¹ JRS XXVII (1937), 180–207Google Scholar, ‘The origins of the second Macedonian war’.

² We are especially grateful to Professor Casson for first raising with us the problem of the reading in the naval clauses in Livy and Polybius and for subsequently reading this article and advising us on many points, particularly in the sections dealing with Hellenistic ships. Many of the topics touched on here will be fully treated in his forthcoming book, Ancient Ships and Boats (Princeton). We have also to thank Sally Humphreys and other readers at the editorial stage of our draft for their valuable comments and criticism.

³ JRS LVII (1967), 1–8Google Scholar ‘The treaty of Apamea (188 B.C.)’.

⁴ See JRS XXVII (1937), 206–7Google Scholar; Walbank, , Philip V of Macedon (Cambridge, 1940), 128Google Scholar.

⁵ Colonies were founded in 194 at Volturnum, Liternum, Puteoli, Salernum, Buxentum, Sipontum, Tempsa and Croton (Livy XXXIV, 45, 1–5); cf. Salmon, E. T., JRS XXVI (1936), 47–67Google Scholar.

⁶ Thiel, J. H., Studies on the History of Roman Sea-Power in Republican Times (Amsterdam, 1946), 255–372Google Scholar.

⁷ Tarn, W. W., Hellenistic Military and Naval Developments (Cambridge, 1930), 127Google Scholar.

⁸ See Thiel, op. cit. (n. 6) 255 ff. and 273, against the scepticism of Sanctis, De (Storia dei Romani IV, 1, 121Google Scholar, n. 16) about these numbers.

⁹ He uses πλοῖα, νῆες and σκάφη (originally ‘ships’ hulls') as general terms for ships including warships, and if in one place he speaks of μακρὰ πλοῖα it does not follow that πλοῖα alone in another passage does not mean precisely the same. At Ecnomus (Polyb. 1, 25, 7) the Romans put out to sea with 350 μακραὶ νῆες κατάφρακτοι to invade Africa. The Punic fleet came to meet them with 330 νῆες κατάφρακτοι; clearly the omission of μακραί in the second passage does not indicate any difference between the two sorts of warship. In Polyb. 1, 52, 6, L. Junius Pullus reaches Syracuse in 249 with 120 σκάφη and 800 νῆες φορτηγοί in III, 41, 2 P. Scipio sets out for Spain with 60 νῆες, while Ti. Sempronius leaves for Africa with 160 σκάφη πεντηρικά; but Scipio's-fleet no doubt also contained quinqueremes and it is even doubtful if Polybius' phraseology can be taken to exclude their all being quinqueremes.

¹⁰ Cercuri are not mentioned in the surviving parts of Polybius, but this is probably due to chance, since they were a widely-used type of non-combat galley. Rome had them in the First Punic War (Diod. XXIV, 1), and earlier Alexander had some on the Hydaspes (Arrian, , Anab. VI, 2, 4Google Scholar). Athenaeus (v, 208 f.) mentions one as an auxiliary service boat for a Syracusan grain ship; cercuri are standard craft for grain transport on the Nile (P. Tebt. 856: 171 B.C.); and finally, according to Appian, Pun. 75, L. Marcius Censorinus crossed over to Utica in 149 with a fleet which included 50 quinqueremes, 100 hemioliai and many ἄφρακτα, κέρκουροι and στρόγγυλοι. Appian probably goes back indirectly to Polybius here (cf. App., Pun. 132, on the sack of Carthage: καὶ τάδε μὲν Πολύβιος αὐτὸς ἀκούσας συγγράφει; Schwartz, RE, ‘Appianus’, col. 220). It may therefore be safely assumed that in the passage from which Livy here derives Polybius had κέρκουροι καὶ λέμβοι.

¹¹ Cf. Polyb. I, 52, 6 (see above, n. 9).

¹² Cf. the 800 of L. Junius Pullus in 249 (Polyb. I, 52, 6).

¹³ Cf. Thiel, op. cit. (n. 6) 273, 287.

