⁷⁷ According to Luebeck, its first modern supporter was Scaliger, relying on Schol., Frogs, 1074Google Scholar. But it existed when De Baif wrote in 1536.

⁷⁸ ἡ μονήρης καὶ διήρης καὶ ἐφεξῆς κατὰ τοὺς στιχους τοὺς κατὰ τὸ ὕψος ἐπ᾿ ἀλλήλοις Should anyone think I am unfair to the scholiasts, I would now refer to the chapters entitled ‘The explaining of obsolete words’ and ‘The explaining of matters of fact’ in Dr. Rutherford's recent volume A Chapter in the History of Annotation.

⁷⁹ Validaeque triremes Quasque quater surgens exstructi remigis ordo

Commovet,…. Celsior at cunctis Brati praetoria puppis Verberibus senis agitur molemque profundo

Invehit et summis longe petit aequora remis.

⁸⁰ σὺ ταῦτα φωνεῖς νερτέρᾳ προσήμενος κώπῃ, κρατούτων τῶν ἐπὶ ζυγῷ δορὸς

We have here a reference to a ship in which the zugite was the most important person, and so not a trireme; and as it is too early for a bireme, it bears out section B; it was a μονήρης δίκροτος. The importance of the zugite here came from his being the stern oar; see n. 30. Is it not possible however that the contrast is between oarsmen and fighting men, with a play upon δορός ?

⁸¹ προσπαρδεῖν ἐς τὸ στόμα τῷ θαλάμακι. Anyone inclined to take this literally should read Julien de la Gravière's remarks in La Marine des Anciens.

⁸² Quam sexdecim versus remorum agebant. A translation of ἑκκαιδεήρης in the corresponding passage in Polybius.

⁸³ This may help to explain Lucan's ‘senis verberibus’ (n. 79) which refers to one hexeres only, and should on the accepted theory be sex verberibus, if it were to refer to the beat of the six banks. It means ‘with sixfold strokes’ ‘strokes worked (or made) by six (men) apiece.’ Lucan's quadriremes have not four ordines, but a fourfold ordo.

⁸⁴ Thuc. 7, 40, 5 the Syracusans in boats ἐς τοὺς ταρσοὺς ὑποπίπτοντες τῶν πολεμίων νεῶν Schol. ὑποδυόμενοι ὑπὸ τοὺς ταρσούς. If the schol. be right, as Bauer supposed, the accepted theory is in a bad way, of course. But the schol, must be wrong. The same phrase in Dio Cass. 50, 32, 8 clearly means driving the ship across the oars so as to break them; he adds καὶ τὰς κώπας συναράσσοντες; and warships could not go under. Cf. Polyb. 16, 4, 10 ἐμπιπτόντων ἐς τοὺς ταρσούς.

⁸⁵ See Bauer 367 n. 1 on the so-called Malay bireme. See also two startling sections of triremes in Kopecky, , Die Attischen Trieren (1890)Google Scholar plates 21 and 22, which he calls ‘sehr beachtenswerte Abbildungen alter Schiffe,’ from Rondelet. ‘Die erste (fig. 21) ist der Abdruck einer Medaille’ etc. On turning up Rondelet (1820) I found, of course, they were Rondelet's own sections, the most worthless of guesswork; of fig. 21 Rondelet does not even pretend to figure, or refer to, any original, but merely labels it ‘after a medal.’

⁸⁶ The two triremes in the Naples Museum, figs. 1676 and 1691 in Baumeister, the first from Pompeii and the other from Puteoli, are, I think, of no great value, as the top oars could hardly reach the water; but the way the oars are laid in threes, one actually upon the other, can be meant for nothing but three oars to a bench all issuing in a sheaf from one opening. The spirited Isis-temple ships, the only ones that give any idea of the general look of an ancient warship, are of no value for the ‘problem.’ I have not seen any representation of the Ulubad, ‘bireme’; but according to B.C.H. 12, 190Google Scholar the oars (14 in number) are in groups of two, side by side. If not a moneres, it would seem to add little to what can be learnt from the Palazzo Spada ship. Two recent discoveries, the ship on a metope of the Treasury of the Sicyonians at Delphi (see Assmann, in Jahrb. 1905, p. 32)Google Scholar, and a graffito on the wall of a tomb near Anfushi bay in Egypt, to which Mr. G. F. Hill kindly referred me (DrBotti, G. in Bull, de la Soc. Archéol. ďAlexandrie (1902) p. 13Google Scholar seq. and Admiral Blomfield ib. p. 37), do not bear on the problem of the oars; though the latter ship (called late Ptolemaic) is interesting as showing a further development of the navis ignifera used by the Rhodians in 190 B.C.

