A. M. Snodgrass kindly read my typescript and M. I. Finley and B. B. Shefton drafts of the second and third sections. I am grateful for their comments and especially those of Shefton, who did not agree with me.

¹ See Corinth i 50 n. 1 and RE ix 2258–9: I assume their collections of references are fairly complete.

² Corinth viii 2, no. 1.

³ Verdelis, N. M.Ath. Mitt. lxxi (1956) 51–9Google Scholar and lxxiii (1958) 140–5; PAE 1960, 136–43 and 1962, 48–50.

⁴ J. G. Frazer had previously reported remains of a ‘tramway’ on the east side of the Isthmus (Pausanias's Description of Greece iii 5): they have now, it seems, disappeared.

⁵ Pliny (NH iv 10) and Hesychius (s.v. ‘Diolkos’) seem to say that the diolkos was from Lechaeum to Cenchreae; if so, they were wrong.

⁶ Strabo viii 335, though if this is meant as the direct distance across the Isthmus, the diolkos would have been rather longer.

⁷ Thuc. viii 7–8; Polyb. iv 19.77–9 and v 101.4; Corinth viii 2, no. 1; Dio Cass. li 5.2 Cf. Thuc. iii 15.1 (preparations in 428 B.C.). Though the diolkos is not mentioned, its use on these occasions is assumed generally and reasonably, since it existed earlier and was available later. On the other hand I do not think that the transport of warships across the Isthmus in 883 A.D. (Georgius Phrantzes i 33: in Corp. Script. Hist. Byz. xx, ed. Bekker) is likely to have been on the diolkos, since by then there had been too long a period of anarchy for a public utility of its kind to have remained serviceable (see also n. 8); still less do I believe G. F. Hertzberg's assertion, for which he gives no evidence, that small ships still used the diolkos in the twelfth century A.D. (Gesch. der Byz. 306).

⁸ Strabo viii 335, Pliny (NH iv 10) ‘Lecheae hinc, Cenchreae illinc angustiarum termini, longo et ancipiti navium ambitu quas magnitudo plaustris transvehi prohibet’. Incidentally, use of the diolkos may have ended in 67 B.C.; first, its track is interrupted near its western end by the modern canal, which here was preceded by the cutting for Nero's canal (Gerster, B., BCH viii [1884] 225–32)CrossRefGoogle Scholar and, secondly, a bridge over a 40–50 m cutting would have been impracticable nor was any trace of a diversion of the diolkos observed in the stretches on either side of the interruption, where—unless spoil heaps prevented it—one might expect a diversion to have started.

⁹ See n. 8. Pliny is unambiguous, and conceivably Strabo's πορθμєῑα were a particular kind of ship (cf. perhaps Hdt. vii 25).

¹⁰ The Latin inscription at Corinth (Corinth viii 2, no. 1) even describes the transport of a fleet in 102 B.C. as unprecedented.

¹¹ Thuc. iii 15.1. His ὁλκοὶ τῶν νєῶν must, I suppose, have been slipways, by analogy with Hdt. ii 54, where ὁλκοὶ τῶν νєῶν survived long after a site had been abandoned (cf. also Hdt. ii 159). This implies that the difficulty encountered by the Spartans was one of structure rather than equipment.

¹² Polyb. v 101.4; cf.fr. 162.

¹³ See n. 8.

¹⁴ Cf. Roebuck, C., Hesp. xli (1972) 127Google Scholar: he thinks the purpose commercial and fiscal.

¹⁵ Further, the paving of the track, which is of a softish stone, shows signs of much wear or other deterioration, but—if my memory is right— not of much repairing.

