¹ By ‘the accepted theory’ in this paper I mean the group of solutions (they are legion) which, though differing in details of arrangement, agree in this, that a trireme had three banks of oars at a substantial interval one over the other, a quinquereme five, a dekeres ten, and so forth, each oar rowed by one man and the lowest bank fairly near the water. (I do not include Bauer, or so much of Assmann as relates to breit-polyereis.) All these solutions rest on a common basis and fall together if that be destroyed. The most important current expression of this theory, beside Mr.Torr's, , is Assmann's hoch-polyeres theory (art. Seewesen in Baumeister, and several papers, notably Jahrb. 1889, p. 91Google Scholar, Zur Kenntniss der Antiken Schiffe), followed by Droysen, Griechische Kriegsaltertümer in Hermann's Lehrbuch; Luebeck, , Das Seewesen der Griechen und Römer, 2 vols. 1890Google Scholar; and Schmidt, , Ueber griechische Dreireiher, 1899Google Scholar; to judge by Luebeck's article biremis, it will be adopted in the new Pauly-Wissowa. Bauer's theory (Griechische Kriegsaltertümer in Müller, 's Handb. d. klass. Alt.-Wiss., 1893Google Scholar, and several papers), that a trireme had a very slight interval between the banks and that ships larger than triremes never had more than three banks but employed more than one man to an oar, is quite a separate matter. Important is Admiral Fincati, 's Le Triremi, 1881Google Scholar; a trireme had three oars to one bench, like a Venetian galley a zenzile. I unfortunately only know this book in Serre's translation, at the end of Vol. 1 of his Marines de la guerre, 1885 and 1891, from which I cite it. I cannot classify Admiral Serre; though accepted, I believe, in France, his views seem to bear little relation to the evidence. Weber, 's book Die Lösung des Trierenrätsels, published 1896Google Scholar, but written much earlier, with many blunders and mistranslations, contains ideas. A trireme had three men to an oar, a quinquereme five, etc. Accepted by Speck, , Handelsgeschichte, 1900Google Scholar. Weber has no monopoly in mistranslations. The best exposition of the accepted theory prior to Assmann is probably that of Cartault, , La Trière Athénienne, 1881Google Scholar. I understand he afterwards agreed with Bauer. While this paper was in the press two important articles appeared: one by Mr. Torr in Dar.-Sagl. s.v. navis, which seems to state his version of the accepted theory more definitely than was done in Ancient Ships; the other by Mr. A. B. Cook in Whibley's Companion to Greek Studies, who favours the Venetian theory, but not very decidedly. References to Torr in this paper are to Ancient Ships unless otherwise stated.

² I have had to notice the boats on Trajan's column, and one or two other matters, and, of course, writers of later date.

³ A trireme a zenzile was one in which three men sat on one bench on the same level, one a little astern of the other, each rowing one oar, the three oars issuing through one opening side by side, and giving the appearance of a bundle of three oars (see Figs. 1 and 2). In the galley a scaloccio several men rowed each oar.

⁴ The monumental evidence is often overrated. Even in the case of the best monuments, one can never say how far the artist may have sacrificed truth of detail to artistic considerations. It will be considered under E.

⁵ However little one wishes to dogmatise, one cannot always be writing in the potential mood and expressing every shade of proper reservation.

⁶ By ‘the larger polyereis’ in this paper I generally mean quadriremes to dekereis both inclusive, nothing over a dekeres being heard of in action.

⁷ A is very old as an opinion. B and a good deal of D (2) are new, I think. C (1) is given correctly by Bauer. D (1) is primarily Weber. In referring in this paper to Bauer's arrangement I mean his arrangement considered physically, i.e., as a slight interval only between the rows, apart from questions like the meaning of thranite or παρϵξϵιρϵσία.

⁸ τῷ τὴν ἐμβολὴν εἶναι κατὰ τὰς πρώτας θρανίτιδας κώπας The only writer known to me who cites this passage is Breusing, , Die Lösung des Trierenrätsels, 1889Google Scholar; and as he could not understand it at all, he said that the words from τῷ τὴν ἐμβολὴν to the end must be a gloss. If one may discard everything as a gloss that does not suit one's own theory, one can prove anything. No one who has seen a bumping race, and watched the cox of the boat in front washing off the nose of the boat behind with his steerage, will have any difficulty in construing the passage. I quote Polyaenus throughout from Woelfflin-Melber. He made considerable use of Ephorus; but according to Melber, , Ueber den Quellen und der Wert der Strategemensammlung Polyäns, (1885)Google Scholar, the passages most material to this paper (5, 43; 3, 11, 7 and 12 and 13; 5, 22, 2) are derived from some earlier work on naval tactics.

⁹ Assmann has to translate it (Baumeister, 1616) ‘neben den hintersten Thranitenriemen,’ which is not in the Greek.

