were necessary, as the temperature had to be much higher than that of the tepidarium, with which it was usually directly connected. According to the specification of Vitruvius the laconicum was circular at its base and roofed with a hemispherical dome, in which was an opening which could be closed at will by raising a disc of bronze which hung beneath it.[1] The hot air from the furnace was, no doubt, carried up the walls in tubular bricks. Of the furnace itself we have probably the remains in the two heavy stone walls, four and a half feet wide, which attach themselves externally to this part of building. The space between them (I), three feet wide, was doubtless utilised for the fire. The curve towards the west seems designed to prevent the proximity of the outer wall on the north interfering with the stoking arrangements.

Beyond the tepidarium lay the caldarium (E), a room twenty-three feet long by sixteen feet broad with an apse at either end. Here also we can but conjecture the nature of the internal arrangements. The wider apse on the south (F) had, no doubt, contained the warm bath corresponding in its position to the cold bath in the frigidarium, while the smaller space in the apse at the opposite end of the room (G) would be occupied by the labrum containing cold water to throw over the bather. The narrow chamber or recess (H), twenty-three feet by five feet, immediately adjoining this room and terminating the building on the west, was no doubt designed to hold the water cisterns of copper, three in number—for hot, for tepid, and for cold water respectively. These were usually circular in form as may be seen in the new baths at Pompeii, or at the Saalburg, at Rückingen, and at Feldberg. They were generally placed somewhere close to the caldarium, and beneath them was the principal furnace of the establishment. At Newstead the exact position of the furnace is somewhat doubtful. No trace of it was discovered projecting from the west end, its usual place in the plans of baths, being the situation where it could be best employed to heat the water of the cisterns which was afterwards conveyed in pipes to the caldarium. It is possible that the walls projecting from the south of the caldarium (L) represent a praefurnium. We find one at Silchester similar in plan and occupying the same position relatively to one of the caldaria of the baths. At Newstead, however, these walls did not appear to belong to the earliest period, although they were in part overlying two blocks of heavier masonry which may have been the remains of the earlier

1 Vitruvius, v. 10 (11) 5; also Marquardt, La Vie Privée des Romains, vol. i. 340.