But in 1899 and 1903 further work was done at Naukratis by Hogarth which led him to the following conclusions. Petrie’s Hellenium is wrongly identified: it is not a walled enclosure: what Petrie took for walls is simply débris of houses[551]. The real Hellenium is to be found in what Petrie called the North Temenos[552]. All Petrie’s evidence for a seventh century Naukratis comes from his scarab factory and his “Great Temenos,” both in the South part of the town, which is marked off by the occurrence there of the burnt stratum already referred to, and is shown by the finds to have been the Egyptian quarter of the town[553]. The Greeks would naturally have separate quarters and occupy the Northern seaward end of the town[554]. The scarabs, it is maintained, may well be of Phoenician make[555]. [Sidenote: (iii) arguments from the vase inscriptions shown to be indecisive.] The early arrival of the Greeks in Naukratis has been thought by Ernest Gardner to be confirmed by the numerous inscriptions, some painted but most (about 700) incised, on the pottery from the site[556]. His arguments were criticized by Hirschfeld and Kirchhoff[557] and have received little support[558]. In some of them the lettering appears very crude and primitive; but this may be due simply to the fact that they are scratched by hasty and unskilled hands. They are not more archaic in appearance than some of the graffiti on vases from Rhitsona (Mykalessos) in Boeotia, of which the earliest must be dated in the middle of the sixth century, while others are contemporary with the finely written signatures of Teisias, who flourished at the end of the sixth century[559]. Gardner is certainly wrong in thinking that the lettering of any of his inscriptions proves a seventh century date. But on the other hand, as well remarked by Edgar[560], all that his critics have proved is that none of the inscriptions are _necessarily_ so early. It by no means follows that they are necessarily not. But even supposing that the Naukratite graffiti are all sixth century, it does not follow that Greek Naukratis was of no importance till then. Both Gardner and his critics and likewise Mallet[561] discuss the inscriptions with too little reference to the particular sherds on which they are inscribed. Thirty years ago, when the study of archaic Greek pottery was still in its infancy, this was perhaps inevitable. But in the present state of our knowledge the style of the potsherds would be a natural starting-point for dating the graffiti. Unfortunately the information on this point given by Gardner is inadequate, and the Naukratite finds have been so dispersed, that the task of collating sherds and graffiti must now wait for someone who can devote to it his undivided time and attention[562].