⁸⁷ Every oarsman will sympathise with Arenhold, , Die historische Entwicklung der Schiffstypen (1891)Google Scholar, when he says bluntly that every monument on which the oars ‘ganz steil in's Wasser tauchen’ is self-condemned. I would like to say the same of every similar reconstruction, and of every monument which shews an oarsman grasping the oar from underneath and with no possibility of getting his feet against anything. Mr.Holmes, G. C. V., Ancient and Modern Ships (1900)Google Scholar, suggests that the monuments shew that the art of rowing was not understood till the Liburnian came in. But some mediaeval pictures also shew the oars at an absurd angle; e.g., C. A. Levi, Navi Venete, pls. 28 and 31; and it seems incredible that any people should row for centuries without discovering the proper angle for the oar to make with the water.

⁸⁸ Graser published it (Arch. Zeit. 1874 vol. 32, p. 71). It is now in the British Museum (Dept. of Gr. and Rom. Antiq.) It is certainly not a drawing of the Lenormaut relief.

⁸⁹ Two slits in the παρεξειρεσία, on which Eins, , Das Rudern bei den Alten (1896)Google Scholar, has based what appears to be an attractive theory of the διέκπλους. I have not seen his book. Torr follows Graser in saying the holes are for ropes for an anchor: but if so they should be further forward. If they are not portholes, this hepteres had seven men to an oar, as the monument shews that no oars could be rowed anywhere except through, or resting on, the παρεξειρεσία.

⁹⁰ With a possible reservation in favour of one or more hexereis in Sicily, n. 51.

⁹¹ Polybius has been so abused for saying that the Romans had no experience of building quinqueremes and required a Carthaginian model, that I feel the utmost diffidence in suggesting that the basis of the story is merely that the Carthaginians had got the new system and the Romans had not. Polybius does not say that they copied a stranded quinqnereme; he says (1, 20, 15) that they built their whole fleet (i.e. quinqueremes and triremes) on the model of a stranded cataphract. Ihne's criticism (Röm. Gesch. 2, 49), that they had Syracusan models to hand, is beside the point. We, for instance, had many English models to hand in the Napoleonic war; yet I have read that we often copied the lines of French prizes.

⁹² According to Diod. 19, 62, three ennereis and ten dekereis were built. This may be an anticipation; anyhow, they did not go into action. Plutarch gives no details of size. Beloch, , Gr. Gesch. iii. 1Google Scholar, 159 n. 1, defends Diodorus' account, as against Niese, and says it is the best picture of a sea-fight of the time that we possess. This seems to overlook the battle of Chios just a century later.

⁹³ Lysimachus' great okteres, the λεοντοφόρος, is said to have distinguished itself in the sea-fight between Ptolemy Keraunos and Antigonus Gonatas; Memnon 13=F. H. G. 3, 534, τὸ ἐξαίρετον ἔφερεν. The change of system obviously came in before this ship was built, whatever Memnon's description exactly means.

⁹⁴ Livy 24, 40 (nuntiantes, Philippum primum Apolloniam tentasse, lembis biremibus centum viginti flumine adverso subvectum) is an apparent instance to the contrary. But we know all about these lembi, which Philip had built on the Illyrian model (Polyb. 5, 109) and which fought so well at the battle of Chios; and they were certainly not biremes (n. 60). The explanation is flumine adverso; they were going up stream and to get more power the oars had been double-banked for the occasion. Double-bank, ‘to provide…with two rowers for each oar’; see Murray's Diet. s.v. I wish to thank my friend Mr. Colin Campbell for calling my attention to this word, which he tells me is still in use, and which aptly explains this puzzling passage. As to Pliny 7, 56, see Appendix.

