¹ Especially Cartault, , Monuments Grecs, XI–XIII, 1882–1884, 33 ff.Google Scholar; Kroker, , JdI, I, 1886, 95 ff.Google Scholar; Assmann, , JdI, I, 1886, 315 ff.Google Scholar; AA, 1901, 98; Denkmäler 1597; Pernice, , AM, XVII, 1892, 285 ff.Google Scholar; JdI, XV, 1900, 92 f.; Torr, , RA, XXV, 1894, 14 ff.Google Scholar; Tarn, , JHS, XXV, 1905, 208 f.Google Scholar; Behn, Ebert's Reallexicon, art. “Schiff.”

The fullest modern treatment is by Köster, , Dos Antike Seewesen, Berlin 1923, 84 ff.Google Scholar; this is unfortunately damaged by lack of references, especially in the case of the illustrations. Limited aspects of the question are to be found, among recent publications, in Marinatos, , BCH, LVII, 1933, 218–21Google Scholar; Richter, , Bulletin of the Metropolitan Museum of Art, 1934, 169 ff.CrossRefGoogle Scholar; Chamoux, , RA, 1945, 55 ff.Google Scholar; Carpenter, Rhys, AJA, LII, 1948, 1–10.CrossRefGoogle Scholar Miltner's article ‘Seewesen,’ in RE, Suppl., V, relies too blindly on Assmann and Köster, and entirely neglects much of the available evidence on Geometric ships; §11 contains, however, some useful details of later developments.

² As Torr observed, p. 14. Add to his examples that Pernice's No. 21 is actually a part of Pernice's No. 2, following Cartault's error: also the confusion mentioned on p. 102 below, n. 18.

³ M. Devambez of the Louvre Museum informs me that more fragments from the Dipylon (presumably still from the Rayet hoard) are to be exhibited shortly; one may hope for more ship-fragments. One small new group (here 19) has now (April 1949) been exhibited. Other fragments have been incorporated in made-up vases, but in some cases in false associations. See n. 18.

⁴ R. S. Young, Hesperia, Suppl., II, passim, esp. pp. 1–6, 68–71, 195–232; Kahane, P., AJA, 1940, 464ff.CrossRefGoogle Scholar; Weinberg, S., Corinth, VII, iGoogle Scholar, and AJA, 1941, 30 ff. Weinberg deals particularly with Corinthian Geometric and ‘Protocorinthian Geometric.’

⁵ E.g., the oinochoe, Athens NM 194, his Pl. XXIII, 2, is assigned to the transitional Strict-Ripe period, i.e., ca. 800; but if the figure-style recalls that of 3 below, vase-shape and the dot-edged snake decoration are more compatible with Late Geometric style.

⁶ By Nottbohm, G. in JdI, 1943, 1 ff.Google Scholar She attributes a number of vases (among which is the bowl in the Brit. Mus., 40 in the list below, which must surely be later) to ‘the Master of the great Dipylon amphora’ (Athens NM 804), to whom she assigns the exclusive use of the so-called oval ornament as a filling-decoration. F. Chamoux, loc. cit., pp. 55 ff., substitutes workshops for individual painters, but follows Nottbohm in choosing particular criteria, especially isolated details of ornament, for the grouping of vases, to the neglect of all others.

⁷ This is the case with (b): I did not see (a).

⁸ All the Geometric examples are clearly warships. Undoubtedly long-distance merchant-vessels, broader in the beam and with high curved stem and stern, were built in the Geometric, as in other, periods; but Geometric artists just did not depict them: see below, pp. 135 f. It is to be noted that the usual Homeric ship is not a true warship, although it is distinguished from the ‘broad merchantman.’

⁹ On the type of Mycenaean ships, see Marinates', S. valuable article ‘La Marine Créto-Mycénienne’ in BCH, LVII, 1933, 170 ff.CrossRefGoogle Scholar

¹⁰ The question of whether this horizontal line above the gunwale actually represents the deck or something else is discussed below, pp. 127 ff. Similarly other terms in this list, such as ‘horns’ to describe the typical appendages to stem and stern, and ‘eye’ to describe the relieved, decorated circle on the bow of many examples, are used without prejudice for certain characteristics of the painted representation; their actual interpretation will be discussed in the next section.

