perhaps, the most simple and symmetrical; and, although details of its construction are awanting, its general arrangement seems plain. The plans of the corresponding houses at Housesteads, Gellygaer, and Lyne are not sufficiently complete to enable us to fill up any of the missing details. The buildings in a similar position at Chesters and at Aesica show remains of hypocausts. Otherwise they add little to our knowledge. At Weissenburg on the Raetian Limes the group of administrative buildings is much better preserved than is usual in the smaller forts in Germany.[1] There the Principia, the Storehouse, and a square house with a central courtyard occupy the same relative position as at Newstead, while a number of the rooms on the side of the house farthest from the rampart have been furnished with hypocausts.
Although the general plan of the Newstead house differs from that of the majority of the ordinary Roman houses in Britain, we find, both at Caerwent and at Silchester, one or two private dwellings which closely resemble it. The best examples at Caerwent are the houses No.111[2] and No. VII.[3] In both of these, the rooms open as at Newstead from the outer side of a square corridor which runs round the four sides of an inner court or garden. In house No.111 the corridor roof has been supported on columns, the sides having probably been open. The ambulatory was paved with tesserae of brick. At Silchester the same arrangement reappears in house No. 1 Insula XIV.[4] All of these are buildings of some pretension, exhibiting traces of the refinement characteristic of private dwellings. Nor is this courtyard type of house confined to Britain. We find it, for instance, at Timgad in North Africa in the precincts of the Forum, perhaps as an official residence. In this case the roof of the corridor is supported by light shafts like a moorish patio. And we find it at Mont Beuvray in France, where the building must date from the time of Julius Caesar or Augustus. In this dwelling, with its central court bordered by columns, M. Déchelette recognises the model of the villae urbanae of the end of the Republican period.[5]
The Officers' Quarters
Of the building which must have occupied the space on the north, corresponding to that occupied by the courtyard house on the south, almost no trace could be found. The surface soil was very shallow, and nearly
1 Der Obergermanisch-Raetische Limes, Lief. 72, Kastell Weissenburg.
2 Archaeologia, vol. 57, p. 301, p1. xl.
3 Archaeologia, vol. 58, p. 222, p1. viii.
4 Archaeologia, vol. 55, p. 216, pl. x.
5 Les fouilles de Mont Beuvray, 1897 to 1901, Paris, 1904, p. 45, fig. 4.
63