⁹⁵ See n. 40.

⁹⁶ Pernice, , Geometrische Vase mit Schiffdarstellung (Jahrb. 1900, p. 92)Google Scholar, on the ship published by the late Dr.Murray, A. S., J.H.S. 1899 (vol. 19), 198Google Scholar. I gather that in 1900 Pernice no longer held the view he had taken in 1892 as to the fragments of dipylon-ships, Figs. 5 and 6 in his article in Ath. Mitth. 17. Assmann claims to have refuted Pernice, , (Arch. Anz. 1901, p. 98)Google Scholar; and his point, that the Dipylon chariots shew one horse beyond and not over the other, is a fair one. But he does not (apparently) deal with the three things that seem conclusive, viz.:—(1) the supposed upper deck has no supports; (2) the supposed upper oars are cut off short on reaching the (supposed lower) deck, i.e., fall on the other side of it; and (3) the steersman is lower than the supposed upper rowers.

⁹⁷ It is well known that almost all beginners will try to draw, not what they see, but what they know to be there. A case exactly in point appears to me to be the idea of some savages, that a drawing in profile represents half a man only. This would meet Assmann's point about the chariots. It is easy to shew the further horse beyond the other, but very difficult thus to shew the further oarsmen.

⁹⁸ Schreiber, die Hellenistische Reliefbilder, Pls. 10 and 23 respectively. See the two together in Dar.-Sagi. s. v. navis.

⁹⁹ For such a hepteres at Venice, Fincati, p. 196. It does not however appear if the oars in the Venetian ship were at unequal levels.

¹⁰⁰ Though I do not accept Bauer's hypothesis of the larger polyereis, I thoroughly agree with his conclusion; ‘[Meiner Hypothese zufolge] ist es unmöglich, den Typus eines Schiffes nach der Zahl der auf einer Darstellung sichtbaren Ruderreihen zu bestimmen’ p. 463).

¹⁰¹ Nothing larger than a dekeres is known to have gone into action; nor does it appear that in mediaeval times more than ten men to an oar were ever known. It is possible that the performance of Demetrius' hekkai-dekeres, which so pleased Plutarch's, authority, (Dem. 43Google Scholarτὸ τάχος καὶ τὸ ἔργον ἀξιοθεατότερον τοῦ μεγέθους), was only a ‘contractor's trial’ with a picked crew and very favourable conditions. Yet Philadelphos' extraordinary fleet (Ath. 203d) cannot have been merely for shew; though the account may he exaggerated, as Beloch supposes. Livy's translation of ἑκκαιδεκήρης (n. 82) seems to dispose of the otherwise attractive view that the higher terms were arbitrary and merely denoted so much extra tonnage.

^101a One of Weber's points is the single line in the water.

¹⁰² Even as late as 1896 Eins is said to have taken Y, and Haack (whose paper I have also not seen) Y and Z, across EE. Since this went to press, I see that the older view is still taken by Torr in Dar.-Sagl, and by MrConybeare, E., Triremes, 1904Google Scholar.

¹⁰³ ‘No sign of Y and Z crossing over the transverse pieces. The surface is much weathered and perished, and they may have done so in very low relief, now lost—or even in paint; no doubt the thing was made far more intelligible by the colouring (of which naturally no trace remains, but it must have been there). AAA are rounded knobs projecting vertically above the transverse strip E, but with their faces in the same plane as the face of E.’…‘I think Torr's drawing (which I have examined since looking at the stone) exaggerates the disturbed surface of the water; there is a raised lump where X meets the water in the case of oarsmen 1, 3, 4. No such lump in the case of Y and Z; but this must not be pressed.’ All these points come out clearly on a cast in the Inner Temple Library, which also shews another point referred to by Mr. Carr Bosanquet, and not appearing in Fig. 3, viz., that X seems to pass over F in the case of oarsmen 3, 6, and 8, as well as 1. The raised lump in the water round X, as compared with the smoothness where Y and Z meet it, is most distinct in this cast. The figure in Baumeister, reproduced by Luebeck, is from a cast in Berlin, but is (admittedly) much touched up and ‘completed.’