¹¹ So e.g., in 6, 14, 15; and in many of the shiprepresentations on seventh-century fibulae. The birds are probably sea-gulls, if anything specific, although two separate types can be distinguished. Birds of prey, too, occur in Geometric art: see Lorimer, , BSA, XLII (1947), 137.Google Scholar

¹² Nottbohm dates the activities of the painter of Athens 804, ‘the master of the great Dipylon amphora,’ to ca. 760–740. The earlier date may be correct.

¹³ Only in this case, and conceivably in 5, the positions of the figures are compatible with Pernice's theory that the apparent deck is really the far gunwale; see pp.127–9.

¹⁴ The attribution of vases to separate painters is dangerous (especially where, as in most of these shipscenes, only small fragments remain), first because on the same vase the figure-scenes and the purely Geometric ornamentation, including filling-ornament, may have been executed by different artists, specialists in each work, although there is no evidence for such specialisation at this date; and secondly because in the eighth century at any rate the manner of representing ships was changing so rapidly that a single painter might well have modified his technique from vase to vase.

¹⁵ For a remarkably similar pile of dead see another crater fragment, Louvre no. A 519; (Hampe, Frühe griechische Sagenbilder, fig. 22; Nottbohm, loc. cit., fig. 2). Chamoux plausibly assigns the two fragments to a common workshop.

¹⁶ As Torr, p. 14, points out, Cartault did not realise that this piece includes the smaller fragment already published separately in Annali (1872), pl. I, 4; Pernice lists only this smaller fragment (his no. 3).

¹⁷ See p. 129 below.

¹⁸ The main Louvre fragment and the largest of the Athens fragments, showing the mast and most of the sail, were found separately in 1872 and 1871 respectively and published by Hirschfeld in AdI, XLIV, 152; from his account it appears that the two fragments were not seen side by side, although this is not made clear. At any rate separate sketches were made and were combined to show a join in Mon. Ined., IX, Pl. xl, 4—a misleading picture which has been widely reproduced (e.g., by Behn in Ebert's Reallexicon, s.v. ‘Schiff’). The sail was made to terminate, in the sketch, at the right-hand end of the left-hand Athens fragment; thus there seemed to be no inequality between the saillength before and after the mast, and a join seemed possible. Soon after discovery the bow-fragment was removed to Paris, evidently by M. Rayet, where the two other small pieces were joined; and the other mast-and-sail fragment fell into oblivion and was eventually declared lost. Meanwhile Cartault took the opportunity of republishing the Paris fragment, extremely inaccurately; Pernice took the republished form to be a new discovery and listed the part and the whole as his nos. 21 and 2. In Athens, a join was made at some time between the mast-and-sail fragment and three others showing the ship's stern, which had been separately published by Pernice as his no. 19, loc. cit., fig. 2. This important join (which would have revealed the inaccuracy of the drawing Mon. Ined., IX, Pl. xl, 4) was never published, nor was a connexion between the two groups re-observed until recently; Mme Karouzou tiells rae that Dr. E. Kunze also commented on it before the late war. The whole story is typical of the muddle caused by the immediate dispersal of the original Dipylon finds.

The atory has recently been complicated by the mounting of the Louvre group of fragments as the bow of a ship of which 17a is the stern. The gap between bow and stern is in any case too small for a typical sail: but the connexion is impossible, for while there are only six vertical divisions of the sail in 14, there are at least seven and probably eight in the sail in 17a. The horizontal lines of the sail were clearly drawn by the artist continuously, in a single stroke: it is impossible that the number of divisions in the same sail should suddenly change. This seems to be a decisive argument against this latest conjecture.

¹⁹ In the Louvre fragment the two deck-struts do not project above the top edging-line, while in the Athens fragments they do: but in 16 also, according to Pernice's drawing, the projections are irregular in the same ship. (The faulty connexion with 17a does not do away with this apparent discrepancy.) The helmet-plumes of the warriors at stern and bow are different; also the under-deck edging-line in the left-hand Athens fragment seems to be of a lighter paint. But the two members of the composite photograph are only juxtaposed for illustrative purposes, and should not be used as an exact criterion.