¹⁰ This passage, unlike the former, is not evidence against anyone but those who accept Assmann's view (based on the monuments) of the παρεξειρεσία as an outrigger or ‘oar-box’ (Riemen-kasten); however, as it is conclusive that Assmann is right on this point, this is not very material. Chabrias' new steering oars were not where the old ones were. The new ones were through the παρεξειρεσία therefore, the old ones were not. But the old ones were in the usual place on the stern of the ship, as shewn by their lifting clear of the water; therefore the old view, that by παρεξειρεσία is meant the stern (and bow) of the ship beyond the oarage, is untenable. The same conclusion is supported by Peripl. Pont. Eux. 3, the waves coming in not only through the oar-holes but over the παρεξειρεσία (where the reference must be to a higher point, not a different point); amd by the frequent references to ships losing part of their παρεξειρεσία in action (Thuc. 7, 34 is a good instance). But the absolutely decisive passage is Polyaen. 3, 11, 13; Chabrias stretches skins over the παρεξειρεσία of each side of the ship (ὑπὲρ τὴν παρεξειρεσίαν ἐκατέρου τοίχου) and nails them to the deck above, thus making a Φράγμα which prevented the waves washing in and the oarsmen looking out. Chabrias here improvised a cataphract. Assmann never really proved his own theory of the παρεξειρεσία at the same time there is nothing in Buresch's, attack on it, Die Ergebnisse der neueren Forschung über die alten Trieren (Woch. für klass. Phil. 1891, No. 1)Google Scholar.

¹¹ In a Rhodian inscription of the first half of the first century B.C. (I.G. xii. fase. i. No. 43) trihemioliai are contrasted with cataphracts, and again triremes with aphracts;. suggesting that the trihemiolia was then a smaller or less important ship than a trireme. The form τριηρημιολία (Ath. 203 d) suggests that Photius is right in calling it a trireme; if so, it was a light trireme evolved from a hemiolia (as to which see n. 22), as the trireme from the pentekontor. The suggestion that it means a ship of 2½ banks is the merest guesswork.

¹² κατὰ μέσον τὸ κύτος ὑπὸ τὸν θρανίτην σκαλμόν Cited by Weber.

¹³ As I shall often have to refer to the battle of Chios, I should note that some writers (e.g. Beloch, Bevölkerung, and Ihne) doubt the accuracy of Polybius' version, obviously diawn from Rhodian sources, that it was a defeat for Philip. But even if so, this cannot affect the details of single events, which are precisely given; for even if the Rhodians wrote up an account of the battle for the honour of Rhodes and Theophiliscus, they would take all the more care to put in details that either did happen or might, consistently with nautical probability, have happened. The account of this battle is hardly affected by Polybius' supposed inaccuracies as to the first Punic war, for which his sources were far different. One cannot go into the case for Polybius in a note; but I would point out (1) that, as to the numbers, no one, I think, haa, yet examined the numbers in the sea-fights generally up to Actram, and the only examination for part of the period that I know of—Kromayer, , Die Entwickelung der röm. Flotte vom Seeräuberkriege des Pompeius bis zum Schlacht von Actium (Philol. 1897), p. 426Google Scholar—accepts the great numbers recorded for the war with Sextus Pompey; (2) that to bring in the population question (Beloch, Serre) is surely to explain obscurum per obscurius; (3) that the real exaggeration is not in the separate accounts of battles, which generally mention ‘ships’ or ‘cataphracts,’ but in the summing-up chapter (1, 63), where Polybius has used πεντήρεις when he ought to have said warships, as appears both from the separate accounts and from the columna rostrata (C.I.L. 1, 195); and (4) that Ihne's objection (Röm. Gesch.² 2, 47) that the Romans had ships before the first Punic war, neglects the obvious explanation that Polybius or his authority means no more in speaking of the creation of the Roman fleet than we might in speaking of the creation of the German fleet—a first serious bid for seapower. See also n. 91.

