the wall did not rise to this height at the edge of the mesa, where it could not have been more than one or two stories high. There is no evidence of the former existence of an inclosed court of any considerable size between the buildings and the cliff, although a passage probably skirted the brink of the precipice, and house ladders may have been placed on that side for ready access to upper rooms. By a series of platforms or terraces, which were in fact the roofs of the houses, one mounted to the upper stories which formed the apex of the pueblo. [Illustration: BUREAU OF AMERICAN ETHNOLOGY SEVENTEENTH ANNUAL REPORT PL. CIX EXCAVATIONS IN THE WESTERN MOUND OF AWATOBI] [Illustration: FIG. 256--Structure of house wall of Awatobi] On the western, northern, and eastern sides the slope is more gradual, and while there are many obscurely marked house plans visible over the surface, even quite near the top of the elevation, they are doubtless the remains of single-story structures. This leads me to suspect that when Awatobi was built it was reared on a mound of soil or sand, and not on the solid rock surface of the mesa. The configuration, then, shows that the pueblo sloped by easy decline to the plain to the north, but rose more abruptly from the south and west. There are low extramural mounds to the north, showing that on this side the dwellings were composed of straggling chambers. The general character of the rooms on the level slope at the western side of old Awatobi is shown in the accompanying illustration (plate CX). The peculiarity of these rooms appears by a comparison with the many-story chambers of the southern declivity of the ruin. Extending the excavations four feet below the surface we encountered a floor which rested on solid earth, and there were no signs of walls beneath it. This was without doubt a single-story house, the roof of which had disappeared. The surrounding surface of the ground is level, but the tops of adjoining walls of rooms may readily be traced near by. The room was rectangular, twice as long as wide, and without passageways into adjoining chambers. The northern, eastern, and western walls were unbroken, and there was nothing peculiar in the floor of these sections; but we found a well-preserved, elevated settle at the southern side, extending two-thirds of the length of the main wall to a small side wall, inclosing a square recess, the object of which is unknown to me. All walls were smoothly plastered, and the floor was paved with flat stones set in adobe. The singular inclosure at the southern corner could not be regarded as a fireplace, for there was no trace of soot upon its walls. I incline to the belief that it may have served as a closet, or possibly as a granary. Its arrangement is not unlike that in certain modern rooms at Walpi.