¹ Vol. xxi (1931), 173–199.

² Hermes xxxiv (1899), 1–54Google Scholar.

³ Vol. x, 100–106.

⁴ Vol. lxviii (1933), 361–383 (all references are to this article unless otherwise stated).

⁵ Tarn thinks that Antony's fleet was organised in six squadrons (or seven including Cleopatra's). Two of these formed the left wing at Actium, one (or two with Cleopatra's) the centre, and three the right. Plutarch's statement (Ant. 65, 1) that there were two commanders in the centre might be an objection to this arrangement; but I am not here concerned to criticise it.

⁶ Based on the assumption that the 5,000 killed (Plutarch, Ant. 68, 1Google Scholar) included rowers. Kromayer takes the figure to apply to marines only, and estimates that about 40 ships were sunk (pp. 368–9). Orosius vi, 19, 12 gives 12,000 killed and 6,000 wounded, but probably Plutarch's figure is to be preferred.

⁷ Both Florus and Orosius apologise, as it were, for Antony's smaller number of ships : ‘sed numerum magnitudo pensabat’; ‘quantum numero cedens tantum magnitudine praecellens.’ It seems likely that the over-emphasis on the larger size of Antony's ships, which appears in nearly all the ancient historians, was due to an attempt to magnify a victory that was in fact won over a greatly inferior fleet. Tarn is no doubt right in saying that Octavian's fleet was substantially the same as that which had defeated Sextus Pompeius. For that Agrippa had built comparatively large vessels.

⁸ Kromayer did not admit this (see p. 37;, n. 1), and Tarn only goes so far as to say that Dio ‘makes Antony claim to have the larger fleet,’ referring to l, 18, 4–5 and 19, 4. But the fact that Octavian does not make a similar claim seems to me significant, See l, 24, 2—the sort of thing a commander whose fleet was in fact smaller might be expected to say. In cap. 28, 6, Octavian clearly admits that Antony's marines are more numerous, and the same notion appears in the account of the battle : after the light of Cleopatra, Octavian's marines ventured to board Antony's ships ‘since they were now equal in number’ (J, 33, 6). A possible objection is to be found in an earlier passage (l, 32, 6) when Dio speaks of two or three of Octavian's ships simultaneously attacking one of Antony's (part of a general description of the battle). But this might be possible, whatever the relative numbers, given the conditions that Dio imagines—the superior mobility of Octavian's ships and Antony's defensive tactics. It does not necessarily imply that Octavian's fleet was the larger.

⁹ Dio says the greater part of the fleet had wintered in the Ambracian Gulf. Whether this is to be taken literally (=250+), as by Kromayer, seems to me questionable. It is doubtful whether Antony could or did bring up any appreciable number of ships after the opening of the campaign. See Kromayer, Hermes xxxiv, 30fGoogle Scholar.

¹⁰ Dio l, 13, 5; 14, 1; 30, 1. The account of Sosius' defeat seems to imply fairly heavy losses. Orosius vi, 19, 7, records an action in connexion with the occupation of Corcyra.

¹¹ Plutarch 64, 1; Dio l, 15, 4. The phraseology in both cases suggests the destruction of a considerable number: ‘He manned the best and largest ships, from triremes to dekereis, … and burnt the rest. ‘he chose the best vessels and burnt the rest.’ For famine, disease and desertion in Antony's camp and consequent loss of rowers see Orosius vi, 19, 5; Velleius ii, 84–5; Plutarch 62, 1; Dio l, II, 2; 12, 8; 15, 3 and 4. For another instance of loss of rowers during the winter Appian BC v, 98, 406Google Scholar. (For his campaign against Sextus Pompeius in 36 Octavian was only able to use 102 of the 120 ships then stationed at Tarentum.) Dr. Cary, who has kindly read this paper in manuscript, reminds me of the attrition of the Athenian fleet at Syracuse in 415–3. The Athenians sent out altogether 176 triremes (94 in 415, followed by 82 under Eurymedon and Demosthenes in 413). Yet in the third battle in the harbour they only used 86, and in their desperate attempt to break the Syracusan blockade 110, including some ‘not for service.’

¹² See his remarks on p. 365.

¹³ Cf. Plutarch 61, and Orosius vi, 19, 6 and 8. Orosius' authority apparently forgot to add the division originally stationed at Tarentum under Agrippa to that originally stationed at Brundisium under Octavian, and consequently reduced the fleet at Actium from more than 400 (the figure given by Florus and generally accepted as correct), to 230 ‘rostratae naves’ and 30 ‘sine rostris.’ Presumably the same source underlies Plutarch's 250 (a round number).

¹⁴ It would, incidentally, imply that the inventor believed that Antony intended to retreat, or the ‘perversion’ is even sillier than Tarn already thinks it. If he believed that, he presumably found it in the epitomator, who therefore found it in Livy.

¹⁵ Agrippa's division raided the Peloponnese and captured Methone before Octavian crossed.