¹⁴ Schol. Frogs 1074 τῷ θαλάμακι· τῷ κωπηλατοῦντι ἐν τῷ κάτω μέρει τῆς τριηροῦς· τῷ θαλάμακι· οἱ θαλάμακες ὀλίγον ἐλάμβανον μισθὸν διὰ τὸ κολοβαῖς χρῆσθαι κώπαις παρἀ τὰς ἄλλας [Γ] τάξεις τῶν ἐρετῶν ὅτι μᾶλλον ἦσαν ἐγγὺς τοῦ ὕδατος. ∥ ἦσαν δὲ τρεῖς τάξεις τῶν ἐρετῶν· καὶ ἡ μὲν κάτω θαλαμῖται, ἡ δὲ μέση ζυγῖται, ἡ δὲ ἄνω θρανῖται. θρανίτης οὖν ὁ πρὸς τὴν πρύμναν, ζυγίτης ὁ μέσος, θαλάμιος ὁ πρὸς τὴν πρῷραν. (I cite down to || from Rutherford's ed. of the scholia (1896); he does not give the latter half, which is therefore not in the codex Ravennas. I cite it from the codex Venetus. In the former half, according to the facsimile published by the Hellenic Society, cod. Ven. omits Γ̅.) Schol. Ach, 162 τῶν ἐρεττόντων οἱ μὲν ἄνω ἐρέττοντες θρανῖται λέγονται, οἱ δὲ μέσοι ζυγῖται, οἱ δὲ κάτω θαλάμιοι. Hesych. θρανίτης ὁ πρὸς τὴν πρύμναν, ζυγίτης ὁ μέσος, θαλάμιος ὁ πρὸς τῄ πρῴρᾳ (so Suidas and Zonaras), Hesych. θαλάμιος ἐρέτης δ κατωτάτω ἐρέσσων ἐν τῇ νηἰ θαλάυιος λέγεται, δ δὲ μέσος ζύγιος, δ δὲ ἀνώτατος θρανίτης. θαλάμιαι κῶπαι οἱ κατωτάτω καὶ οἱ ταύτην ἔχοντες τὴν χώραν θαλάμιοι λέγονται. Suidas. θρανίτης λεώς τῶν γὰρ ἐρεττόντων οἱ μὲν ἄνωθρανῖταιλέγονται, οἱδὲμέσοιζυγῖται, οἱδὲκάτω θαλάμιοι. Etym. Mag. θαλαμίδιοι κῶπαι ὁ κατώτατος ἐρέτηςθαλάμιος λέγεται, δ δὲ μέσοςζύγιος, ὁ δὲ ἀνώτατοςθρανίτης. Eustath. 1818, 52 ἔχειδέ, φησίν (Pausanias), οὗτος (thranite) τὴν ἄνω ἕδραν, τὴν δευτέραν ζύγιος, τὴν τρίτην θαλάμιος. 640, 11 θαλαμῖται καὶ θαλάμακες ἐρέται οἱ ὑπὸ τοὺς θρανίτας. Lastly Pollux 1, 87 καλοῖτο δ᾿ ἂν καὶ θάλαμος οὗ οἱ θαλάμιοι ἐρέττουσι τὰ δὲ μέσα τῆς νεὼς ζύγα, οὗ οἱ ζύγιοι καθῆνται τὸ δὲ περὶτὸ κατάστρωμα θρᾶνος, οὗ οἱ θρανῖται. —There is another scholion on Frogs 1074, given by Zuretti, , Scolii al Pluto ed alle Rane d'Aristofane dal codice Veneto 472Google Scholare dal codice Cremonense 12229, L, 6, 28: τρεῖς τάξεις ἦσαν ἐν τῇ τριήρει· οἱ μὲν πρῶτοι θρανῖται καλούμενοι, οἱ δὲ δεύτεροι ξυγῖται, οἱ δὲ τρίτοι θαλαμακες.Read with Eustath. 1818, 52, this illustrates the use of πρῶτος as sternmost in Polyaen. 5, 43 above.

¹⁵ Unless it be Ar., Mech. 4Google Scholar, discussed under F.

¹⁶ That ἄνω and κάτω mean ‘astern’ and ‘in the bows’ has often been asserted but never proved. I believe it is correct, but my reason for thinking so is given in B; it has nothing to do with the Schol., on Frogs 1074Google Scholar. If it be correct, all the ἄνω and κάτω passages given in the note are disposed of conclusively.

¹⁷ This forced Assmann to explain e.g. a dekeres as constituted by three superposed triads, each triad consisting (in superposition) of a thranite, zugite, and thalamite; with a lonely thalamite on the top. The τεσσερακοντήρης is legimate evidence so far as it goes. Since the inscription about the τριακοντήρης was found, no one can suppose it to be a bad joke of Callixenus’: and the idea that it was a kind of flat-bottomed river barge (Assmann, Droysen, Torr) seems to me to be disposed of by the fact that Philopator had such a barge (the θαλαμηγός of Ath. 204 d. seq.). If any one will read Athenaeus consecutively he will see that he puts side by side three monsters of three different types; the τεσσερακοντήρης (long ship), the θαλαμηγός (ποτάμιον πλοῖον), and Hiero's ship (round ship). The height of the tesserakonteres, on which rests the ‘Mississippi steamer’ theory, is given to the top of the ἀκροστόλιον, which (pace Liddell and Scott) is not the gunwale, see Torr, 68. Those who treat δίπρωρος as ἀμφίπρωρος have forgotten the old Calais-Douvres; and the twin hull was only the logical outcome of the common practice of lashing two ships together to get a steady platform.

¹⁸ If there was a visible triple beat on any view, what becomes of the stock comparison with the wings of a bird?

