_a_ is what might be termed a door-sill, a smooth oval stone, evidently from the drift, probably dioritic, at all events a dark-green hornblende rock. In the present instance one was not long enough to fill the gap left between the walls, and two were superposed. I saw no traces of wooden lintels or sills. These doorways appeared to be generally about 0.50 m.--20 in.--above the floor, but if we deduct 0.20 m.--8 in.--for the clay (measure having been taken from the timbers), 0.30 m.--12 in.--will remain as their approximate height over the chambers. The few doors that I could observe are all in the longitudinal walls, and none of them in the transverse; that is, they all open from east to west. But not all the longitudinal partitions have doorways. It cannot, therefore, be admitted that every transverse row was occupied by one family, still less that the family apartments were arranged longitudinally. I rather suspect that this arrangement was vertical, or perhaps vertical and transverse. This surmise is given, however, for what it may be worth. Windows I could not find, although small apertures undoubtedly existed in all the outer walls, both for light and for air. The chambers being all very much ruined, the lower ones filled with the stones and decayed ruins of the superposed stories,--of these stories themselves but part of the walls, denuded and often twisted, remaining,--I have not been able, with one single exception, to secure or even see any of what we would call the "furniture." Small fragments of grinding-stones (_metates_) are sparsely scattered over the entire ruins, otherwise the only object of daily use as articles of furniture met with by me has been a hearth, which I found or dug out _in situ_, in room _I_, and which, complete, forms part of the collections sent by me to Cambridge. The place where this hearth was situated is marked on the diagram in room _I_. It stood on the floor against the north wall, and is composed of three plates of stone, originally ground and polished (as the specimen found in building A will show, which is a fragment only), and, judging from new fragments found, of diorite or other hornblende rock. There are three plates,--a basal one, 40 m.--16 in.--long and 20 m.--8 in.--wide, and two sides, placed vertically east and west of the base,--all three resting against the north wall of the room. Pl. III., Fig. 4, is a diagram of the room, the floor timbers, and the hearth. The basal plate was covered with 0.10 m.--4 in.--of very white ashes, which I have also secured, and the rear of the hearth, which is formed by the original "first coat" of earth daubed over the wall, is thoroughly baked by the heat produced in front of it, as the samples sent will show.[111]