¹⁶ On this point see Kromayer, Hermes lxviii (1933), 372Google Scholar.

¹⁷ Orosius vi, 19, 8.

¹⁸ ‘Florus’ ‘Orosius’ and ‘Plutarch’ are, of course, meant to include their authorities.

¹⁹ Tarn, p. 189, n. 1. Ferrabino in Riv. di Fil. ii (1924), 433–72Google Scholar, liii (1925), 131 f. For a criticism of Ferrabino's views see Kromayer, , Antike Schlachtfelder iv, 662–71Google Scholar.

²⁰ Dio l, 13, 2.

²¹ Incidentally it would have been very risky for Antony's right wing to make a deliberate gap, on the chance of Cleopatra coming up in time to it. The result might easily have been an enemy break-through, with Antony cut off from Cleopatra. For an interesting parallel (in a land battle) cf. the battle of Mantinea in 418 B.C. (Thuc. v, 72). (I am indebted for this point to Dr. Cary. See, however, Woodhouse, W. J. in BSA xxii, 1916–1918, 51 f.Google Scholar, for a defence of Agis' tactics.)

²² The reading of the first two words is uncertain, and not important for the present purpose.

²³ See Tarn, p. 174 f., for references.

²⁴ I assume for the purpose of the present paper that these lines do in fact refer to Actium. For a recent exposition of the view that they do not, see M. Wagenvoort's article in Mnemosyne lix, 1932, 403–421Google Scholar. He thinks that the Epode is in fact two poems, the first (written before Actium) ending at l. 20 or 22; the two lines under discussion refer to Sosius' ‘eruptio.’ For another recent discussion of the Epode see Wilkinson, L. P. in CR xlvii, 1933, 2 ffGoogle Scholar.

²⁵ What else would they do ? Retreat—if possible—was preferable to immediate surrender. That it was in many cases possible—with a refuge so near as the Ambracian Gulf—there can, in my opinion, be no reasonable doubt. To follow Antony southwards was a more difficult matter, in view of the encircling movement of Octavian's right wing. But some who had seen the end of the battle did succeed in following him (Plutarch 6y, 5). Florus' phrase (ii, 21, 6) : ‘Caesaris naves … illas … ad arbitrium dissipavere ‘would most naturally be taken to mean ‘routed,’ ‘put to flight.’ And compare Vergil (Aeneid viii, 705 f.) : ‘omnis eo terrore Aegyptus et Indi, omnis Arabs, omnes vertebant terga Sabaei. ‘Plutarch's ἀπεῖπε (68, 1) might mean either flight or surrender. But compare the expression used in cap. 67, 5 above : ἤδη δὲ καὶ τῶν στρογγύλων πλοίων οὐκ ὀλίγα καὶ τῶν ϕίλων τινὲς ἐκ τῆς τροπῆς ἠθροίζοντο πρὸς αὐτούς.

²⁶ Hence Bentley made them retire ‘in portum aliquem Actio propinquum ad sinistram’ (quoted by Tarn, p. 174, n. 3). But what ‘portus’ was this ?

²⁷ ‘Some of his ships refused to fight and returned to harbour (words indicating something towards the left came somewhere in this clause)’ (p. 17). ‘We can now fill up the gap I left in Octavian's despatch, the meaning of the crucial sentence, however expressed, was this: “He tried to turn us with his right, but most of his ships refused to fight and the left returned to harbour”’ (p. 195).

²⁸ Suet. Div. Aug. 17, 2Google Scholar; Orosius vi, 19, 11.

²⁹ CAH x, 105.

³⁰ Cf. Dio l, 11, 5 f.; Res Gestae 25, 3.

³¹ Tarn thinks Plutarch's statement (68, 1) that the battle lasted till the tenth hour was quoted from Augustus' memoirs (p. 177).

³² There is no lack of stories to Augustus' discredit (which Actium, on any hypothesis, was not) that were not hushed up. For his tolerance and ‘civilitas’ cf. Suet. Div. Aug. 51, and 54–6. See Kromayer, p. 383, for further criticism of the alleged forgery.

³³ Hermes xxxiv, 29 f.; lxviii, 383.

³⁴ Hermes xxxiv, 25 and 29–30.

³⁵ Dio l, 31, 4.

³⁶ Compare the battle between Calvisius and Menecrates off the Campanian coast in 38 B.C. (Appian BC v, 81–3, 342 ff.Google Scholar).

³⁷ Kromayer in Hermes xxxiv, 41 fGoogle Scholar.

³⁸ Dio l, 31, 4.

³⁹ Plut. 65, 4; Kromayer, , in Hermes xxxiv, 43Google Scholar.

⁴⁰ Plut. 65, 5; Servius on Aeneid viii, 682Google Scholar; Dio l, 31, 5.

⁴¹ See Kromayer, 377 f.

⁴² Compare Agrippa's intention after the battle of Mylae, and the result of his failure to carry it out (Appian BC, v, 108, 445 f.Google Scholar).