¹⁹ δίκροτοι αὐτῶν (i.e. τῶν μακρῶν νεῶν) τὰς κάτω κώπας οὐκ ἐπι πολὺ ἔξω ἔχουσαι τοῦ ὕδατος

²⁰ Curtius, Diodorus, and Justin are silent. Plutarch, (Alex. 63)Google Scholar says he built πορθμεῖα κωπήρη καὶ σχεδίας

²¹ For cercurus see Torr s.v.; a type equally suited for warfare or commerce, but always-reckoned among the small craft of a fleet; he has a lot of evidence. Weber's idea that a, cercurus was a trireme is a mere mistranslation of App., Pun. 121Google Scholar.

²² App., Mith. 92Google Scholar the pirates originally (πρῶτον) used myoparones and hemioliai, later (εἶτα) δικρότοις and triremes, i.e., when they organised themselves. This is conclusive; and overrules Hesych. ἡυιολία· ἡ δίκροτος ναῦς, where the definite article makes nonsense anyhow. I want to make this clear, because the accepted explanation of ἡυιολία is a ship with 1½ banks. There is not a shred of evidence for this; it rests on the fact that ἡυιόλιος means 1½. I might say that hemiolia means a ship of 1½ squads, which has at least the support of Photius s.v. υὗ τὸ ἡμιόλιον μέρος ψιλὸν ἐρετῶν ἐστὶ πρὸς τὸ ἀπ᾿ αὐτοῦ μάχεσθαι The certain thing is that it was a pirate ship (Arr., Anab. 3, 2, 4Google Scholar, App., Mith. 92Google Scholar, Phot, s.v.), and a typical one (Theophr., Char. 25, 1Google Scholar), and could be classed with the little myoparones, which were certainly single banked (evidence Torr 119); it was a favourite for surprises (Diod. 19, 65, Polyaen. 4, 7, 4); and the latter passage also shews it was small, the object of Demetrius being to display the minimum of force. Pirates, whose heads depended on their speed, would not go in for fancy arrangements of oars.

²³ The word occurs in the active sense once, in a chorus (Eur., I. T. 407)Google Scholar, δικρότοισι κώπαις, of the Argo, (a traditional single-banked ship, Ap. Rhod. 1, 394 seq.), where it refers to the beat of the oars on either side of the ship. This shews that in Euripides' time it cannot have been a technical term for the beat of two banks on the same side of the ship.

²⁴ The same causes which compelled the Venetians to divide the crew of a trireme into 3 squads and work as a rule in relays (Fincati p. 167) would have compelled the Greeks also to do this. Part of a crew did row alone, (Thuc. 3, 49; Polyaen. 5, 22, 4; Xen., Hell. 6, 2, 29)Google Scholar; but these passages do not shew which part. If, however, when not in action, one squad only rowed at a time, as at Venice, it is explained how the Athenian horse-transports, with 60 oars only, kept up with triremes.

²⁵ i.e., when used as technical terms; for Thuc. 7, 65 (the Syracusans covered with hides τὰς πρῴρας καὶ τῆς νεὼς ἄνω) might refer to the upper works of the ship. As to οὐκ ἐπὶ πολὺ ἔξω ἔχουσαι τοῦ ὕδατος, the forward oars would of course suffer most in the bad water. But it may be that these triakontors, built for a river, were even lower in the water than usual, and anyhow they would be heavily laden. Some were lost going down from Patala.

²⁶ ‘In the Odyssey κατά is the regular word for motion inwards, ἀνά for motion outwards;’ Mr.Myres, J. L., J.H.S. xx. p. 140Google Scholarsq. For later Greek, Mr.Oppé, A. P., J.H.S. xxiv. p. 225Google Scholarsq. Mr. G. F. Hill kindly furnished me with these references. If the ship was generally entered from the stern, this would explain why κάτω should be fore and ἄνω aft; and at Athens anyhow she would be entered from the stern, if launched bow first; see Prof.Gardner, E. A., Ancient Athens, p. 553Google Scholar. This is also borne out by the ordinary term for ‘to come forward,’ ἀναφέρειν τὴν κώπην, which shows that ἀνά is motion toward the stern.

²⁷ Incidentally, this may suggest that Schol., Frogs 1074Google Scholar represents a genuine tradition, i.e., one descended from a time when men knew the technical meaning of κάτω for of course I do not suppose that the Scholiast knew this, any more than Eustathius, and all that I can attempt to shew is what the word meant to Arrian, or rather to Ptolemy.

²⁸ Bauer, , Neue Philol. Rundschau 1895, p. 265Google Scholar, ‘in Schräg vom Hinterschiff zum Vorschiff abfallender Linie augeordneten Ruderpforten,’ which may well be right. It is clearly shewn in the Venetian triremes in Fig. 1. See Aesch., Agam. 1617Google Scholar, and n. 80.

²⁹ The inclination of the παρεξειρεσία to the long axis of the ship (n. 118) would furnish another explanation. The longest oars of the tesserakonteres were thranite oars, as the reference to the lead shows (Ath. 203 f); but as we have no idea how she was arranged, it is useless either suggesting explanations of this or drawing deductions from it as to triremes. All her thranite oars were not of the same length.

