near Upchurch, and there are three vessels spoilt in the firing (but still usable) in the British Museum, and another, a true waster, in Maidstone Museum.
   Though this evidence is meagre, it is sufficient perhaps to show that pottery was manufactured in the lower Medway region; but, at the same time, it seems doubtful if manufacture there was on a larger scale than in any other well-populated settlement, even allowing for the fact that much of the evidence may be some feet deep in river mud, and at present inaccessible. In the present state of our knowledge, the matter cannot be settled definitely, but there are several points that, in the meantime, are worth consideration.
   It has already been said that the coarse pottery recovered from the marsh-land includes several distinct fabrics, in addition to that popularly known as ‘Upchurch’ ware, and, although many of the types may be dated in the late 1st or early 2nd centuries, there are others that are as late as the 4th century. This assemblage of wares and types is, in fact, just what would be expected from any well-populated region. If no undue emphasis is placed on the discoveries of early antiquaries, the extensive layers of sherds may be regarded not solely as kiln refuse, but as general occupation-levels, varying in depth with the levels of the Roman land-surface. The more recent discoveries of Mr. Peter Gray may probably be regarded as the sites of wattle and daub huts, such as have been found on the banks of the Thames; and Mr. Gray has in addition found traces of more substantial buildings on Slay Hills Saltings. See Top. Index, p. 168. Cf. also for Thameside occupation, Roman London (Hist. Mons. Comm. Inventory), p. 13.
   The fact that some of the pottery shows Continental affinities must not be overlooked. Occasionally a platter of bitumen-coated Belgic ware makes an appearance, but the examples are not numerous enough to postulate a local Belgic industry, and they must be regarded as imports. Certain features in the fumed grey ware—notably the carinated or angular profile—are likewise Belgic traits, but whether they imply an import trade or a local derivative manufacture cannot at present be determined.