¹⁴ The Polybian equivalent of rostratae naves is uncertain, for ἔμβολος, the Greek for rostrum, only occurs twice in a naval context in the surviving parts of Polybius (VIII, 4, 4; XVI, 5, 2; it is found at VI, 53, 1–2 and passim in the plural, meaning the Rostra in the Forum), and there is no derived adjective to correspond with the fairly common Livian phrase rostratae naves. However, Casson suggests plausibly that rostratae naves may be Livy's convenient term for ‘combat galleys’ (as opposed to non-combat craft, which would include not only onerariae but also sine rostris speculatoriae, cercuri and some lembi); if this is so, Antiochus will have brought back both νῆες κατάρακτοι and combat ἄφρακτα.

¹⁵ Appian's account (Syr. 22), also based indirectly on Polybius, is a little different, for he makes Polyxenidas' total 200 (διακοσίαις ναυσί, κουφοτέαις τῶν πολεμίων παρὰ πολύ). Thiel, following Kromayer, accepts Appian's figure and assumes a MS. error in Livy (c for cc); the suggestion was originally made by Sigonius. They argue that once the Roman and Pergamene fleets had combined to give a total of 105 tectae and nearly 50 apertae (Livy XXXVI, 43, 13 (Polyb.)), Polyxenidas can hardly have been ‘occasione pugnandi laetus’ if he had only 70 tectae and 30 apertae, all minoris formae. But Polyxenidas may have been confident in his seamanship, and if in fact he had 130 open ships, it is odd they play no part in the later battles (as Thiel points out, against his own view). More decisive is the fact that when the Syrians hoped to face a Roman fleet of 81 tectae and 20 apertae Polyxenidas refers to their numbers being approximately equal (Livy XXXVI, 43, 5, quoted in the text above), which would make little sense if the Syrians had 70 tectae and 130 apertae. Livy's figure is therefore to be accepted.

¹⁶ Thiel, op. cit. (n. 6) 274, n. 345; Kromayer, , Antike Schlachtfelder II, 157Google Scholar, n. 4.

¹⁷ It may be asked: could Hellenistic triremes be νῆες κατάφρακτοι? In one passage Polybius seems to allow the possibility that they might. In v, 62, 3 he describes how Antiochus III took over various war materials in Coele-Syria, including 40 πλοῖα; of these 20 were κατάφρακτα, ἐν οἶδὲν ἔλαττον ἦν τετρήρους, and the rest were τριήρεις, δίκροτα and κέλητες. This may be simply an explanatory phrase, viz. there were 20 cataphracts, that is, none less than a quadrireme. (That quadriremes were νῆες κατάφρακτοι is clear not only from this passage, but also from Polyb. XXXIII, 13, 1–2: Athenaeus set sail with 80 νῆες κατάφρακτοι of which five were Rhodian τετρήρεις (τετρήρεις Z, Dindorf; τριήρεις was Ursinus' false reading)). But it reads rather as if some of the κατάφρακτα could have been smaller than quadriremes, and Casson remarks, in a letter, that in lighter navies (such for instance as the Rhodian), which put triremes in the line, such triremes were still κατάφρακτοι, as they had been in the fifth and fourth centuries. However, he insists that it would have made nautical and naval nonsense were Antiochus to have tried to use 70 triremes as ships of the line in 190. On the whole question see L. Casson, op. cit. (n. 2).

¹⁸ On the difference between these craft see Polyb. 1, 63, 8.