²⁰ There is a further vertical line from the lower edge of the sail, between the mast and the right-hand corner, which together with the faulty drawing in Mon. Ined., IX, led Assmann, , JdI, 1886, 315Google Scholar, to suppose that this was the sheet, and that the sail had a lower boom; and that since lower booms are seen only on Egyptian ships the Dipylon type of ship was Egyptian in origin. Part of this misleading theory is reproduced by Pernice.

²¹ The earliest recorded use of white paint in Attica for figure-painting is on the Benaki amphora of the end of the eighth century: see Cook, J. M. in BSA, XXXV, 168Google Scholar; but it is used as simple enhancement of ornament perhaps as early as the middle of the century; it is not uncommon on Cretan and Corinthian Geometric, cf. Payne, , Perachora, I, 54.Google Scholar

²² ‘Cataphract’ is taken as implying the broadening of the deck-struts to provide lateral protection for the rowers. See below pp. 129 and 137.

²³ Köster's pl. 29, also described as Dipylon ware, is not even Geometric, but Corinthian, of the late seventh or early sixth century. See p. 122 below.

²⁴ Cf. a Late Geometrie bowl in Copenhagen, CVA Denmark, II, pl. 72, 4a.Google Scholar

²⁵ AM, XVII, 1892, 300. On the whole Furt-wängler's view seems the more probable, since the two left-hand oars below the hull are at a slightly different angle from the others, and are also directly below the one extremity, suggesting that they are steering-oars. But the object grasped by the seated figure cannot possibly be the handle of either of these two oars; rather it is a spear.

²⁶ It is difficult to see Payne's (PV, pp. 9 f.) ‘…extreme clearness with which the rhythm of the repeated forms is allowed to produce its effect.’ The I figures seem to me to be particularly wooden, and the whole composition to lack life and movement. Contrast 40.

²⁷ There was no consistently followed convention of perspective at any stage of Geometric figure-painting, but different subjects were treated by different methods: thus the fact that on the other side of this same vase the farther horse is shown ahead of, and not above, the nexarer one, does not preclude the painter's using the vertical and not the lateral plane to represent depth in the case of the rowers. In the case of horse-and-chariot scenes the method of perspective used here was frequently employed on Dipylon vases, which, however, for other subjects—e.g., in prothesis scenes, where mourners are sometimes shown over, for behind, the bier, and in scenes where rows of corpses are shown one above the other-prefer to represent depth vertically.

²⁸ To the arguments used by Tarn, , JHS, XXV (1905), 208Google Scholar, n. 96 (the upper beams have no adequate support; the supposed upper oars do not reach the water; the steersman's view is blocked by the upper bank of rowers) against this ship being a real bireme may be added the general criticism that no ship whose silhouette really looked like this could ever have put to sea: it would have overturned at the first wave. Yet the Geometric artist was capable of making straightforward silhouettes fairly accurately, as long as he did not try any fancy tricks with perspective. Incidentally one of Tarn's arguments is not conclusive, for in Group III E, too, struts are never shown for the deck,. because the rowers' heads come where the struts should be.

²⁹ It should be remarked that the strange picture of a ship on the Late Geometric (?) bowl, Athens NM 12221, published by Laurent, , BCH, XXV (1901), 143 ff.CrossRefGoogle Scholar and fig. 2 (ref. in Poulsen, Dipylongräber, p. 100, n. 6) is false, on the testimony of Prof. Beazlev: see Benton, S., BSA, XXXV, 75Google Scholar, n. 6. Further, the clay model of a ship described by Köster as ‘from Rhodes’ (op. cit., fig. 19), and implied by him, p. 88, to be earlier than 500 B.C., is actually of Hellenistic date and uncertain provenience. It is in the library of Corpus Christi College, Cambridge: Torr in Daremberg-Saglio s.v. ‘Navis’ (fig. 5269) described it as Rhodian, also without warrant. The interesting bronze profile of a ship from the Idaean cave (Karo, ARW, 1905, Usermheft, pp. 62 ff. and Pl. 1; AM, 1920, pp. 132 f.; Kunze, Kretische Bronzereliefs, 42), which Karo dated to the eighth century and called Ripe Geometric, is in fact probably later. The shapes of the animals from other fragments of the same stand are not at all Geometric (although this does not in itself argue against an eighth-century date), and the warship resembles those of the Boeotian fibulae or the Artemis Orthia relief rather than Geometric types (although there is a resemblance in general lines to 38): the straight bar fixed at an acute angle to the stern, and projecting below it, apparently a form of decoration, is paralleled, e.g., on the plate-fibula in Thebes, Hampe no. 140, Pl. 6—it is certainly quite different from the horn or regularly curved extension which are the only known Geometric appendages to the stern. The ship has five rowers rowing over the gunwale: the captain, with shield, stands in the helmsman's position; behind him stands a woman. Similarly the penteconter-type ships on the Rhodian plates B.M. A719, 720, illustrated in Torr, loc. cit., figs. 14 and 15, are still unaccountably described as Geometric. These plates cannot be earlier than the late seventh century, on grounds of style as well as ship-type. From the second half of the seventh century comes the representation of the bows of a long-ship, with ram (apparently strengthened by two bands), straight horn, eye, and fore-platform, on a gem now in New York: AJA, 1942, 489, fig. 1.

