¹ Kromayer, J. (‘Die Entwickelung der römischen Flotte,’ Philologue 1897)Google Scholar, who has gone into the figures for the civil wars, accepts 300 for the fleet of Sextus Pompey also. But this seems to me impossible; for Pompey's 3 squadrons at Mylae, totalling 155, are described by Appian as constituting the larger part of his fleet; after losing 30 at Mylae and some at Tauromenium he cannot have had more than 250 at the most at Naulochus, for building between the two battles was out of the question. This would give a total of about 550 ships in action.

² Kromayer, J. in Hermes 34 (1899) p. 1 Google Scholar. If Octavian had 400, and Antony 170 (plus 60 Egyptian), over 600 ships were engaged. But the figures for the Actium campaign are very uncertain.

³ See J. Beloch, Die Bevölkerung der gr. röm. Welt; also Die Bevölkerung Italiens in Alterthum in Beiträge zur alten Geschichte, vol. 3.

⁴ The following helps one to realise what such figures mean. On a population of 42 millions, the British Navy has a personnel of 121,983 (including coastguard and marines), and mobilised 319 vessels of all sorts for the manoeuvres of June—July, 1906; while in crew and troops two quinqueremes carried about the same number as one battleship.

⁵ Meltzer, has stated this (Gesch. der Karthager, vol. 2, p. 568, n. 49)Google Scholar. But he makes no application of it to the numbers.

⁶ The view that Rome, prior to 260 B.C., used Greek ships only, seems again coming into prominence, no doubt as a reaction against Mommsen; see e.g. Speck, E., Handelsgeschichte, 3, i, § 715 Google Scholar. But it is demonstrably wrong. Rome controlled no Greek ships before 327 B.C. (treaty with Neapolis); while the first treaty with Carthage, which cannot be later than 348 B.C., presupposes Roman warships.

⁷ Not, of course, manned by Romans.

⁸ Liv., Per. 12 Google Scholar; App., Samn. 7, 1 Google Scholar.

⁹ Liv. 40, 26.

¹⁰ Liv. 41, 1.

¹¹ Mommsen thinks the Tarentines had to give up their ships after the war with Pyrrhus; but probably this was not the case (see Niese, , Grundriss d. röm. Gesch. in Müller, 's H. d. k. A. III, 5, (1906), p. 70, n. 4)Google Scholar, for Tarentum kept its independence and had a number of ships in the second Punic war. Even if they did hand over some ships, the Romans, as was their custom, probably burnt them.

¹² Polyb. 1, 20.

¹³ The number appears (quite clearly) from App., Syr. 22 Google Scholar combined with Liv. 36, 42, and is presumably that of Polybius.

¹⁴ I do not mean that they had anything to do with duoviri, who are not heard of during the period of the great naval wars.

¹⁵ Liv. 35, 20. In the affairs of Greece and Syria Livy is supposed to represent the substance of Polybius fairly accurately, and for the naval war with Antiochus the way in which Appian agrees with and complements Livy makes this almost certain. If the 100 quinqueremes of 260 B.C. be from Fabius those of 192 B.C. are not.

¹⁶ Two Massiliotships joined Cn. Scipio in 217 B.C. (Polyb. 3, 95 = Liv. 22, 19); and in 211 B.C. four joined the propraetor M. Junius Silanus (Liv. 26, 19).

¹⁷ Naturally I attach no importance to the fact that Floras 1, 18, 7 says the Romans built 160 ships in 260 B.C.

¹⁸ Reuss, F., Philologue 60 (1901) p. 102 Google Scholar, who has made the latest examination of Polybius' sources, gives Chs. 20–24 as all from Fabius. But such a result seems to me merely to condemn his method; for the break in the sense of the narrative between Chs. 21 and 22 is patent to anyone.

