηγός remaining after the ships were sent off.

⁹⁸ Hence the fleet is a στρατός and its camp a στρατόπεδον (H. 7, 124, etc.). One is reminded of the fleets of the Roman Empire. Unfortunately we have no information as to the relations, on a Persian ship, of the trierarch to the commander of the marines, that terrible crux of the later Roman fleet. Artemisia appears as mistress in her own ship: yet, though the marines were few compared with those on a Roman vessel, they were of an alien and dominant race. One would like to know how Darius solved the problem. The fact that Achaemenes, after landing his Egyptian marines, took his fleet home, may show that his position differed somewhat from that of the other στρατηγοί and that he as a satrap was not merely a general of marines. But it might also mean that he shipped Persian troops in their place, with a view to possible disaffection in Egypt.

⁹⁹ Taking the 110 Greek ships at 150 rowers and 18 marines, they could land some 18,000 troops of all sorts. If we take each of the three Persian fleets at say 80 ships (they can hardly have been stronger by now) we get, at 20 marines per ship, 4,800 troops, or say 4,000, for some were not there (H. 8, 130). Tigranes had what remained of his army corps, perhaps originally 10,000 (n. 27; not 60,000, as H. says), and the Persians were encumbered by some 12,000 armed and disaffected Ionian rowers. The extreme weakness of their position is apparent.

¹⁰⁰ Hauvette, l.c. 195, has shown that H. did not get his figure here from Darius' stelai on the Bosphorus.

¹⁰¹ See Meyer's, reconstruction of the narrative of Callisthenes of Olynthus in his Forschungen, ii. pp. 1Google Scholarseq., Die Schlacht am Eurymedon.

¹⁰² For recent discussions of this system see Ginzel, F. K. in Klio, vol. i. pp. 349–380Google Scholar, and C. F. Lehmann-Haupt in ditto, pp. 381–400.

¹⁰³ Sosylos is at least evidence for this much, when, in referring to the Carthaginian navy, which he knew, he says that the Phoenicians always do so and so.

¹⁰⁴ I am indebted here to the chapter on the Carthaginian navy in Meltzer, Gesch. d. Karthager, vol. ii.; and for what follows I refer once for all to my paper in J.H.S. xxvii. (1907), 48.

¹⁰⁵ This is only a combination (Meltzer, ii. 234), but a good one.

¹⁰⁶ Polyb. 1, 53, 9.

¹⁰⁷ My conclusion (J.H.S. xxvii. 57), that the (successful) object of Rome in the first Punic war was to keep afloat a fleet of 20–40 ships more than Carthage, ought to be expressed differently. They aimed at maintaining four divisions to the Carthaginian three. These divisions were not necessarily of the same strength as the Carthaginian, but there is little evidence for the strength of a Roman division in the first Punic war, and possibly it was not constant.

¹⁰⁸ In case anyone should think the whole question of these divisions fanciful, I append a few figures from the Roman navy, taken from the mass of material in Livy, Polybius, and Appian. From 218 to 214 a Roman division (as in the first Punic war) fluctuated between 60, 55, and 50. In 214 Rome answered the Carthaginian mobilisation of 215 with a decree for a (standing) fleet of 150 quinqueremes in home waters (Livy 24, 9), and henceforth the Roman division was 50 ships of the line. The two standing fleets from 214 to 206 were, Sicily 100, Adriatic 50. In 208 two additional special squadrons of 50 quinqueremes each were formed for Italy and Sardinia. After 206 Rome laid up ships fast, and the figures fall. War against Philip (198): 100 tectae, 50 apertae (probably allies), and lembi (Liv. 32, 21). Against Antiochus, first 100, then 50, quinqueremes ordered; not all built; at sea in 191, one division (50) under Livius, with a half-division (25) taken over from Atilius, and allies (Liv. 36, 41). Against Perseus, 50 quinqueremes ordered (Liv. 42, 27). Against Carthage in the last war (App. Lib. 75), 50 quinqueremes, and allies. A complete analysis of the second Punic war is really conclusive. Livy omits the scouts from the divisions, or gives them separately, as being generally allies.

¹⁰⁹ See Mülder, D. in Klio, 7, 29Google Scholar, already cited.

¹¹⁰ Frag. 5 in Kinkel, Epic. Graec. fragmenta. Also Choerilus in Pauly- Wissowa (Bethe)

¹¹¹ As Lehmann-Haupt, C. F. in Klio, 2, 338Google Scholar, n. 2.

¹¹² What Aeschylus' unlikely figure of 207 for these means can hardly perhaps be ascertained. It may relate to something else and have got transferred.

¹¹³ Mr.Wells, J., The Persian Friends of Herodotus (J.H.S. xxvii. 1907, p. 37)Google Scholar.

¹¹⁴ The speeches of Demaratus and Achaemenes belong after the battle.

¹¹⁵ The information may have only reached H. at second or third hand, of course. It need not, either, have been exclusively Halicarnassian; he has some Samian details about Salamis, which, however, Mülder (l.c.) attributes also to Choerilus.

¹¹⁶ Cf. Polyb. 12, 25^f, of Ephorus, ἐν τοῖς πολεμικοῖς τῶν κατὰ θάλατταν ἔργων ἐπὶ ποσὸν ὑπόνοιαν ἐσχηκέναι μοι δοσεῖ with illustrations. This is pared away by Schwartz in Pauly-Wissowa s.v. Ephoros (vi. i. 11). But I think we may agree with A. von Mess, l.c. p. 406, that the question of Ephorus' sources for this period is more complex than is usually supposed.

¹¹⁷ It is always possible that the number of the northern fleet was preserved in the traditions of Cyme, and that Ephorus, with his known local patriotism, adopted that tradition. This would explain his radical divergence from Herodotus over the one fleet.