¹⁹ It has been suggested to us that the difference between maior and minor forma, corresponding to the general build of the hull, may depend upon oarage. In Athen. v, 36, 203d, describing Ptolemy II's fleet (it may be from Callixeinus, but Jacoby in FGH 627 seems doubtful) τὰ μέγιστα τῶν πλοίων include categories ranging from τριακοντήρεις down to πεντήρεις, and are distinguished from τὰ…ἀπὸ τετρήρους μέχρι τριηρημιολίας, which are διπλάσια τούτων; and this may suggest a substantial difference between quadriremes—perhaps developed by some minor modification of the trireme such as a second man on the zeugite oar—and quinqueremes with a more radical reorganization of the rowing arrangements, with several men to an oar. This, it is pointed out, would involve oars of a larger and different pattern and, hence, a different build of hull. Without entering into the controversial question of the rowing arrangements in quadriremes and quinqueremes we would observe, however, that this suggestion does not account for the main problem here, viz. that of the 70 decked ships bigger than triremes and so almost certainly quinqueremes—but minoris formae. Hence the difference in build can scarcely be directly related to the number of men to an oar. (Tarn, op. cit. (n. 7), 130 n. 1, also thought the difference was one of oarage; but he does not face the problem of these 70 tectae naves minoris formae either. In fact he does not mention them.) We would tentatively suggest that with the introduction of ships rowed with several men to an oar, the hull was made much heavier than before, but that experience showed that the new arrangement could be achieved with ships of less heavy build than had at first been thought. Lionel Casson has pointed out to us the analogy of the Roman warships of the First Punic War; there too one can detect a difference between the Roman quinqueremes of Ecnomus, built to carry the corvus (cf. Polyb. 1, 22, 3), and those of the Aegates Islands, built after the model of Hannibal's blockade-runner at Drepanum (cf. Polyb. 1, 47, 10; 59, 8).

²⁰ In Appian (above, n. 15) they are simply κουφότεραι τῶν πολεμίων παρὰ πολύ.

²¹ On this see Thiel, op. cit. (n. 6) 275; Casson as above, n. 19.

²² Another example of Livy's inconsistency in rendering technical terms is to be found in a comparison between XXXII, 32, 9 and XXXV, 26, 1; in the former passage he renders Polybius' πρίστις; (Polyb. XVIII, 1, 1) by navis rostrata, yet in the second he is content to transliterate the Greek word pristis.

²³ Thiel, op. cit. (n. 6) 275–6, rightly points out that Antiochus' forces were no match for the united fleets of Rome and her allies, when these were fully deployed (though, as we have seen, his assumption that Polyxenidas' ships of the line in 191 were triremes does that admiral some injustice). He also criticises Antiochus for ‘a certain lack of maritime energy and of sound, solid maritime traditions’ (since Phoenicia had become Seleucid only in 200 and Cilicia only in 197). But he also points out the Roman difficulties which led to their fighting Side and Myonnesus against odds, and our point is not so much the real threat presented by Seleucid naval power, as the nuisance value it had had throughout the war.

²⁴ See Hammond, N. G. L., JRS, LVIII (1968), 1–12Google Scholar, especially 7, n. 24; on Illyrian piracy see Dell, H., Historia XVI (1967), 344–58Google Scholar. The idea of a geographical sailing limit appears in earlier Carthaginian treaties with Rome (first treaty, Polyb. III, 22, 5–6; second treaty, Polyb. III, 24, 11); and in the treaty between Rome and Tarentum (Appian, , Samn. 7, 1Google Scholar). The Romans, having suffered under it, now adapt it to their own purposes.

²⁵ Plutarch, Flam, 9, 6 says ‘ten’; but Polybius (confirmed by Livy) is the better source.

²⁶ Livy XXXIII, 30, 5 adds that the ‘sixteen’ was ‘inhabilis prope magnitudinis’, but Plutarch (Dem. 43) says that Demetrius' ‘fifteens’ and ‘sixteens’ were swift and effective craft. By 197, it is true, this surviving specimen may have become somewhat decayed (cf. De Sanctis, op. cit. (n. 8) IV, 1, 95, n. 184), and its special mention in the peace-terms may not imply a great naval concession. Casson reminds us that Philip had not used it in his naval campaigns of 201. See further Thiel, op. cit. (n. 6) 251, n. 263; Tarn, op. cit. (n. 7) 132 ff. Tarn is, however, wrong in identifying Philip's ‘sixteen’ with Ptolemy Ceraunus' ‘eight‘, for this ship was from Heraclea (Memnon, FGH 434 F 8 § 5) whereas the ‘sixteen’ was built in either Macedonia or Greece (Plut., , Dem. 43, 3Google Scholar). Walbank's note on Polyb. XVIII, 44, 6 (Commentary on Polybius II, 612) should be corrected accordingly.