¹⁹ Gr. Gesch. vol. 3, i. p. 677 n. 1. If this be so, it is noteworthy that Philinus knows nothing of the boarding-bridge (corvus, κόραξ); and no doubt Ihne was right in suspecting the traditional account of this machine, which is not heard of after Ecnomus, and which seems part of the deliberate introduction by Fabius of an element of wonder into this war: for, after all, boarding and πεζομαχία were the oldest form of sea-fighting known, and the Carthaginians would have been delighted with an arrangement that would have prevented more than two Romans coming aboard at once. Now the Athenians had used grapnels in 413 B.C., and they occur commonly in the second Punic war; and the κόραξ was probably an improved grapnel on a pole, like Agrippa's ἄρπαξ. Dion Cassius so understood it, for Zonaras speaks of the Romans using χεῖρας περικόντους σιδηρᾶς and Appian so understood it, for whereas Agrippa's ships at Naulochos use the ἄρπαξ (described App. b. c. 5, 118), at Mylae these same ships have κόρακας (b. c. 5, 106). The supposed boarding-bridge must have been taken by Fabius from the bridge or ladder of the sambuca used by Marcellus in attacking Syracuse, no doubt a real machine, as it appeals again (somewhat altered) in the sieges of Rhodes and Cyzicus by Mithradates, (App. Mith. 26 Google Scholar compared with 73). Some sort of ladder for boarding occurs, however, at the battle of Cumae, (App. b.c. 5, 82)Google Scholar, if this be the meaning of Appian's unique use of καταῥῥάκται

²⁰ About the proportion refitted after Ecnomus.

²¹ Viz, battle of ch. 21, 30, Mylae 50, Tyndaris 18, Ecnomus about 100 (64 + over 30), Hermaea 114, Aegatcs Insulae 120; 432 altogether.

²² The consul hurries after the enemy as an easy prey, going forward with 10 ships; they surround him and sink 9, but, pursuing the flagship, become engaged with the main Roman fleet, and lose 8 sunk, 10 taken.

²³ Polyaen. 8, 20; Romans 200, Carth. 80.

²⁴ The date is fortunately not material here, for either year is open to serious objection. For a summary of the arguments, see Reuss, u. s.; also Beloch, , Gr. Gesch. 3, ii, 234 Google Scholar, whose reasons for 254 are hardly convincing. The difficulty is this: 255 gives no time for the siege of Clypea, and does not explain why the Hermaea triumph fell in 253; while 254 makes the Romans first waste a year before succouring their beaten troops (though the fleet was ready), and then send out the fleet, not under the consuls (as on all other occasions in this war), but under the consuls of 255–4 as proconsuls, without any apparent reason; for the very different proconsular squadron-commands of the 2nd Punic war can hardly be cited in support.

²⁵ Polyb. 1, 29, says they refitted the prizes. If he means all, which is unlikely, then only forty-four were taken, and the lower of the two numbers hereafter discussed for the Carthaginian fleet at Ecnomus becomes even more probable.

²⁶ The number 350 given for Hermaea does not shew that the storm number should be 464, because, ceteris paribus, the smaller number is to be followed. But as a fact Eutropius, or some scribe, had the curiosity to add up, and Eutropius does give 464 for the storm; and Meltzer hereupon suggests that the real Polybian tradition may have been 464, a suggestion which is out of the question for at least three separate reasons: it prefers the easier version and the larger number, and corrects a good early writer with a sound text by a poor and late compiler. Reuss, , u. s. and Speck, , Handelsycsch. 3, ii, § 824–5Google Scholar, follow Meltzer; but one cannot write history merely by taking the line of least resistance.

²⁷ Possibly the number 114 comes from the column. Perhaps, too, the reason why Philinus (Diodorus) gives the Roman loss at Drepana as 117 was to shew that the Carthaginians had had a full revenge for Hermaea.

²⁸ There is an exact parallel in Appian's account of the war against Autiochus; he dismisses in two casual lines the very important defeat of Hannibal by the Rhodians at Side, which Livy gives at length; and this in a war where his general agreement with Livy over the naval operations is most marked.

²⁹ If any Roman ships were lost at Hermaea, the number lost in the storm would be fewer, 80 anyhow remaining.

³⁰ This must be a common source of confusion in classical (as in modern) fleet numbers. No figures in antiquity are more exaggerated than those of transports; after such numbers as 3,000 and 1,600, the writer who confined himself to less than four figures must have been astounded at his own moderation.

³¹ The difficulty, of course, all through (money apart), both at Rome and Carthage, must have been, not ships, but men to row them.

³² The Romans are described as in wedge formation, not in line.

³³ Even at Drepana, Polybius says, there was some ramming, and some ships settled.

³⁴ Assuming that they did have as many as 200 ships at Hermaea.

³⁵ Polyb. 1, 49, πάντι τῷ στάλῳ. I mention this as it is commonly assumed that he left a squadron at Lilybaeum, a most useless proceeding, as the blockade was a failure anyhow, and his striking force insufficient.