¹⁰⁴ Even Mr. Tori's storehouse of quotations fails here. Herod. 5, 33 (which I shall come to presently) is certainly not such evidence. Pollux' τρήματα is quite satisfied by openings in the παρεξειρεσία; and none other appear on the Praenestine and Palazzo Spada ships, and perhaps I may add on the prow of Samothrace. (The portholes are however low on the Delphi ship, which is a moneres; but the gunwale is low also). θαλαμία is not connected with thalamite, technically, and does not mean the thalamite ports, but any port (Ar., Ach. 553Google Scholar) or any opening (Ar., Peace, 1232Google Scholar).

¹⁰⁵ Torr, p. 45 who takes AAA as the thalamite portholes, about one foot above the water, but points out the difficulty of squeezing in the rowers.

¹⁰⁶ See section D (d), and n. 75.

¹⁰⁷ Polyaen. 5, 22, 2. Note that the oars were not merely drawn inboard, but taken right out. The same manœuvre in Polyaen. 1, 47, 1; 3, 11, 3; excerpta Polyaeni 57, 9. This is obviously dead against the portholes being covered with leather bags, the only alleged support for which is the Praenestine ship. There is no proof that the Athenian ἀσκώματα were such; the only passage is Zonaras, who shews his ignorance by saying that the ἀσκώματα were fastened to the oars, the Athenian lists shewing that they were fastened to the trireme. Pollux 1, 88 τὸ πρὸς αὐτῷ τῷ σκαλμῷ δέρμα ἄσκωμα is more likely to be correct. But I suppose that the notion that the most intelligent people in the world first ‘honeycombed’ the sides of their triremes with holes larger than a man's head, and then covered the holes with leather bags to keep out the water, will die very hard. Why some of the text book writers believe that the oars were put out from the inside, blades first, instead of having the handles passed in from outside, is to me a puzzle. It also seems to me to be a grave question whether oars could be rowed at all through the sides of a boat as light as a trireme without pulling her to pieces in a short time.

¹⁰⁸ See n. 11.

¹⁰⁹ I do not mean more than ‘at the same time.’ We cannot for instance prove that the arrangement of the fourth century was that of the first. See however under F. Some writers assume a new arrangement of oars to explain each monument.

¹¹⁰ I am bound to refer to this controversy, on which so much has been written, in Germany, and which has produced the greatest gem of the whole trireme-literature, the theory that the ‘thalamites’ may have taken 4 strokes and the ‘zugites’ 2 to the ‘thranites’ 1, because a pianist can play in three-time with one hand and four-time with the other. Given more than a certain proportionate difference in length, it is matter of mathematical demonstration, as well as practical knowledge, that the oars cannot be rowed together by one man to each oar so that each oar should do its best and each man pull his weight, i.e., his own and his share o f the ship's; and therefore each added bank after the first means a relative loss in power, owing to the disproportionate increase in dead weight. Schmidt here almost takes up the position that, if practical oarsmanship forbids his deductions, so much the worse for practical oarsmanship: the ‘thranites’ had ‘erheblich längere Riemen.…Um diesen Schluss kommen wir nun einmal nicht herum, wir mögen uns drehen und wenden, wie wir wollen. Die namhaft verschieden langen Riemen, also auch alle ihre Konseqnenzen, sind feststehende Thatsache’ (p. 17; italics mine). Once more, whatever thranite means, there is no evidence of any sort that the thranite oars were much longer than the others.

¹¹¹ Bauer remained of the same opinion after examining the original; see his review of Schmidt, in Neue Phil. Rundschau, 1900 p. 301Google Scholar.