³⁰ C.I.A. vol. 2 part 2, 791 1. 56— θρανιτίδων τούτων ἀποφαίνει ὁ δοκιμαστὴς ζυγίας Δ. If the number of the θρανίτιδες that were ἀδόκιμοι were extant, we might have something to go on as to the relative lengths; for as most oars go at the leather, or point of contact, then if only a few could be used as ζυγίαι we should know that any theory (like Assmann's explanation of the Lenormant relief) which made the zugite oars less than two-thirds of the length of the thranite, was, on this ground alone, untenable. The higher pay of the thranites probably had nothing to do with the length of the oar, (that is a Scholiast's guess), but was merely one sign of the greater consideration they enjoyed; and the primary reason no doubt (apart from any question of their more probably being burgesses) was that it depended largely on them, as the stern oars, whether the boat was ‘together’ and kept her pace. Great importance was attached to the manning of the stern benches in a mediaeval galley, as Jurien de la Gravière shews. The Athenian lists do not really prove anything at all as to the relative length of the oars, as we do not know why those ten were condemned; and we have no right to make them mean that all thranite oars were longer than all zugite oars, still less that they were much longer.

³¹ See Kroker, , Die Dipylon-vasen (Jahrb. 1886), wich whose account (p. 106Google Scholarseq.) of the first evolution of the warship I agree, as against Pernice's, criticism in Ath. Mitth. 17 (1892), p. 306Google Scholar.

³² It does not occur in the Athenian lists, and plays no part in battles again. I do not mean it was not built at all; Mithridates e.g. had a few, and see Polyb. 1, 20, 14 (the Italiot states), 25, 7, 1 (Egypt).

³³ Athenian lists; Arrian l.c. and 7, 19; Polyaen. 3, 9, 63; etc.

³⁴ If indeed the triakontor was not originally a pirate, Thuc. 4, 9.

³⁵ See n. 22.

³⁶ See n. 11.

³⁷ Demetrius had lemboi at the siege of Rhodes (Diod. 20, 85), but we do not hear of them in action (if Diodorus be correct neither he nor Ptolemy put μονήρεις into line at Salamis), and so cannot say if they were the Illyrian lemboi or not. Polyb. 1, 53, 9, and 3, 46, 5 (Hannibal crossing the Rhone) add nothing, and earlier mentions of lemboi refer to ship's boats. Polybius is clear as to Philip's fleet of lemboi being almost a new thing (5, 109, σχεδὸν πρῶτος τῶν ἐν Μακεδονια βασιλέων) and as to his tactics at the battle of Chios being new. We may conclude that if he was not actually the first to introduce the Illyrian lembos he was the first to perceive its possibilities and to use it in a fleet action.

³⁸ Polyb. 21, 45 μηκέτι ἐχέτω πλὴν ί καταφράκτων· μηδὲ τριακοντάκωπον ἐχέτω ἐλαυνό· μενον κ.τ.λ. Livy 38, 38 has run the two together (neve plures quam decem naves actuarias nulla quarum plus quam triginta remis agatur habeto), while App., Syr. 39Google Scholar mentions cataphracts only.

³⁹ See post, n. 94 as to Philip's ‘lembi biremes,’ and ‘double-banking.’

⁴⁰ Precisely the ‘galeotta’ of Furtenbach. No doubt someone experimented with biremes before triremes were invented. But these experiments remained without effect (witness the silence of Herodotus, Thucydides, and the Athenian lists, and indeed of all writers prior to Caesar) and have nothing to do with the biremes known to history, which appear first in the 1st century B.C. See under E.

⁴¹ It may be objected that the bireme of Octavian's time was a ‘Liburnian.’ Biremes are mentioned in history earlier than Liburnians, which is all I require; but it is as well to be clear about the Liburnian. In origin, it was another of the light swift pirate-craft of the Adriatic (App., Ill. 3Google Scholar), if indeed it was not the lembos under another name; and the fact that under the Empire the Liburnian was built, first as a bireme (App., Ill. 3Google Scholar, Lucan 3, 534—note Lucan's ‘crevisse,’ it had grown) and later as a trireme, etc. (Veget. 4, 37), which nobody doubts, only shews that there were biremes of two different builds running parallel, the Liburnian bireme evolved from a Liburnian and the dicrotos bireme evolved from a triakontor (just as earlier there were the trireme and the trihemiolia); see C.I.L. 5, 1956 which mentions a ‘bicrota’ called Mars and a ‘Liburna’ called Clupeus. When Appian (Ill. 3) says that in his time light δίκροτα were called Liburnians he shews, either that the two builds had become confounded, or (more probably) that he was ignorant of the process by which the δίκροτος bireme had been evolved, and that for him δίκροτον was simply ‘bireme.’

