I have been involved for many years with the material of Greek origin found in the late 19^th – early 20^th century excavations of Flinders Petrie, Ernest Gardner and David Hogarth at Naukratis. The site is situated on the western, Canopic branch of the Nile delta, in the district of the new capital city of Sais, for which it functioned as an international port (Fig. 1). Before those excavations it was known principally from the description given by Herodotus (Histories II, 178–9) in the later fifth century BC as a Greek trading post, administered by representatives of several Greek cities, mainly from the East Aegean (Möller 2000). More recently my work has formed part of a larger project based at the British Museum, under the lead of Dr. Alexandra Villing, undertaking not only a full review of those early finds, but also further investigations at the site itself (Naukratis: Greeks In Egypt 2014).

The crux of the retrospective part of the project is that the material was widely scattered to museums, roughly seventy. In fact, all my own places of previous employment, Oxford, University College Dublin and UCL possess such material – Oxford (and thence Dublin) received its share of the finds especially from Hogarth’s later seasons, with only a small portion from Petrie and Gardner’s excavations on behalf of the Egypt Exploration Fund (now Egypt Exploration Society: EES). As Petrie and Gardner were long-serving professors at UCL, they must be the source of the material here – some 150 Greek pots (or rather fragments thereof) in the Classical Archaeology collection and eight in the Petrie Museum (on which more below), though many other artifacts from the site reside there. Work with the material is now drawing to its conclusion. The results have steadily been put online: British Museum (BM) objects going live immediately on the Museum’s Collections Online Catalogue and non-BM pieces, alongside BM ones, in a dedicated Naukratis Online Research Catalogue (Villing et al. 2013–15). The proceedings of two colloquia are in course of publication.

So far, the work of the team has demonstrated some clear conclusions, not previously recognised, by far the most important being the amount of local Egyptian material from the site, giving the lie to the entrenched notion that Naukratis was essentially a Greek foundation. We should see it rather as a town in which non-Egyptian traders were given an enclave and prospered, at least in part living a ‘Greek’ way of life but also engaging closely with their Egyptian neighbours. Geophysical survey has moreover suggested that, rather than being archaeologically ‘exhausted’ following the early large-scale fieldwork, there are still substantial untouched areas to yield stratigraphic sequences (Thomas and Villing 2013), much needed despite Petrie’s pioneering work (Petrie 1886). In the 1970s and 80s an American team, led by Albert Leonard Jr and a Lancastrian, Willy Coulson, did find some undisturbed levels, but their results have needed some significant amendment in respect of both the local topography and pottery chronology (Coulson 1996; Leonard 1997 and 2001).

While noting the new insights into the character of Naukratis, I should also mention a similar re-evaluation of another site excavated by Petrie in the Delta, Tell Dafana, whose finds have been similarly scattered. The re-study and publication of the material by the BM is now achieved (Leclère and Spencer forthcoming), and two points are worth stressing: first that it can no longer be thought a fortress manned by Greek mercenaries, but clearly was an Egyptian town dominated by a huge temple complex. Secondly the character of the Greek material from the site – found mostly within the temple enclosure –, while largely contemporary with the heyday of Naukratis in the sixth century BC, is very different in character, not focused on the drinking vessels that predominate at Naukratis, but consisting mostly of storage jars, plain or decorated (often elaborately) – storage jars used in (and probably dedicated for) the Egyptian cult, the deity not identified.

That takes us to the material from Naukratis, and here I focus on that of Greek origin, and in particular material with epigraphic content, almost wholly ceramic, with which I have been dealing, and which was also a major fascination for the early excavators of the site. Of note is the fact that in most classes of what was regarded in the 1880s as plain pottery, only sherds with inscriptions were retained; e.g. Petrie talks of large quantities of Cypriot storage amphoras - basket-handled jars - but only a dozen sherds, inscribed, were kept. The limited new excavations completed in May 2014, in an area untouched to date, revealed, not unexpectedly, a fuller range of material.