Of a Corinthian Late Geometric pyxis (AM, 1903, 198, K 65) Johansen, VS, 30, n. 1, remarks: ‘M. Pfuhl montre peut-être trop d'imagination quand il dit que le motif … représente un navire. Le fragment conservé est sûrement inintelligible.’ Mr. Dunbabin suggests that a Geometric sherd from Chios (BSA, XXXV, pl. 35. 20) may be part of a ship; but this again seems doubtful.

³⁰ I am indebted to Mr. James Brock for information on this crater and permission to include it here, and to use his photograph and drawing before the full publication of the vase.

³¹ The Argive manufacture of the crater was proposed by Furtwängler, , BPW, XV (1895), 202Google Scholar, and is accepted by Pfuhl, MuZ, I, 111 and others; but in this case the sigma should be and not An eastern origin has also been suggested, and I understand that this theory has been revived on the strength of the fighting-top with warrior's head, which is best paralleled by the invaders' ships on the Medinet Habu reliefs (cf. Lorimer, , BSA XLII, 124 f.)Google Scholar, if it really is a fighting-top; but the rectangle from which a head protrudes on the fibula Hampe 62a is exactly similar to a rectangle without head on another similar fibula, London 3204 (Cat. Bronzes, 372); in both cases this rectangle lies on the after side of the mast and in the latter case looks very like some sort of decoration rather than a solid structure capable of bearing a man's weight. This may be the correct explanation here and with the Egyptian ships at Medinet Habu: nothing else suggests an eastern origin for the Aristonothos crater. On grounds of ship-type a western origin is far more probable, and Ducati, , Mélanges d'arch. et d'hist. 1911, 33CrossRefGoogle Scholar, may not be so far from the truth in suggesting Cumae; for the down-pointing ram which is a peculiar characteristic of the right-hand vessel also occurs in the ship on the ivory situla, of about the same date, from the Pania tomb near Chiusi (Mon Inst, X, pl. 38a; Randall-Maclver, Villanovans and Early Etruscans, 250); so, too, does the vertical post above the stem and the sharply curved extremity of the stern-piece. There is no deck in the Etruscan vessel, nor is there anything resembling a fighting-top, but the general similarity is striking. It is at least possible that the representation is of an encounter between a Greek warship from one of the Italian colonies and an Etruscan trading-vessel. The light buff clay of which the crater is made is very similar to the clay of much of the early Etruscan ware, which may argue for an Italian source, as may the vase's provenance of Caere. Miss L. H. Jeffery kindly informs me that she believes the lettering of the inscription to be not incompatible with a painter from Magna Graecia itself, and observes that the Euboic abecedaria on the Formello vase from Veii and the bottle from Caere (Roehl, Imag.³, 81, nos. 31 and 32) nevertheless show a shaky four-stroke as on this crater, instead of the usual Euboic

^31a The Karatepe stone relief of a ship, of which a photograph appeared in The Times of February 4, ^x949 shows a type evidently wide-spread in the seventh century, similar to (d) above and to the Chiusi ship, v. n. 31.

³² See Cook, J. M., BSA, XXXV, 206.Google Scholar

³³ Munich 2044. FR, Pl. 42; Pfuhl, MuZ, III, fig. 231; Beazley, ABS, 30, no. 21; Technau, Exekias, pl. 5.