¹⁶ It has been suggested that something more ought to be said about this limit and so I offer a vague surmise. In 412 B.C. it was presumably triremes that were transported on the diolkos (Thuc. viii 7–8). In 217 B.C. hemioliai and undecked ships were transported, but the cataphracts were sent round the Peloponnese (Polyb. v 101.2–4; cf. fr. 162), presumably because they could not be transported; what kind or kinds or warships these cataphracts were is not stated, but one might expect that some were pentereis. Yet, according to some students, triremes and pentereis could be housed in the same sheds (e.g. Morrison, J. S. and Williams, R. T., Greek Oared Ships 286Google Scholar, though on 183 D. J. Blackman is non-committal), and consequently their dimensions must have been much the same: if so, the determining factor here for transport on the diolkos should be weight—and unladen weight, since warships did not carry cargo. By this reasoning the loading limit on the diolkos was between the weight of an empty trireme and that of an empty penteres; and though we do not know what those weights were, we do know approximately the dimensions of triremes—about 35 m long and 5m wide (ibid. 285). Students more knowledgeable about ships than I am might be able to work out what sizes of merchant ships correspond to the trireme and the penteres, allowing first for the trireme being of exceptionally light build and secondly for the lead sheathing of the hull that seems to have been or become usual in merchant ships (de Vries, K. in Bass, G. F., A History of Seafaring 49)Google Scholar. From the data collected by Casson, L. (Ships and Seamanship in the Ancient World 183–90)Google Scholar my very tentative guess is that merchant ships which could carry a load of around 200 tons could have been transported on the diolkos but without their load, and that it would have been a very small ship that was not too heavy when fully laden. This chain of argument is, of course, very tenuous and also assumes that the efficiency of the diolkos remained constant.

¹⁷ So G. F. Bass kindly told me. For a short account of the Kyrenia ship see M. L. Katzev in Bass, op. cit. 50–2.

¹⁸ Vries in Bass, op. cit. 49.

¹⁹ That cargoes had to be unloaded is the opinion also of Verdelis, N. M. (ILN 19 Oct. 1957, 649–51)Google Scholar and of Burn, A. R. (The Warring States of Greece 60)Google Scholar; but Verdelis's routing of cargoes by way of Lechaeum and Cenchreae is unnecessary and Burn's statement that in favourable winds sails were used to help ships up the gradients seems to me unlikely because of the need for stability. I owe these last two references to A. M. Snodgrass.

²⁰ It is, I suppose, possible that the original purpose and use of the diolkos was to transport cargoes and not ships and that that was why the Spartans had to construct ὁλκοί in 428 B.C.

²¹ C. M. Stibbe gives a helpful conspectus in Lakonische Vasenmaler. For Ras el Bassit see Courbin, P.Rev.Arch. 1974 175–7—Google Scholarhis krater is presumably plain ware, but he says nothing diagnostic about the other Laconian; there was also one piece of Etruscan bucchero. Though I know of no Laconian in Chios, there is rare Chiot imitation of Laconian (Lane, E. A., BSA xxxiv 186)Google Scholar.

²² Stibbe (op cit.) gives, on my rough count, nearly 120 attributed pieces to Samos, 15 to Tocra and 11 to Tarentum; Naucratis, where the quantity of pottery found was embarrassingly large, gets 17. The exemplary statistics for Megara Hyblaea estimate for a period of two and a half centuries over 15,000 Corinthian pots, over 2000 Attic, and 180 Laconian—mostly plain ware (Vallet, G. and Villard, F., Mégara Hyblaea ii 9 and 127–9)Google Scholar: incidentally, for Etruscan bucchero the figure is 150, which I think should be augmented by some items assigned to ‘Ionian bucchero’.

²³ Hayes, J. W. in Boardman, J. and Hayes, J., Tocra i 87–95Google Scholar and ii 39–41. C. M. Stibbe has in hand a comprehensive study of Laconian plain ware.

²⁴ Vallet and Villard (op. cit. 127) say that plain kraters predominate at Megara Hyblaea and that this is normal in the West (cf. Dunbabin, T. J., Western Greeks 240—on Sicily)Google Scholar. For Tarentum Stibbe observes that the fine ware was relatively infrequent (Meded xxxvii [1975] 14).

