No CrossRef data available.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 08 May 2015
Since W.K. Pritchett's discussion in 1965 of a Hellenistic site on the sandbar at Pylos there has grown a considerable body of support among scholars for the existence of the sandbar in 425 B.C. Nowadays it links the southern tip of the promontory of Coryphasion, which the Athenians occupied, to the mainland near Gialova, and to its north lies Osmanaga lagoon which it cuts off from the main body of Navarino Bay. The promontory is linked at two points to the mainland, at the sandbar and at the dunes to the north which curve round the bay called Voidokoilia. Thucydides in his account of the Pylos campaign at the beginning of Book Four does not mention a sandbar, and supporters of its existence generally assume that his text can be read so as to accommodate it in his account. This I believe is mistaken, for the plan attributed to the Spartans of besieging the fort only makes sense as described if there was no sandbar to use as a land bridge. The aims of this article, therefore, are first to remind readers that the plain meaning of the text at 4.13.1 strongly implies that there was no sandbar in Thucydides' day (this is the older view, clearly expressed by A.W. Gomme); second to show that Pritchett's influential inference from the Hellenistic site to a substantial fifth-century sandbar is not compelling; third to show that attempts to read the text in such a way that it is compatible with the sandbar's existence fail; and fourth to show that there is as yet no reliable scientific information to prove that in the fifth century the sandbar existed in any way that is relevant to the Pylos campaign. In these circumstances there is no reason to reject the plain meaning of the text. Further, this meaning is consistent with what few indications there are about the condition of the sandbar in the fifth century.
1 See Pritchett, W.K., Studies in Ancient Greek Topography, Part I (Berkeley and Los Angeles 1965) 7–12Google Scholar. Among those following Pritchett are Wilson, J.B., Pylos 425 BC: A Historical and Topographical Study of Thucydides' Account of the Campaign (Warminster 1979) 55Google Scholar; Lewis, D.M., ‘The Archidamian War’, in Lewis, D.M., Boardman, J., Davies, J.K., Ostwald, M. edd., Cambridge Ancient History ed. 2, vol. 5 (Cambridge 1992) 414 n. 119 with the map on p. 415CrossRefGoogle Scholar; Hornblower, S., A Commentary on Thucydides, vol. 2, Books IV-V 24 (Oxford 1996) 159Google Scholar; and Rhodes, P.J., Thucydides History IV, I-V.24 (Warminster 1998) 209Google Scholar. The lagoon area is also shown as dry ground by Talbert, R.J.A. ed., The Barrington Atlas of the Greek and Roman World (Princeton and Oxford 2000), map 58Google Scholar, a view which I believe is mistaken. Maps adequate for the purposes of this article are readily available, e.g. in Wilson; Lewis; Grundy, G. B. ed., Murray's Small Classical Atlas ed. 2 (London 1917), map 13 GGoogle Scholar; Gomme, A.W., A Historical Commentary on Thucydides, vol. 3, Books IV-V 24 (Oxford 1956) 487Google Scholar; and the Penguin Classics translation.
2 I have restricted the relevance of the sandbar to the single, crucial question of bringing up siege engines. However, the availability of an approach over the sandbar for infantry also seems to me to render the limited landing space on Coryphasion of dubious use to the Athenian ships, being as it would be subject to Spartan attack, especially after the warfare restarted on the breakdown of negotiations, when there were many ships to be moored.
3 Gomme (n. 1)483.
4 μάλιστα. at 4.13.1 means that landing was either definitely possible or most feasible at that point; ἄν is a conjecture which does not materially affect interpretation.
5 The description of the passage south of Sphakteria as that ‘towards the other mainland’ (4.8.6) strongly implies that the mainland with which this is contrasted is that to the north of Coryphasion.
6 Cf. Yalouris, N.. ‘Hellenistic cemetery of Gialova Palaeonavarino (Koryphasion)’, Arch-aiologikon Deltion 21 (1966) 164-5Google Scholar, for a brief account of the excavation. Pritchett does not claim to have discovered the site, which had received brief notices in archaeological reports in immediately preceding years. A list of archaeological sites in the lagoon and sandbar area is given in Table 1 on pp. 190-1 of Kraft, J.C., Rapp, G.R. Jr., Aschenbrenner, S.E., ‘Late Holocene Palaeogeomorphic Reconstructions in the Area of the Bay of Navarino: Sandy Pylos’, Journal of Archaeological Science 7 (1980) 187–210CrossRefGoogle Scholar.
7 Cf. Yalouris (n. 6), 165; A.R. (1967–1968) 10–11Google Scholar: ‘Fourth century use of the same area is attested by grave stelai used in the construction of the later graves’. However, since the area is the sand spit all materials will have had to be brought in, and it is at least as likely that ready-made slabs from older cemeteries elsewhere were brought in along with new material as that the slabs were available on the spit. Thus there is no sure indication of fourth century use of the site.
8 100 years is in effect the interval between the Pylos campaign and the traditional start of the Hellenistic period. It is a sufficient time to use to demonstrate the shakiness of Pritchett's inference, but in my view there is probably a larger interval, perhaps of 150 years or more, between 425 B.C. and the start of the cemetery. I draw this conclusion from the confidence with which reference is made to Hellenistic remains in the area by contrast with Late Classical remains; e.g. Pritchett (n.1) 7, on sherds from the area: ‘Most of the identifiable fragments seem to date from the third century B.C.; one or two may be as early as the fourth’; cf. n.7. Subsequently to the excavation on the spit the settlement of Pylos on the northern slope of Coryphasion was excavated by Korres, G. (Ergon 1983, 41-2Google Scholar). This is described as a Hellenistic village whose cemetery is located on the spit. Late Classical or Hellenistic statuettes and terracotta votives were found in the village (A.R. 1983–1984, 41-2Google Scholar). Pylos was perhaps a community with perioikic status: see Shipley, G. in Hansen, M.H. (ed.), The Polis as an Urban Centre and as a Political Community: Symposium August 29-31 1996 (Copenhagen 1997)265Google Scholar.