¹¹² Fincati seems clear that no zenzile galleys larger than triremes were in use at Venice; but it is generally asserted, on Pantera's authority, that quinqueremes a zenzile were used. In Pantera's time the zenzile galley was only a memory. A thing might however be feasible with the shorter Athenian oars that was not so with the Venetian. How many difficulties would be avoided if one could only agree with Beloch, (Gr. Gesch. 2, 470)Google Scholar that the Athenian lists do not really prove that the oars of a trireme were used for a quadrireme.

¹¹³ καθάπερ οἶμαι κἂν ταῖς τριήρεσι τὰ πέρατα τῶν κωπῶν εἰς ἴσον ἐξικνεῖται, καίτοι γ᾿ οὐκ ἴσων ἁπάντων ὄντων· καὶ γὰρ οὖν κἀκεῖ τὰς μέσας μεγίστας ἀπεργάζονται διὰ τὴν αὐτὴν αἰτίαν. ―πέρατα cannot of course refer to the handles, which did not, and could not, come εἰς ἴσον on any conceivable theory, except Graser's.

¹¹⁴ This (the word αἰτία) is conclusive against μέσαι here meaning amidships, whatever theory we adopt as to the trireme; for the oars amidships would not have to be the longest to make the ends come level; indeed if they were the longest the ends would not come level. It seems equally conclusive against Conybeare's view that μέσαι means the middle of three superposed banks.

¹¹⁵ The explanation is substantially Fincati's, though he does not apply it to Galen. He says they had two zenzile triremes at Venice, in one of which the oars formed one even line in the water. To the same effect is Aristot., de part. anim. 4Google Scholar, 10—the handle of the κώπη μεσόνεως traverses a greater space.

¹¹⁶ The chapter is too long to cite in a note.

¹¹⁷ The argument under B, C, D, and E is independent of the meaning of ‘zugite.’

¹¹⁸ This was the object of the telaro in the mediaeval galley, and of the first importance, as Jurien de la Gravière points out. It gave the boat, seen from above, the look of a parallelogram with two projecting ends; see the frontispiece and pl. 7 in Furtenbach, also the rearmost trireme in Fig. 1, ante. If I am right, then the παρεξειρεσία itself, though possibly inclining (as from stern to bow) somewhat toward the long axis of the ship, must have been somewhat broader at the bow end than amidships; and this agrees well with Thuc. 4, 12, where Brasidas falls wounded and swooning on to the παρεξειρεσία and does not roll off.

^118a I take μεσόνεοι to be a technical term; something like vogue-avante.

^118b Even Assmann, now doubts it; Jarhb. 1905, p. 89Google Scholar.

¹¹⁹ Fincati could at least claim that his boat would go: according to a writer in the Academy, 1883 p. 219, it attained the great speed of 9 miles an hour, i.e., nearly three-quarters of the pace of an average University crew from Putney to Mortlake. Unfortunately I have never seen any details of what the boat exactly was.

¹²⁰ So far as we have gone, there has been nothing to lead one to distinguish the Roman trireme from the Greek. It is however just possible that in Polyb. 1, 20, 15, we have a reference to a trireme a scaloccio; the Romans, he says, built their whole fleet (quinqueremes and triremes) to a Carthaginian model; and if, as suggested in this paper, the quinqueremes had 5 men to an oar, these Roman triremes may have had 3. This would only accord still further with what happened at Venice, where triremes on both systems are said to have been built. But even were this so, the scaloccio trireme (if I am right as to Galen's meaning) was not the one that survived in the Aegean. At Venice, the galleys a scaloccio killed the trireme a zenzile.

¹²¹ The length of the oars, for instance. It might be attractive guesswork that the bench rose a little from the ship's side inboard and that the oars had separate portholes very close together; this would much resemble Bauer's theory, I think, and might be a useful subject for experiment. It has been suggested by Mr. Cook, whose citation of the τρίσκαλμοτνᾶες of Aesch., Pers., 679Google Scholar, for the zenzile trireme is most happy, as a reference to Fig. 2 (ante will shew.