²⁵ Hayes, loc. cit, has some references; and Shefton, B. B. in Dunbabin, T. J., Perachora ii 382–5Google Scholar and 539–40 lists aryballoi and kraters (supplementing Mingazzini, P., Coll. Castellani i 186–8)Google Scholar. For Tarentum (and some other sites) see Stibbe's paper cited in n. 23.

²⁶ Odessus: krater, (BIA Bulg. xxx [1967] 168Google Scholar fig. 16a). Istria: perhaps two kraters (Histria i pls 38. 750 and 89.l.2). Berezan: a few kraters and aryballoi (Kinch, K. F., Vroulia 127Google Scholar; Leningrad, Inst. of Archaeology, Photo-Archive ll, 10060–53043; diary of Skadovsky in Leningrad, State Hermitage Archive 1900.1—dyelo 37). Olbia: hydria (in Leningrad). Cape Tuzla (Taman peninsula): perhaps an aryballos (N. P. Sorokina, Tuzlinsky Nekropol fig. 4.3). Sinope: several aryballoi (JDL lxxiv [1959] 123 n. 1: C. M. Stibbe tells me that these have now been returned to Sinop) cf. also perhaps Boysal, Y., Arch.Anz. 1959 20Google Scholar no. 13. On Odessus, Istria and Olbia I am indebted to B. B. Shefton, for Berezan and Cape Tuzla to J. G. F. Hind. Hind, whose knowledge (often first-hand) of the South Russian finds is exceptionally full, confirms the extreme rarity of Laconian there.

²⁷ For Rasel Bassit seen. 20 and there may bea few sherds from Al Mina (Robertson, C. M.JHS lx [1940] 20Google Scholar, fig. 8 l–0).

²⁸ Jdl lxxiv (1959) 123.

²⁹ T. J. Dunbabin once argued that the export of Laconian to Sicily was by way of Corinth, since the finds in Sicily and Perachora were of similar types (Western Greeks 240); but the similarity no longer appears peculiar, Laconian finds in the Aegean and Africa are unlikely to have passed through Corinth, it is hard to explain why and how Corinth should have procured Laconian wares to market, and the whole process is unnecessarily complicated.

³⁰ Mon. Piot xlviii 1 (R. Joffroy).

³¹ Some students claim the Vix krater for Corinth (but see Jeffery, L. H., LSAG 191–2Google Scholar and 375 on the alphabet). Anyhow, a Corinthian origin would not affect my argument.

³² So, for instance, Joffroy (op. cit. 51–4; L'Oppidum de Vix 142–54 and especially 151) is apparently for Etruria and Boardman, J. (The Greeks Overseas ²213)Google Scholar for Marseilles. It is worth considering how the krater was transported overland, even dismantled: it seems to me that it must have been crated and either put in a cart or slung on two poles to be carried by porters.

³³ It could be argued that the krater was looted by a tribe living nearer the coast and by one means or another passed on to Mont Lassois. If so, the risk of damage would have been still greater and the shortness of time becomes still more troublesome.

³⁴ So, for example, Villard, F., La Cér. gr. de Marseille 141–2Google Scholar.

³⁵ For reasons similar to those given in n. 33 it seems unlikely that the krater was a gift to some nearer tribe which was then passed on.

³⁶ Joffroy, , L'Oppidum de Vix 120–3—Google Scholarabout 25 Greek pots of the sixth century, though these are a very small part of the finds of pottery (ib. 152).

³⁷ Picard, C., Latomus xix (1960) 426Google Scholar n. 1, offhandedly championed an itinerant trader.

³⁸ B. B. Shefton suggested to me that a native of Mont Lassois would not have ordered a volute krater (as the Vix krater is) since that type was then unfamiliar in Gaul; but an itinerant trader could have described it in words and sketches.

³⁹ Even if the krater was made at Mont Lassois by an imported crafts man, the procedure for procuring him must have been equallly indirect, though the risk of course might have been limited to his person.