9 Pritchett (n.1) 8.
10 See Zangger, E., Timson, M.E., Yazvenko, S.B., Kuhnke, F., Knauss, J., ‘The Pylos Regional Archaeological Project: Part II: Landscape Evolution and Site Preservation’, Hesperia 66 (1997) 549-641, on p. 584CrossRefGoogle Scholar.
11 See Zangger et al. (n.10)558.
12 Essays in Greek History (Amsterdam 1994) 145–177Google Scholar.
13 Logistics: Pritchett (n.1) 24. The explanation was similarly criticised by Wilson (n. 1) 55. In general, it is true that belief in the feasibility of a landing at the wall facing the harbour is attributed to the Spartans with their recent attempts on the western coast of Coryphasion in mind, for their failure to land there is the subject of the first part of the same sentence (4.13.1). Thus their thought might be: ‘(We could not land there, but) we definitely can here’. However, if this is taken to leave open the possibility that an approach could be made across the sandbar, then the connection strongly suggested in the text between the landing and a siege operation is left inert. Further, the contrast between the two sites for landing is without a practical point, since there would be no need to use the ships at all at the second site. The contrast is only effective if approach must be made by ship in both cases.
14 Wilson (n.1) 55-7. My criticism of this explanation applies also to that of Strassler, R.B., ‘The harbour at Pylos, 425 B.C.’, JHS 108 (1988) 198–203CrossRefGoogle Scholar, who, in positing the site of the harbour of Pylos as a cove at the south east corner of the headland, accepts the sandbar from Pritchett (198 n.2) and the wall facing it from Wilson (200).
15 There are obscurities and omissions in Thucydides' account of the campaign, but the clearest errors of fact relate to measurements, in particular the size of Sphakteria and the width of the southern entrance into the harbour, with the attendant estimate of the relative widths of the northern and southern entrances. Estimates of distance are notoriously inexact in ancient authors, and it is not clear that here they have any bearing on the reliability of other parts of the narrative. Apart from this special class of errors, it is hard to point to other mistakes which might lead to grave distrust of his view of the terrain. Scholarly opinion is divided on whether or not Thucydides had personal experience of the topography; a recent commentator, Rhodes (n.l) 208, writes: ‘His topographical information (which seems more reliable now than it did to earlier generations of scholars) suggests that he may have been to Pylos; alternatively or additionally he had a very well informed source…’.
16 Loy, W.G. and Wright, H.E. Jr., “The Physical Setting’, in McDonald, W.A. and Rapp, G.R. Jr., The Minnesota Messenia Expedition: Reconstructing a Bronze Age Regional Environment (Minneapolis 1972) 36–46Google Scholar; on p. 46.
17 McDonald and Rapp (n. 16) 240.
18 Kraft et al.(n.6) 199.
19 Kraft et al. (n.6) 199: ‘The archaeological sites of Hellenistic and Roman age … found on the west end and centre of the sand barrier require that it was in its present position by 2300 to 2500 B.P.’
20 Pritchett, however, in his later treatment of the campaign (n.12), is inclined to think that the Spartans would not have attacked from the north in view of the steep terrain, but from across the spit (165). He thus continues to affirm the existence of the sandbar, while implying (176) that his account is consistent with that of Thucydides, but he gives no further consideration to the question of disembarkation in an attack on the wall facing the harbour, and he does not refer to the article of Kraft and his co-workers.
21 Zangger et al. (n.10) 558 with reference to Haralambous, D., ‘Geomorphologische Untersuchung in der Bucht von Navarino, I Teil’, Praktika of the American Academy 34 (1959) 92–96Google Scholar. All three of the major studies thus reject the notion that the lagoon area was dry land in the fifth century B.C.
22 Zangger et al. (n.10) 576. One of their aims was to ‘establish a reliable radiocarbon chronology for the pollen sequence; …. at the end of our study we had obtained one of the most complete sedimentological and palynological Holocene sequences in southern Greece, including an accelerator-dated radiocarbon chronology stretching back to approximately 7400 B.P.’ It is to be noted that more detailed studies of the PRAP results are to be published separately, so that a clearer picture may emerge (554).
23 The PRAP study (Zangger et al. (n.lO) 585) subtracts 560 years before calibration from the radiocarbon age of all shells and bulk-sediment listed in the tables, to compensate for hard water and reservoir effects; though the authors say that all the dates obtained by the subtraction seem plausible, this plausibility is hardly going to allow distinctions of the precision we require for our purposes. Further, the dates nearest to Thucydides are averages within a range of almost 400 years (633, Table 11), and the authors draw attention to the error margins associated with radiocarbon dates (594).
24 Yialova is identical with Gialova, and Palaeonavarino with Coryphasion.
25 Zangger et al. (n. 10) 558-9.
26 Pritchett (n. 12) 146.
27 Pritchett (n.1) 8 notes that several nineteenth-century maps show one or more exits through the sandbar from the lagoon to the bay.