³⁶ Polybius says the prizes were taken to Carthage. Some were damaged; if we take the same proportion as after Ecnomus, 60 to 70 at the most would be worth refitting, and Adherbal would have remaining just about the same number of ships to tow them.

³⁷ The Fasti shew that he was Claudius' colleague. Polybius speaks as if he were his successor and sailed the next year (248 B.C.), but it seems reasonably clear that the naval operations under both consuls form one connected sequence and took place in the same year, 249 B.C.

³³ Pol. 1, 52. If, in fact, auy ships joined him from Lilybaeum, they must have been sent off before the battle of Drepana.

³⁹ The 300 of Polybius may be another instance of confusion due to transports or ships other than warships; for Diodorus says the Romans sailed to the siege of Lilybaeum with 240 long ships and 60 cercuri, i.e. 300 vessels. Cercuri occur in the Roman navy (or the Roman service) in both the second and third Punic wars; Liv. 23, 34; App., Lib. 75 Google Scholar.

⁴⁰ At first sight the loss in the first storm would seem to be 170 Roman ships and 114 prizes; for the prizes would, of course, have been cast off when the storm broke. But as rostra were forthcoming for Aemilius Paullus' column, one division must have got to harbour with its prizes; consequently the Roman ships lost were more than 170.

⁴¹ The 200 ‘ships’ with which Hasdrabal crossed to Sicily (Poly b. 1, 38) are obviously transports. To suppose that they were warships makes nonsense of the events before Lilybaeum in 250 and 249 B.C., more particularly of the Romans laying up part of their fleet, the account of which is very circumstantial. It is no objection to this that Hasdrubal did get his large army across; the Romans had no naval base facing Africa, and even if they had had, the command of the sea (such as it was with galleys) rarely, if ever, prevented an army crossing in ancient times. Pompey commanded the sea absolutely as against Caesar; so did the liberators as against Antony and Octavian; yet in each case the Adriatic was crossed in force.

⁴² Meltzer, vol. 2, p. 336.

⁴³ There is another, perhaps a better, way of getting at the Carthaginian fleet of 241 B.C. In the war with the mercenaries the Carthaginians had nothing but triremes and pentekontors (Polyb. 1, 73); they had therefore lost all their quinqueremes at the Aegates Insulae, including presumably the Roman prizes. Suppose all the 120 ships lost to have been quinqueremes, the swifter ships alone escaping. 120 is 50 per cent. of 240, 60 per cent. of 200, 70 per cent. of 170, 80 per cent. of 150. But we know that in 219 B.C. the Carthaginian fleet of Spain contained 88 per cent. of quinqueremes, (post, p. 9); it is therefore most unlikely that their fleet of 241 B.C. contained as few as 50 per cent., or even 60 per cent., and we come back to this, that a fleet of not over 170 cannot be far from the mark. Of course, if the 120 ships lost were not all quinqueremes, the argument is even stronger.

⁴⁴ It 200 ships or so was in fact Carthage's effective limit, the limitation must have had to do with the crews, of which we know little. It has nothing to do, for instance, with the number of the νεώρια at Carthage being 220; for, apart from Utica, the Carthaginians had the control of the docks built by Agathocles at Hippagreta, (App. Lib. 110)Google Scholar; and besides, a fleet could at a pinch winter ashore anywhere (e.g. Liv. 36, 45).

⁴⁵ The only time, except in Spain, when a Roman squadron seems to have been outnumbered by a Carthaginian was after Syracuse had joined Carthage; Marcellus, in 212 B.C., had to offer battle to Bomilear with an inferior force.

⁴⁶ Polyb. 3, 41 = App. Ib. 14 = Liv. 21, 17. Livy adds 20 celoces.

⁴⁷ Liv. 23, 32.

⁴⁸ The number captured at New Carthage, Polyb. 10, 17.

⁴⁹ i.e. 150 for Sicily and the Adriatic (Liv. 24, 11), plus 35 in Spain, the original squadron of Cn. Scipio. There were also in Spain 25 prizes taken from Himilco (Polyb. 3, 96 = Liv. 22, 19), but it appears from Polyb. 10, 17, that these were not fitted out.

⁵⁰ Liv. 25, 27. I find it impossible to make out from Livy's confused narrative (24, 36; 25 25; 25, 27) whether there were or were not 20 other Carthaginian ships in the harbour of Syracuse with Epicydes.

