I must now describe a peculiar characteristic of the domestic economy of a Roman house in Britain, and in the other western and northern provinces of the empire. The Romans did not warm their apartments by fire lighted in them, as was the case in the middle ages, and in modern times, but by hot air circulated in the walls. The floor of the house, formed of a considerable thickness of cement, was laid upon a number of short pillars, formed usually of square Roman tiles placed one upon another, and from two to three feet high. Those of the largest of the hypocausts yet found at Wroxeter were rather more than three feet high. Sometimes these supports were of stone, and in one or two cases in discoveries made in this country, they were round. They were placed near to each other, and in rows, and upon them were lain first larger tiles, and over these a thick mass of cement, which formed the floor, and upon the surface of which the tessellated pavements were set. Sometimes small parallel walls, forming flues instead of rows of columns, supported the floors, of which an example has already been found in the excavations at Wroxeter. Flue-tiles,—that is, square tubes made of baked clay, with a hole on one side, or sometimes on two sides,—were placed against the walls end-ways, one upon another, so as to run up the walls. These arrangements,—which were called hypocausts, from two Greek words, signifying _heat underneath_, and were used in Italy and Greece chiefly for warming baths, are represented in _plate_ IV., _fig._ 7, where AA is the floor of cement, BB the pillars supporting it, and CC the flue-tiles running up the wall of the room. They had an entrance from the outside, somewhat like the mouth of an oven, and fires being lighted here, the hot air was driven inward, and not only filled the space under the floor, but entered the flue-tiles by the holes in the sides, was carried by them up the inside of the wall, and no doubt had some way of escape at the roof. The ashes and soot of the fires have been found in the hypocausts at Uriconium, just as they were left when the city was overthrown and ruined by the barbarians. The ashes are chiefly those of wood, but considerable remains of mineral coal have been discovered. These hypocausts must sometimes have become clogged and out of order, and it would be necessary to cleanse them, as people in aftertimes cleansed chimneys. A sort of alley across the middle of the large hypocaust last-mentioned was probably intended for this purpose. It communicated with another hypocaust adjoining it to the north by a doorway, and this other hypocaust was entered by a rather large archway at the foot of the steps already mentioned. People appear to have been sometimes satisfied with having the hot air merely under the floor, and the flue-tiles were not always used. Comparatively few of them, indeed, have been yet found in the hypocausts of Uriconium. * * * * *