⁴² This passage is a good instance of one which explains equally well on any theory and is useless to cite. Other good instances are Polyaen. 5, 22, 4 and the drowning thalamites of App., b.c. 5, 107Google Scholar.

⁴³ Strabo 7, 325. ἀνέθηκε Καῖσαρ τὴν δεκαναΐαν ἀκροθίνιον, ἀπὸ μονοκρότου μεχρι δεκήρους. He uses the word to mark the fact that the trophy began, not only with a μονήρης, but with the smallest kind of μονήρης.

⁴⁴ λαβόντα τῶν ναυτῶν ἕκαστον τὴν κώπην κ.τ.λ. Bauer alone has put this correctly. As regards triremes, the passage is conclusive as against Weber (three men to an oar) who has to mistranslate it, and Serre (three banks, but in action only the top bank rowed by three men to an oar), for then Brasidas would not have troubled to take the other oars with him on a mere raid. The large number of oars for a trireme given in the Athenian lists also certainly presupposes one man to an oar. Weber has to say a trireme carried two spare sets, which (apart from the question of weight) is improbable, seeing that the account of battle after battle assumes that a ship with a crippled ταρσός is out of action. The spare oar question is not, however, easy; see e.g. the Hippia, (C.I.A. vol. 2 part 2, 802 c. 6)Google Scholar which is said to have a ταῤῥὸς δόκιμος (not, however, ἐντελὴς δόκιμος) though five oars are broken. Probably Assman's solution is the best (reviewing Schmidt, in Berl. Phil. Woch. 1900Google Scholar, No. 43); the περίνεῳ oars were deck sweeps, carried for use in a ship left crippled. I may add that, with a παρεξειρεσία half carried away, no spare oars but deck sweeps would (on the view I take of a trireme) have been of much use. Possibly however a trireme rowed 25 groups of 3 oars each side, and carried some half dozen spare oars of each class.

⁴⁵ And if we had, it would be a cataphract, and so could not shew any rowers.

⁴⁶ C.I.A. vol. 2 part 2, 812 a 35: οὗ[το]ς τὴμ μὲν τετρήρη ἀποδέδ[ωκ]εν τὰ δὲ σκεύη ὀφείλει διὰ τὸ [ἐπὶ] πεντήρη κατασταθῆναι. The σκεύη here include the ταῤῥός which had been previously mentioned. 812 c 143 seq. Ἡδεῖα. …. [οὗ]τος τὴν τριήρη ἀποδέδω[κεν διὰ τὸ ἐπ]ὶ τετρήρη καθεστηκέναι [τὰ δὲ σκέυ]η ὀφείλει (here follow the συντριήραρχοι) σκεύη ἔχουσι ξύλ[ινα ἐντελῆ]… There is another passage to the same effect, and the filling up of the lacunae is quite certain. Incidentally, this disposes of every reconstruction of a trireme which cannot be expanded into a quinquereme.

⁴⁷ Rarely mentioned, and only in the last extant list.

⁴⁸ Diod. 14, 41.

⁴⁹ Here we undoubtedly meet Assmann's breitpolyereis.

⁵⁰ I shall find it convenient to talk of ships of the fourth century, prior to Antigonus' fleet, simply as ships of the fourth century. It will not create any confusion. For our purpose the third century begins with Salamis.

⁵¹ Aelian, V.H. 6, 12Google Scholar: Dionysius II. had a fleet of 400 ships, hexereis and quinqueremes: this is of course impossible, and it must mean ‘including hexereis and quinqueremes’; see Diod. 16, 19. Even so, the statement as to hexereis is extremely improbable, seeing that Alexander never had anything larger than a quinquereme. Very possibly Dionysius II. had built one hexeres on the fourth century system (whatever it was), as a ‘royal ship.’ The statement of Pliny, N.H. 7, 56Google Scholar, that Alexander invented the dekeres, is valueless; see Luebeck 1, 17 n. 6 and Droysen 272 n. 3, who give the evidence as to Alexander's fleets. It is precisely what would get stated about Alexander, and is on a level with Curt. 10, 1, 19, the 700 heptereis carried over in sections to the Euphrates; this last is refuted, were refutation necessary, by Arr., Anab. 7, 19Google Scholar, who gives the correct version (from Aristobulus).

⁵² Many writers have assumed, on the ground of practical necessity, that in the larger polyereis more than one man rowed one oar; but that is another matter. Serre and Weber try to shew that Ap. Rhod. 1, 396 means two men to an oar; but there is no foundation whatever for this. The passage, a straight-forward one, had already been correctly explained by Cartault.—Possibly the Delos ship of Paus. 1, 29, 1 would be in point, if one knew what the passage meant; but I cannot translate it, and Frazer's translation ‘decked for nine banks of oars’ conveys no meaning to me. Pausanias had of course heard of higher values, and therefore the ship was abnormal in some way; νικήσαντα does not mean ‘larger than’ but ‘more curious than.’

