end before the dyke was made. All the objects found in the settlement were Romano-British, and it is doubtful if it continued to be inhabited much after the end of the 4th or early 5th century.'

There is in the Colchester Museum (Joslin Coll. Grave Group No. 75) a small bowl of red ware with stamped ornament of demi-rosettes very similar to fig. 2, pl. II. It was found with two single-handled jugs of a type that has been dated to the end of the 4th or beginning of the 5th century, with the conical bung-shaped foot similar to those from the well (figs. 4-5, pl. IV.).

The fourth vessel of the group is a small bulbous beaker probably of 4th century date. Mr. A. G. Wright, of the Colchester Museum, who has kindly given a description of the vessels states that this beaker resembles type 86, pl. LII., of the Silchester pottery referred to the end of the third century, but that the Colchester beaker is of rather later type on account of its smaller straighter foot. It resembles fig. 6, pl. I. The two jugs are decorated in white slip, the pattern on one being nearly identical with that of fig. 2, pl. III.

A number of fragments of the red coated ware were found in excavations on the Roman site at Pevensey in 1906-8, the vessels found of this ware seem to closely resemble those from the well and are ornamented in the same style, both with stamped rosettes and painting in white slip. Altogether a very large quantity of pottery was found including some of the 4th century "Marne" ware. The coins found ranged from Gallienus (254) to Gratian (375).2

Thus it seems that, while it is not proved on the evidence as yet available, there is at least considerable reason for suspecting that this ware was in use in Britain during the last half of the 4th century; if this conjecture is eventually confirmed our British wares cannot have been imitated from the Continental examples of 5th century date. M. Déchelette speaks of the Continental stamped ware with which the British stamped wares have been compared, as belonging rather to the Merovingian than to the Roman period although some was made in Gaul before the fall of the Empire. After reviewing the evidence in detail M. Déchelette sums up by saying "Il résulte de l'ensemble de ces faits que la poterie de la Gaule à décor estampé peut être attribuée au V siècle."

It is a remarkable fact that all the Continental finds of this particular stamped ware recorded by M. Déchelette, and in comparison with which our British finds seem to have been dated, are in the south far away from any parts near Britain, namely, Bordeaux, Narbonne, the Auvergne, Marseilles, Poitiers, Provence, and at Yverdon (Vaud), in Switzerland. M. Déchelette makes a special point that the area of distribution of "ces curieuses poteries estampées" coincides with that of the Gothic territory, all the finds having been made where Gothic influence was predominant, as

1" Excavations," II., pp. 14. 64, 152, 119, fig. II. : p. 142, fig. 9; p. 149; figs. 7 and 8.

2" Sussex Arch. Coll.", vol. LI. and LII.

at Bordeaux, Poitiers, Toulouse, Narbonne, etc., and that a vain search had been made for it in Frankish or Burgundian regions.'

While the various continental finds all bear a strong resemblance to each other and form apparently a distinctive and characteristic group, no piece showing these distinctive characteristics seems ever to have been found in Britain. While some of the rosette-like ornamentation bears à considerable resemblance to that found in Britain, the group as a whole as compared with the British group as a whole seems very unlike.

The majority of that found on the continent is black, rarely red, and the black glaze is spoken of as having been laid on with a brush, whereas the British examples are nearly all red or reddish, and do not as a rule seem to have been treated with a brush. All, or nearly all, the vessels found in Britain seem to have been bowls of various types while on the continent flat open dishes seem to occur frequently. M. Déchelette states that while the group of pottery is much alike with regard to technique and ornamentation generally, it can nevertheless be divided into two distinct series, one symbolic bearing Christian symbols and made for religious or sepulchral use, the other non-symbolic for domestic use. The ornamentation on both series seems altogether much more elaborate and varied than that on British examples; it includes figures of animals, human heads, arcading, medallions, various leaves, and personal names. With the possible exceptions of a fragment found at Silchester, bearing a rude figure of a cross and dots within a circle, and a cross ornament on a sherd from Ashley Rails, symbolism is apparently unknown on British examples.

Thus both as to date, area of distribution, and detail of form and orna. mentation, it is difficult to see how the British ware could have been derived from this particular group of southern Gaulish pottery described by M. Déchelette, as seems to have been generally assumed. May not the British group rather have been derived independently and directly from imported Samian ware, the forms of the bowls is certainly taken from these models and it seems that the rosette may well have been copied from the same originals. It has been already suggested that the "multiple arch ' design that occurs on some of the pieces from Ashley Rails is derived from the egg and tassel ornament so often seen on Samian bowls (“ Ashley Rails," p. 21). The rosette or star-like ornament also appears on many of these bowls, and even the figure of a cross within a circle is not unknown.

1 M. Déchelette perhaps overlooked a small bowl with rosette-stamped ornament found at Etaples, illustrated by C. Roach Smith in “ Collectanea Antiqua," vol I., pt. IV., fig. I., and quoted by Heywood Sumner in "Ashley Rails," p 19.

46

2 See "Catalogue of Roman Pottery in the British Museum," M1009, 1015, 1031, 1178, 1274, 1351, 1441, and many others; for cross and circle see M2548, fig. 250. For M. Dechelette's account of the continental stamped ware see 'Les Vases Ornés de la Gaule Romaine," vol. II., p. 327. From evidence derived from his excavations at Ashley Rails Mr. Heywood Sumner has already suggested that the dating of the British rosette stamped ware might need revision. Ashley Rails, p. 16.