Official statuary of the Roman period in Britain is now so rare as to be almost wholly unexpected. The few well-known examples have long since been published, and only gleanings remain. But these gleanings are valuable, since they offer a fresh clue to the past existence in official surroundings of important Imperial statues on a wider scale than has hitherto been proved.

Before attempting a conjectural restoration of the first altar-screen it was necessary to try and recapture the plan of the fourteenth-century sanctuary and the positions of the two lateral altars which are known to have existed on the north and south sides of the high altar, also to trace the changes of floor level which have taken place during the intervening centuries.

The well-known bronze figure of a centaur and rider (pl. viii) now in the Royal Albert Memorial Museum at Exeter was found on the beach at the foot of the cliffs (near two geological faults) about 200 yards east of the little river Sid by some fishermen in 1840.

The only other recorded discoveries of Roman objects at Sidmouth consist of a few coins on the sea beach—a ‘first brass’ of Vespasian, a Commodus of A.D. 183, and a small brass of Constantine I (as well as a Bactrian copper)—and others in Sidmouth or the neighbourhood, while pottery of the first, and perhaps second, century A.D., ultimately from the Early Iron Age hill-fort at High Peak, two miles south-west of Sidmouth, has also been noted. The nearest site of any Roman structure is at Seaton, some nine or ten miles to the northeast, where traces of a Roman house have been observed. Until something more substantial has been discovered we can only conclude that the little harbour at the mouth of the river Sid in the Roman period was visited by a few travellers, but not extensively inhabited. It may even be that the bronze object, though of Roman date, was brought by a post-Roman traveller, or decorated a bowsprit when Sidmouth harbour was in use at a later date.

The object comprises a group of a centaur with rider and charging beast set on a pedestal which rests on a shank, now broken; at the top of the shank and immediately below the pedestal is a heavy hook-like bracket. The whole measures 7 in. high (from the top of the centaur's head to the bottom of the broken shank) and 5 in. at its maximum width. It has been cast and is hollow, and is very much worn by the action of the sea, a pebble being still lodged in the mouth and another under the arm of the centaur. The shank also is hollow, about 1½ in. deep, and filled with lead, perhaps in modern times.

In view of the very number—possibly more than three thousand—of English alabaster devotional (as distinct from those made as parts of monuments) carvings which have survived, it is somewhat astonishing how extraordinarily small a proportion of them have hitherto been datable otherwise than on intrinsic grounds. We have, indeed, many references to such carvings in dated or datable documents, but in extremely few cases have we been able to identify the particular objects with which those references are concerned. So far as I know, the only ones which have as yet been connected definitely with a dated document are the ones in the reredos, depicting events in the story of St. James, preserved intact in the cathedral of Santiago de Compostela. Almost invariably an English alabaster carving is datable only within somewhat wide limits, the criteria most usually applied being those laid down ’provisionally’ (but now commonly accepted as practically correct) by Prof. Prior nearly thirty years ago, and in accordance with the inherent characters of the sculpture and the applied coloration. Any discovery of data relating to a particular alabaster carving is, consequently, of considerable moment to us, not merely in relation with the object with which the data are immediately concerned, but much more importantly as giving us means for dating other similar carvings more accurately than within Prior's provisional ‘Periods’ (even though modified by the general terms ‘early’ and ‘late’) of about forty years each. In the case of the Santiago retable there can be no doubt at all as to the direct connexion between it and the existing document relating to it; in the case of the present fragments there is, unfortunately, in the only records I have found associated with them, an element of uncertainty, although there appears to be a very strong presumption that the medieval record to which I shall refer relates to the altar whose retable those fragments adorned.

The following are some identifications that eluded the editors of the 1875 edition. This fine armorial is not widely known to English readers; for with few exceptions its contents are exclusively continental, and in the great majority German, arms. Its compiler was a wealthy amateur armorist of the fifteenth century, of the Patriciate of the city of Constance, in which he discharged several important offices of the municipality. He was also a considerable traveller, and made the pilgrimage to the Holy Land, leaving an illustrated account of it. The armorial, which he completed in 1483, is the work of a joyous and accomplished heraldic draughtsman, with a great facility for design and the fertile invention of damascening patterns. Conrad Grünenberg sets out all his coats as complete achievements, with helmet, crest, and mantling in colour. They number over two thousand, and constitute the fullest and best-executed epitome of German heraldry of his time. But he was omnivorous for coat-armour and scavenged widely throughout the medieval world for his collection, from Scandinavia to Africa, from Ireland to Russia, and beyond to the East and to lands of fable.

The difficulty of accurately dating the vessels known as ‘mortaria’ has often been felt by excavators, and Mr. J. P. Bushe-Fox has done a great service in his classification of mortaria rims, admirably illustrated in his First Report of the Wroxeter excavations in 1912 (Soc.Ant. Lond. 1913, figs. 19 and 20); but he himself stated (p. 76) that ‘one of the chief difficulties has been the lack of any publication in which a sufficient number of specimens has been brought together showing clearly the different types and their variations’. Even his series did not begin earlier than Vespasian; at Margidunum, however, I found examples from the Claudian occupation (A.D. 48) down to the end of the fourth century, and this site therefore has furnished a continuous series of a large number of examples from datable layers, many of them stamped with potters' names, illustrating the changes in development of this vessel during a period of 350 years.