³⁴ The contents of the grave from which this fibula comes range from a Late Geometric-transitional amphora to a Late Protocorinthian diadem (Payne, Necrocorinthia, 224); but the presence of Group III Geometric features in the ship, especially the forward horn, suggest that it is not very far removed from the Geometric period. See also p. 147, n. 64.

³⁵ See n. 29 for the seventh-century (probably) bronze relief of a ship from the Idaean cave, and some other seventh-century ships.

Miss S. Benton has recently made the valuable suggestion that the four gold fibulae in the Elgin Collection at the British Museum (Hampe 88. 1–4, pl. 7) are earlier than the Boeotian type, and may belong to the eighth century. The two smaller ones (Hampe 88. 3 and 88. 4) are, on stylistic grounds, as Hampe stated, earlier than the two larger. No. 88. 3 (here Fig. 8) has on one side a rather careful sketch of a ship with large forward horn, fore- and after-platforms, mast-support, and high, curving stem topped by a box-like protrusion. In essentials similar to other ships on fibulae, this representation however gives the impression of being drawn from life. Typo-logically there is nothing against it belonging to the end of the Geometric period.

³⁶ See also Cohen, Lionel, ‘Evidence for the Ram in the Minoan Period,’ AJA, 1938, 486 ff.CrossRefGoogle Scholar Cohen finds no evidence, and sides with Evans and others in taking the higher extremity of Minoan ships as the stem. His refutation of Marinatos' arguments does not seem to me entirely final, and he is forced to admit in some cases a projection of some sort at the bow. The latest contribution to this subject is Gibson, Charles E.'s ‘The Origin of the Raæ’ Mariners' Mirror 33 (1947), 164 ff.CrossRefGoogle Scholar, which appeared after I had written this article. Gibson agrees with Cohen that the ram is not found in Minoan ships. His suggestion that it originated as a left-over from dug-outs with planking amidships, to form a projection at both stem and stern, so: is perhaps far-fetched, but his explanation of the retention of the ram (p. 165) tallies with that given above.

³⁷ So Miltner, loc. cit., p. 906; Marinatos, loc. cit., p. 215; Cohen (in the case of B, iv), loc. cit., p. 493; Carpenter, loc. cit., 3 f. It is noteworthy that the northern ships on the Medinet Habu reliefs are ram-less; but the ‘Peoples of the Sea,’ like the Achaean assailants of Troy, used merchant-ships rather than warships as their voyage was a long one and they had to carry large quantities of stores.

³⁸ In one case only-40, at the end of the Geometric period-the single protrusion forward has become a definite tooth, or προεμβόλιου

³⁹ Cf. the τροποί in Pollux 1.88: οἱ δὲ περὶ τὴν στεῑραν ὲκατέρωθεν περατεινόμενοι τροποί κτλ and the περιτόαια, Pollux 1. 89: τὰ δὲ περὶ τὴν πρύμναν προύχοντα ξύλα περοτόναια καλεῑται.

⁴⁰ There are no grounds for the common assertion (e.g., Cohen, loc. cit., p. 493, n. 1) that ships were always beached stern-to. It depends mainly on the fact that the Greek ships at Troy were so beached in Homer: but this was because they had to be prepared for a quick get-away, a motive which is present in other stern-to beachings in Homer. When a rapid landing was required, however, ships were rowed straight on to the beach without unnecessary manoeuvres by oar, as in Od. 13, 113–15, when Odysseus is landed by the Phaeacians. There the ship is driven so hard on to the beach that only the stern half is left in the water.

⁴¹ That the ram was composed of more than one beam is suggested by later practice. Most important is Dal Pozzo's sketch (Rumpf, Römische Fragmente, 15) of a lost portion of the Attic trireme relief. In structural details the sketch may be considered accurate: the ram is clearly shown as composed of extensions of the main horizontal timbers. This relief may be dated to the very beginning of the fourth century. See also Koster, fig. 28 and Pl. 40.

⁴² Although if it is only a rail, with no more to support than, perhaps, a canvas side-screen, the struts seem unnecessarily numerous.