The totality of the inscribed ceramics amounts to some 2,700 graffiti and dipinti, most of them Greek votive dedications that tell us much about the religious life of the site and about the local and visiting Greek traders and travellers that worshipped in the sanctuaries of Naukratis. In addition, perhaps 1,800 stamps are preserved on pottery, almost all on transport amphoras; the city symbols and names (of producers or overseeing magistrates) paint a colourful picture of the wide trade contacts the city maintained also throughout the Hellenistic period. That ‘perhaps’ is needed since their identification is a difficult process, not easily described in brief. In basic terms, Petrie drew 1100 stamps in his excavation journal for 1884–5, as was only recently realised. While most of them entered the collections of the British Museum, none was registered with the rest of the material in 1886, save one piece. However, they remained in the Museum’s collections, amidst numerous stamped amphora handles from other sites. Today, having studied them closely, we are able to assign 700 or so to Petrie’s excavations at Naukratis because of their appearance, most notably because they have been trimmed back, as Petrie mentioned in his diary, in order to fit them in the boxes for transit back to London; reassuringly many of these 700 can be equated with Petrie’s drawings. In addition, there are about 75 known from the EES distribution lists in other museums, Montreal to Warrington.

The stamped amphorae belong very largely to the later period of the life of Naukratis, after the foundation of Alexandria. Illustrated here (Fig. 2) is an earlier stamp, curiously overlooked by both the excavator, Ernest Gardner, and by subsequent scholarship, no doubt partly because it was physically kept separate from other Naukratis material in the BM. It is a unique cartouche stamp on a locally made amphora of Greek type, highly likely to be of the pharaoh Amasis. The other illustration (Fig. 3) is of one of few intact handles that Petrie did not trim, now in the Petrie Museum - from a common enough Rhodian jar, made in the workshop of Aristokrates, of c.220–180 BC.

The sherds with graffiti and dipinti (or rather, drawings of the inscriptions but not the sherds themselves) were in large part published by the excavators, and indeed accurately enough. The spread is from c.600 to c.150 BC, with the majority of the first two centuries. Some 50% had remained unpublished, including virtually all the UCL material. As it is still the largest body of such evidence from any Greek site – Athens as a whole perhaps excepted – it can be used as a basis for statistical analyses more confidently than the sets from most sites. The major drawback is that we have little idea of the precise provenance of much of it, though most must come from the various Greek sanctuaries of the site – of Aphrodite, Apollo, Hera, the Dioskouroi, and the various cults assembled in the large conglomerate sanctuary called the ‘Hellenion’ that was also the administrative headquarters of the Greek port officials – and virtually no knowledge of the totality of sherds excavated. Gardner remarks ‘I have roughly estimated the number of good fragments I have recovered from this layer [in the sanctuary of Aphrodite] at 150,000: this fact will give some notion of its richness’ (Gardner 1886: 181).

What can be said is that the overwhelming majority of the texts are in Ionic script and dialect, a modicum from Aeolic Lesbos and a tiny number only from Doric areas, Knidos, perhaps Rhodes and, further west, Aegina. The pots are dedicated to a range of deities, though mostly to Apollo and Aphrodite, and are overwhelmingly drinking vessels; cups for Apollo and deeper chalices and kantharoi for Aphrodite, the most prominent divine recipients. More curious are the dozen ‘kitchen’ mortaria (grinding bowls) from Apollo’s sanctuary, a limited number of female dedicators (literary references describe their understandable presence at this port-of-trade) and the much debated fact that about half the dedications to Aphrodite were painted on the pots before firing; the debate centres on whether the pieces were ordered and made on Chios, or made locally at Naukratis with clay brought from the island, a scenario now largely discounted. Eight of the twelve pieces in UCL are such pre-firing dipinti; one (Fig. 4) appears to be a dedication by a pair of men whose names appear on similar material from the island of Aegina, while another (Fig. 5) is unusual not only in being a much larger piece, lid or large bowl and providing a previously unattested personal name, Thymogethes, ‘Glad-at-heart’, but is an example of the diaspora of material – one sherd in UCL, the other in the British Museum.

It was the fifth century historian from Halikarnassos, Herodotus, whose description of Naukratis inspired Petrie’s original work. Two sherds bear an owner’s graffito by ‘Herodotus’, but the temptation to equate the name with the historian is to be resisted since the typology of the pair lies on either side of his lifetime. Another famous writer with a Naukratite connection is Sappho, the poetess from Lesbos, whose brother according to Herodotus traded with Naukratis (Histories 1.35). Intriguing that a recently published papyrus text, found no doubt in the more arid areas of Egypt to the south, is part of a poem in which she hopes for his safe return with his cargoes (Burris et al. 2014).