⁵³ b.c. 4, 85 (battle between Sextus Pompey and Salvidienus); οὔτε ἑστῶτες βεβαίως ὑπὸ ἀηθείας, οὔτε τὰς κώπας ἔτι ἀναφέρειν δυνάμενοι.

⁵⁴ ‘Vogue dans laquelle la force sur l'aviron est produite presque tout entière par le poids du rameur, qui, monté debout sur la pédague ou sur le banc qui précède, se jette en arrière, et, tirant à lui son aviron, va tomber assis sur son propre banc.’ The lead may have been used to meet the difficulty of the oars being of different proportionate lengths inboard. How this was met in a Greek trireme does not appear; the only actual reference to lead is with regard to the thranite oars of the τεσσερακοντήρης

⁵⁵ The length of the περίνεῳ oars, 4.4 m., is the only one actually known, but this supplies a kind of limit. Schmidt has an interesting attempt to work out the measurement from the data as to the Athos canal in Herodotus and Demetrius of Skepsis; he makes the longest oars in a trireme 3.3 m. outboard.

⁵⁶ There are of course a great many references to spurting, and the common name for it, ῥοθιάζειν, implies a fast enough stroke to make a good deal of splashing. The celebrated feat of an Athenian trireme, which swung round a merchantman and rammed her pursuer (Thuc. 2, 91) implies a quick lively stroke and a power of backing water on one side only quickly and forcibly. And the fact that a crew could only last a short time in action (e.g., Polyaen. 3, 10, 12, Diod. 13, 77, Frontinus 2, 5, 47) conclusively implies a fast stroke. Chabrias, training rowers for a trireme, trained them sitting; Polyaen. 3, 11, 7: and cf. Aristophanes' reference to ‘that which fought at Salamis.’

⁵⁷ See n. 110.

⁵⁸ There is a fine picture of a mediaeval quinquereme, with 5 men to an oar, on Pl. VII. of Furtenbach's, Architecturu Navalis, 1629Google Scholar; with a huge outrigger, and the oarsmen on their feet. A good description of such a quinquereme in Bigge, , Der Kampf um Candia in den Jahren 1667–1669Google Scholar (Kriegsgeschichtliche Einzelschriften, Heft 26, 1899), p. 130: the men worked in three relays, as in a trireme. I owe the reference to these writers to the kindness of Mr.Anderson, W. C. F.. For the scaloccio galleys generally, see Admiral Jurien de la Gravière, Les derniers jours de la marine à rames, 1885Google Scholar; the different strokes in use (none rowed sitting) are described p. 231 seq., the best of commentaries on Appian and on Lucan, , Phars. 3, 543Google Scholar.

⁵⁹ Quinqueremes run ashore and the crews depart, Polyb. 1, 51; 3, 96; etc. Attalus' royal flagship at the battle of Chios (size not given, but following the usual Hellenistic practice [see too Beloch, , Gr. Gesch. iii. pt. 2, p. 428Google Scholar n. 2] it would be the largest he had, and he had quinqueremes) runs ashore and the king and his crew departed (ἀπεχώρησε); Philip tows her off uninjured (Polyb. 16, 6 and 7). Diodorus 20, 47, Demetrius sails to Cyprus and draws his ships ashore and surrounds them with a palisade and ditch; he had heptereis and hexereis, and no preparation made for drawing them up. Frontinus 1, 5, 6, Duilius' ships (quinqueremes anyhow) cross a boom at Syracuse. Ath. 204 C, the dock of the tesserakonteres was only four cubits deep. Livy 30, 25 is not against this; the quinquereme there was damaged because driven ashore at full speed.

⁶⁰ Lembos small and cannot have had more than one bank: Livy 34, 35, and evidence collected by Torr s.v.

⁶¹ ἐμπιπτόντων αὐτοῖς τῶν λέμβων ποτὲ μὲν εἰς τοὺς ταρσούς … ποτὲ δὲ πάλιν εἰς τὰς πρῴρας …, κατὰ δὲ τὰς ἀντιπρῴρους συνπτώσεις ἐποίουν (the Rhodians) τι τεχνικόν.

⁶² One of the wonders of Hiero's ship was the water-screw invented for her by Archimedes, Ath. 208 f.

⁶³ Sed neque rostro ferire celeritate subterlabentem poterant neque transilire armati ex humilioribus in altiorem navem.

⁶⁴ First mention of turres, battle of Chios (201 B.C.) Polyb. 16, 3, πυργούχων (unless πυρσούχων be the correct reading). The best commentary on Livy here is the battles of Mylae, and Naulochos, in App. b.c. 5Google Scholar. At Mylae, though some of Sextus' ships carried towers, they were on the whole much lower and lighter than Agrippa's; and the point of Sextus' epigram (108), that he had been storming forts, not fighting ships, was Agrippa's turres. He gave orders τι προσθήσειν ἐς τὸ τῶν νεῶν ὕψος, and by the height of the ships turres are clearly referred to, for at. Naulochos all his ships carried turres, and could only be distinguished from Agrippa's by the war-paint (121). This seems to shew that altiorem is quite satisfied by turres. The accounts of Actium shew the difficulty of boarding ships carrying turres.