⁴³ If it is shown (as I hope to show in this article) that Geometric ships had partial decks along all their length, there can no longer be any occasion for the common suggestion, improbable on other grounds, too, that this phrase means that the triremes only had platforms fore and aft and no longitudinal deck. Clearly, they had some sort of deck for their whole length, but not over their whole width; in any case, not a watertight one: see pp. 138 f.

⁴⁴ The vague description of the construction of this craft makes it impossible to assert categorically that it was a raft, rather than a crude version of a conventional sailing-boat: the latter seems to me the more probable.

⁴⁵ This is confirmed by 6, where one ship has mast and sail, and the other, otherwise almost identical, has not (the missing piece of this ship would not leave room for a sail).

⁴⁶ In the sixth century the bows of long-ships were frequently made to resemble a boar's head, with eye, snout (the ram) and horn. MissLorimer, H. L.'s remark (BSA, XLII, 125)Google Scholar that the same intention is present in some Geometric and even Bronze Age ships is at the least debatable. B, v, shows that horns and ram do not necessarily go together.

⁴⁷ See Miltner, RE, Suppl., V, 916: some later ships had two eyes each side of the bow. M. follows Assmann in accepting the apotropaic explanation. Of the Geometric ‘eye’ he remarks strangely: ‘Auf der Schiffen der Dipylonzeit fehlt das Auge durchweg.’

⁴⁸ ἄφλαστον Il., 15, 717; notice that it is low enough for Hector to grasp from ground level, although I think this point should not be stressed: ἄκρα κόρυμβα Il. 9, 241; the sense demands that this, although a plural, should refer only to the stern. Hesychius equates with ἄφλαστον.

⁴⁹ E.g., J. Charbonneaux in Préhistoire, 1932, 230–46, who holds that the Dipylon ships were copied from Syrian models, by way of Cyprus, Rhodes, or Crete.

⁵⁰ Evans, , PM, II, 240 and 241Google Scholar, n. 1, maintains with some force that the apparent ram is here actually at the stern, and represents a fixed steering-device. Marinatos, loc. cit., pp. 182 ff., opposes this hypothesis, I think rightly.

⁵¹ The Acropolis bireme-sherds, 31 and 32, might conceivably represent two-bank ships; they come right at the end of the Geometric period, and may well be throw-backs of the seventh century.

⁵² E.g., Attic, London B 436; Etruscan, London B 60; cf. Morrison, J. S., Mariners' Mirror, XXVII (1941), 14 ff.CrossRefGoogle Scholar and CQ, XLI, 1947, 124. Tarn, , JHS, 1905, 209Google Scholar and Mariners' Mirror, 1933, 60, denied that a oar-arrangement necessitates a true bireme system, i.e., with two super-imposed banks of rowers. But in these vases rowers are only shown for the near side upper-bank oars (Morrison, , CQ, 1947, 124 n. 3Google Scholar, points out the inaccuracy in Torr pl. 4, 17, and Köster pl. 44); the other oars must belong to other rowers well below gunwale level, if not directly below the other rowers. It is hard to see how they could be using the same thwart, as Tarn suggests.

⁵³ The old interpretation of Aesch. Ag. 1617 f. as referring to a bireme is no more correct than the solution which takes δορός as = tiller. The statement of the unknown Damastes in Pliny NH, VII, 207, that ‘biremem … Erythraeos fecisse’ does not help much, and anyway the following clause shows that Pliny's transmission is unreliable.

⁵⁴ Masts are shown, e.g., on the cup by Exekias cited in n. 33 on p. 122.

⁵⁵ The largest number of rowers to be assumed on Geometric ships is thirty-eight or forty, on 38 and 40. These ships may well be exceptional. The fifty oars of some Homeric ships are quite feasible in that far more heavily constructed type.

⁵⁶ See Tarn, , Mariners' Mirror, 1933, 59 f.Google Scholar

⁵⁷ Thuc. i.49.1: Thuc.'s words make it clear that the employment of these tactics as late as 433 B.C. was exceptional.

⁵⁸ The invention of the water-tight deck may have been very much earlier than this, although if so it was not generally adopted. The Haghia Triada clay model (Marinatos, op. cit., p. 174, no. 23), of the LM II period, has an almost water-tight deck.