²⁷ See above, n. 9.

²⁸ Cf. Thiel, op. cit. (n. 6) 249, n. 255, who shows that Philip must still have possessed between 20 and 25 warships (and many lembi); the 15 to 20 must have been destroyed or (Thiel suggests) handed over to Pergamum and Rhodes.

²⁹ To some extent the treaty with Teuta is a fore-runner, for she was restricted in the type and number of craft she might send south of Lissus; but this was a small-scale settlement and there was no absolute limitation on the number of ships she might possess.

³⁰ Moore, J. M., The Manuscript Tradition of Polybius (Cambridge, 1965), 140 ff.Google Scholar, 171 ff.

³¹ We cite Büttner-Wobst (ed.), Polybii Historiae IV, in ap. crit. ad loc., and have consulted Dr. Moore, to whom we owe our thanks.

³² Mommsen, Th., Römische Forschungen II (Berlin, 1879), 533Google Scholar.

³³ McDonald, A. H. (ed.) Titi Livi a.u.c. (O.C.T.) v, lib. 31–35 (Oxford, 1965)Google Scholar, praef. viii–xxxvii.

³⁴ Loc. cit. xxxvii–xxxix.

³⁵ Gronovius and Meibomius as cited by A. Drakenborch (ed.), Livius (1743), in his note on XXXVIII, 38, 8.

³⁶ Madvig, J. N., Emendationes Livianae (ed. 2, 1877), 547–8Google Scholar, followed by subsequent editors.

³⁷ Mommsen, loc. cit. (n. 32) 533.

³⁸ To anticipate objections on the grounds that Polybius avoided hiatus, we refer to Biittner-Wobst, (op. cit. (n. 31) IV, xxiv): ‘Zitiert Polybios aus anderen Autoren oder führt er Urkunden wörtlich an, so weicht er bewusst von diesen Gesetzen ab’ (i.e. from his rules for the avoidance of hiatus).

³⁹ Moore, op. cit. (n. 30) 151.

⁴⁰ Appian's source is uncertain. The general problem is discussed by Gabba, E., Rend. Line.⁸ XII (1957), 339–5Google Scholar ‘Sul Libro Siriaco di Appiano’ (with previous views and bibliography on 339—40, n. 1); see also, on Appian's procedure in the Macedonica, Meloni, P., Il valore storico e le fonti del libro macedonico di Appiano (Rome, 1955)Google Scholar, summary on 216—22, with the criticism of Gelzer, M., Kleine Schriften in (Wiesbaden, 1964), 280–5Google Scholar. It is by no means certain whether Appian is here following Polybius even indirectly (so Gabba, and earlier, E. Schwartz, RE, ‘Appianus (2)’, 219–21; direct use seems excluded). If he is, his καταφράκτους may indicate that the corruption was in some texts of Polybius (though not in Livy's) at an early date; but it could of course have come in via an annalist intermediary. There are, however, other possibilities, and in any case what Appian says is so far from Polybius and so absurd—he grants Antiochus 12, not 10 ships, and says that he was allowed these in order to make war on his dependants and could have more if he was attacked—that it would be rash to regard him as affording independent support for καταφράκτων in our text of Polybius.

⁴¹ Niese, B., Geschichte der griechischen und makedonischen Staaten III (Gotha, 1903), 171, 244Google Scholar; G. De Sanctis, op. cit. (n. 8) IV, 3, 113; ‘di far eseguire gli articoli del trattato concluso dopo la battaglia di Magnesia concernenti le navi e gli elefanti da guerra che Antioco Epifane, con la tacita tolleranza del senato romano, aveva dimenticati’.

⁴² above, nn. 9, 17, 25.

⁴³ Mommsen, op. cit. (n. 32) 534.

⁴⁴ loc. cit. (supra, n. 35).