⁶⁵ Classis Antonii centum Septuaginta navium fuit quantum numero cedens tantum magnitudine praecellens. Nam decem pedum altitudine a mari aberant.

⁶⁶ Battles of Mylae, and Naulochos, in App. b.c. 5Google Scholar; and see n. 64; Dio Cass. 48, 47, 4 and 49, 1, 2: the evidence is overwhelming that for a few years there was a great race in building; not only as regards height, but more especially in weight and thickness, see Plut., Ant. 65, 66Google Scholar. I do not know why it is believed that Octavian had only light ships at Aetium. He had the fleet with which he had crushed Sextus; up to hexereis, Floras 2, 21 (4, 11). Plut., Ant. 62Google Scholar is responsible for the other view; probably adopted to rub in the moral.

⁶⁷ Whether Sextus in fact ever spoke of τειχομαχῆσαι or not, it became a commonplace; see τειχομαχία in Plut., Ant. 66Google Scholar.

⁶⁸ Polyaen. 3, 11, 13 (if the rowers sprang up in a hurry they might upset the ship), presumably refers to a trireme; nor do I lay stress on Lucan, , Phars. 3, 665Google Scholar: if she took in drowning men she might turn over.

⁶⁹ quum inferrentur, demittere in aquam remos ab utroque latere remiges stabiliendae navis causa jussit. App., Syr. 22Google Scholar, gives 3 Syrian ships, not 2,and says that it was they who tried to grapple Livius.

⁷⁰ Size not given, but the flagship of any Hellenistic monarch was always the largest obtainable.

⁷¹ This is presupposed by the pace at which a, fleet could be built; for which there is plenty of evidence (no doubt sometimes exaggerated) beside the first Punic war. Elaborate arrangements for building were not required; Dio Cass. 48, 49, Octavian built ships ἐν πάσῃ τῇ παραθαλασσίῳ Ἰταλίᾳ ; and no doubt the building of the Argo in Ap. Rhod. 1, 363 seq. is copied from current Egyptian practice.

⁷² n. 61 (continues) αὐτοὶ μὲν γὰρ ἔμπρῳρα τὰ ρκάφη ποιοῦντες ἐξάλους ἐλάμβανον τὰς πληγάς, τοῖς δὲ πολεμίοις ὕφαλα τὰ τραύματα διδόντες ἀβοηθήτους ἐσκεύαζον τὰς πληγάς.

⁷³ Like a modern racing yacht. See Frontinus 1, 5, 6; Duilius, to get his ships (including presumably quinqueremes) over the boom at Syracuse, shifts the troops aft, thus raising his bows, and goes at the boom at full speed, shifting the troops forward again at the critical moment. If this be true, a ship with bow raised and stern depressed, i.e. with every angle altered, could still get on a good deal of pace. It has, I understand, been demonstrated that a torpedo boat rushing an ordinary floating boom at full speed may be expected to ‘jump’ it without doing herself any serious injury.

⁷⁴ Cf. Arr., Anab. 2, 19Google Scholar. If this be so, it implies that the ballast was easily got at during action.

⁷⁵ Those who speak of a row of portholes of 10 inches (·25 m.)(Assmann) or any such height above the (normal) waterline cannot really have thought what this would mean. Leaving practical considerations aside, the waterline was no more a constant quantity then than now. Polyb. 1, 60–62, the Carthaginian ships were much hampered by being loaded down with corn and stores which Hanno had trusted to put ashore before engaging. Diod. 20, 49 and 83, Demetrius mounts on the prows of his ships great catapults (τοὺς τρισπιθάμους τῶν ὀξυβελῶν), and of course ballasted the sterns accordingly. So Duilius', corvi. App. b.c. 5, 121Google Scholar, Sextus Pompey's men throw over the turres when escaping, shewing that they had been too low in the water. See too an appendix to Kromayer's, article in Philol. for 1897Google Scholar, before cited, on the great numbers of troops that could be carried at a pinch.—I do not give cases, like Marcellus' sambucae before Syracuse, where the ships were not in action: though Marcellus' quinqueremes could still be rowed: Polyb. 8, 4(6).

⁷⁶ Quinqueremis Romana, seu pondere tenacior, seu pluribus remorum ordinibus scindentibus vortices, quum facilius regeretur, duas triremes suppressit, etc. (For ordines remorum see under E). A little before, Livy had said she was slower than a trireme. Fincati p. 158: according to Nicolo Surian (1583) a quadrireme a scaloccio could beat a trireme a scaloccio but not a trireme a zenzile. It is just possible that these triremes were a scaloccio (n. 120) and owed their pace to the greater skill of the Carthaginians; but I think most improbable.