⁵⁹ Even if the epithets ἐï̑σαι and ἀμφιέλισσαι are taken to refer not to the profile but to the section of the ship, had there been such a distinctive feature as a ram it would have earned a distinctive and unmistakable epithet of its own. This argument ex silentio has already been used in the case of the eye; it is I think a sound one, because the great variety of epithets for-ships, some of them rather weak, shows that the poet was anxious to give as much variety as possible to the otherwise similar descriptions of sea-passages. Where external details are meticulously noted (e.g., the epithets μιτοπάρηος, κυανόπρωρος) it is not likely that the opportunities of eye or ram would be overlooked.

⁶⁰ In summarising the archaeological evidence he deals somewhat inconclusively, perhaps, with the biremes on black-figure vases, nor does he mention in this connexion the Geometric apparent biremes.

^60a Professor Carpenter in a recent valuable article (AJA, LII, 1948, 1 ff.)Google Scholar, of which the main point is that the Greeks only penetrated the Black Sea when they could build ships which could be rowed fast enough to overcome the current through the Bosporus, holds that Ameinocles' invention was the Penteconter proper. But 38, 39 and 40 are clearly not true penteconters of the type of the François Vase ship (Fig. 10), which is roughly contemporary with the Phocaean voyages.

⁶¹ Our extant fragments cannot have come from much fewer than thirty-five separate vases, of which almost thirty are from the Dipylon area. Some of these have other scenes—land-fighting or lines of chariots or foot-soldiers—as well as ship-scenes, on others bands of decoration; and ekphora and prothesis-scenes can occur in conjunction with any of these scenes on funerary ware.

⁶² While the wild suggestion of A. S. Murray (JHS, 1899) for 40, that it represents the start of a ship-race, part of the funeral games of a nobleman, entirely fails to account for the more numerous fighting scenes.

⁶³ Loc. cit., p. 94.

⁶⁴ pp. 6 f.; his grave-group 2, Thebes, contained three figured fibulae, one with a ship representation, and a very Late Geometric amphora, but also a Late Protocorinthian diadem. Gp. 3, Rhitsona graves 6 & 75, possibly dating back to 700, contained three fibulae, but all unfigured. These are the only attested early grave-groups,

⁶⁵ BSA, XXXV, 206: ‘… these twins are nothing more than the creation of artists faced with the difficulty of filling a space too broad for a single figure and too narrow for two. The representation of a single body and extremities in duplicate would naturally suggest itself to artists already accustomed to paint horses in that convention …’

⁶⁶ Supported by von Salis, Theseus und Ariadne.

⁶⁷ Cf. Il. 24, 671 f., Achilles holds Priam's wrist to allay his fear; II. 18, 594 (cf. Hymn to Apollo, 196), the dancers on the shield of Achilles hold each other's wrists, not the hands as we do: cf. Kunze, Kret. Bronzereliefs, pl. 48, 70 b. That this was the ordinary practice in Greek round dances is by no means certain; thus on a Cypriot Iron Age amphora (BSA, XXXVII, Pl. 8 b); on a fragment from the Amyklaion (ibid., p. 69); and on three Protoattic hydrias (BSA, XXXV, pl. 39, 43, 45), it is the hands, not the wrists, which seem to be grasped: cf. in general Kunze, op. cit., pp. 212 f. Here rather it betokens special affection-the dance may have been a betrothal-dance, for the girls are ἀλφεσίβοιαι and wear garlands.

⁶⁸ Cf. C, d, where the scene is more probably one of arrival. The rings from Mochlos and Tiryns (PM, II; figs. 147b and 142) are irrelevant except perhaps as prototypes of an artistic genre.

⁶⁹ The ship, at this time, must have been by far the most striking, because the largest and most complex form of machine. Architecture was at this stage extremely simple; the chariot, the plough or the potter's wheel were simple organisms compared with the rather highly developed Geometric ship. Ship construction often appears to survive breaks in cultural traditions. Purely domestic scenes were not favoured by Geometric painters, otherwise the loom would have made a complex and satisfying Geometric subject.

⁷⁰ J. M. Cook, Director of the School, gave me the greatest assistance in writing this article as also did T. J. Dunbabin, who has brought much extra material to my notice; my thanks are also due to Mme. Karouzo of the National Museum, Athens, and to M. Devambez of the Louvre Museum; and to Prof. J. D. Beazley and R. M. Cook for reading through the article and making valuable suggestions.