EXCAVATIONS ON ROMAN SITES IN BRITAIN. 407 



wells, and for trial trenches to ascertain the relation of the town well to 

 the rectangular street plan. But the general character of the season's 

 work at Silchester may be gathered from the summary which follows. 



The excavations of 1905 extended over the six months from May 22 

 to November 1 8, under the constant supervision and direction of Mr. Mill 

 Stephenson. 



It was proposed to complete the investigation, which was begun so 

 long ago as 1892, of the portions of Insulm V. and VI., and of the un- 

 numbered insula south of them, which underlie the grass field. So far as 

 Insulce V. and VI. are concerned, this investigation has been completed, 

 but owing to the unexpected depth of soil which was found overlying 

 the buildings, in Insula VI., time did not allow of the ground to the south 

 being even trenched. The excavation of the two insulce bi'ought to light 

 quite a number of interesting buildings. 



In Insula V. the north-west corner was filled by a sti'ucture of regular 

 plan, but uncertain use, the main feature of which was a large pillared 

 hall or workshop, with store-rooms at the end and a corridor or portico in 

 front. In late times the building seems to have been degraded to other 

 purposes. South of, but detached from it, was another structure of 

 similar plan. The chambers behind this, however, were apparently living- 

 rooms, and give a more domestic character to the building. From this 

 second building there extends southwards a somewhat puzzling series of 

 chambers of several dates as far as a large edifice that occupied the south- 

 west corner of the insula. The main features of this were a great hall (?) 

 with long and short corridors facing the streets, and a narrower corridor 

 at the farther end, beyond which, again, was a square building subdivided 

 into small rooms. The nature and object of this extensive group of build- 

 ings is under investigation. The rest of the insula was devoid of build- 

 ings, except along the northern margin, where there were laid open the 

 foundations of a nice little house of the corridor type with some interest- 

 ing remains of mosaic pavements. 



The western margin of Insula VI. is mostly occupied by the major 

 part of an L-shaped building at the north-west corner which was examined 

 in 1892. This has now been fully traced, and proved to consist of, ap- 

 parently, a series of shops covered by a corridor or colonnade along the 

 street fronts. At the south end there has been subsequently built on a 

 second row of chambers, apparently as a series of drying-rooms, though 

 the hypocausts have been destroyed. The northern wing of the block 

 is noteworthy for having been built over an extensive layer of jaw-bones 

 of oxen. 



The remainder of the north side of the insula is almost entirely filled 

 with the foundations of a large mansion of somewhat interesting character. 

 It originally consisted of a fair-sized corridor house, standing north and 

 south, with mosaic floors. To the east of this was afterwards added a 

 courtyard enclosed by corridors, beyond which was built a second house 

 on a somewhat larger scale, with fine mosaic pavements, (fee. A room at 

 the south-east angle is remarkable for the remains of a wooden steeping 

 tank sunk in the floor. In a corridor of one of the main chambers a 

 human skeleton was found, laid in a rudely made grave against the wall. 

 To the east of the house just described was a narrow courtyard with a 

 wide entrance gateway on the north, and shut off from the street on the 

 east by a strong wall. 



On the southern m.argin of the iiisula are the remains of another 