⁵¹ Liv. 26, 19.

⁵² Liv. 27, 22. The 80 ships of Scipio that he mentions are the original 35, the 18 taken at New Carthage, and the 25 taken from Himilco and never fitted out.

⁵³ Polyb. 10, 17.

⁵⁴ In estimating this tremendous effort (280 ships), it must be remembered that the Romans were now in part using slave rowers (Liv. 24, 11; 26, 35); and that some of Scipio's crews were pressed Spaniards (Polyb. 10, 17 = Liv. 26, 47.)

⁵⁵ Mylae 50, Sardinia 20 to 40, Tyndaris 18, Ecnomus 100, Hermaea something over 114, say 130, Aegates Ins. 120: or about 438 to 458 all told. Polybius adds another 30 or so for the supposed battle before Mylae.

⁵⁶ Boodes takes 17, Tyndaris 9, Ecnomus 24, first storm (with Hermaea) 170, second storm 27, year of Drepana 223 = 470. Add some 10 for Mylae, and an unknown loss at the Aegates Insulae. Polybius gets his figure by counting the Hermaea prizes again, and adding another 123 for the second storm.—Refitted prizes make the Roman and Carthaginian totals overlap to some extent, perhaps 70 to 80 ships.

⁵⁷ Perhaps I may give one instance of what I mean. Battle of Corycus, 191 B.C.; Livy and Appian agree exactly as to the Roman fleet, but Livy gives Polyxenidas 100 ships ( 70 cataphracts, 30 apertae), while Appian, (Syr. 22 Google Scholar) gives him 200. Now Appian, who himself wrote on the Roman navy, does not, as a rule, throw naval numbers about anyhow; (for the proof of this see Kromayer's article in Philologus before eited, passim); and the explanation must be that Polyxenidas' battle fleet of 100 ships was accompanied by some 100 light craft, which Livy has not given. As we fortunately know that 10 years earlier Philip V. had fought in these waters with a fleet containing 150 lembi to 53 cataphracts, we can see that Appian is probably right, especially as Polyxenidas was engaging an enemy 151 strong; and for once we completely justify the larger number.

⁵⁸ Polyb. 1, 53, 9, οἱ προπλεῖν εἰθισμένοι λέμβοι

⁵⁹ Polyb. 1, 20; Duilius' column.

⁶⁰ Polyb. 1, 20; 1, 73.

⁶¹ Polyb. 1, 47.

⁶² For instance, Mareellus' fleet before Syracuse is 60 quinqueremes (Polyb. 8, 4 (6) = Liv. 24, 34); but he has a quadrireme (Liv. 25, 30) and 2 trirenies and 3 smaller craft (Liv. 26, 39). (I do not say that the smaller craft are reckoned in the 60.) Again, Liv. 21, 49 and 50, the praetor M. Aemilius cuts off and captures 7 Carthaginian quinqueremes, with 1,700 milites nautaeque, i.e. about 242 to a ship; the rowers on a quinquereme were more than that, according to Polybius.

⁶³ Liv. 26, 24, the treaty with Aetolia. That quinqueremes were in fact sent appears from Liv. 27, 32, where on 15 of the ships Sulpicius ferries 4,000 troops over the Gulf of Corinth, giving an average of 266 on a ship, which Kromayer says is the highest to be found.

⁶⁴ Polyb. 10, 19 = Liv. 26, 51.

⁶⁵ Polyb. 15, 2; Liv. 28, 17; 29, 9; 30, 25; 30, 26; 31, 11.

⁶⁶ I think there is no instance of any other type being built.

⁶⁷ Polyb. 3, 33 = Liv. 21, 22.

⁶⁸ If I may venture on one modern parallel, the place of the quinquereme at this time was exactly that of the 74 under Nelson; the quadrireme and trireme corresponded to the smaller ships of the line of 60 or fewer guns, the ‘light craft’ to frigates and brigs, while the heptereis and dekereis of the Hellenistic powers took the place of the ships of 110 and 120 guns built by France and Spain. Though both quinquereme and trireme fought in the line, the fact that Livy classes triremes among ships minoris formae, as opposed to the quinqueremes, &c., majoris formae (37, 23; 36, 41), shews some well-marked distinction between the two other than mere size; no doubt the line of division is between the galleys with little oars rowed by one man and those with great oars rowed by several men, a distinction which to a spectator would be most conspicuous.