On the southern portion of this impregnable site and grouped about a point where nearly the whole drainage of the mesa top collects, are found the village remains. The Zuñis stated that the houses were distributed in six groups or clusters, each taking the place of one of the abandoned towns. Mr. Frank H. Cushing [4] was also under the impression that these houses had been built as six distinct clusters of one village, and he has found that at the time of the Pueblo rebellion, but six of the Cibolan villages were occupied. An examination of the plan, however, will at once show that no such definite scheme of arrangement governed the builders. There are but three, or at most four groups that could be defined as distinct clusters, and even in the case of these the disposition is so irregular and their boundaries so ill defined, through the great number of outlying small groups scattered about, that they can hardly be considered distinct. There are really thirty-eight separate buildings (Pl. LX) ranging in size from one of two rooms, near the southern extremity to one of one hundred and three rooms, situated at the southwestern corner of the whole group and close to the western edge of the mesa where the foot trails reach the summit. There is also great diversity in the arrangement of rooms. In some cases the clusters are quite compact, and in others the rooms are distributed in narrow rows. In the large cluster at the northwestern extremity the houses are arranged around a court; with this exception the clusters of rooms are scattered about in an irregular manner, regardless of any defensive arrangement of the buildings. The builders evidently placed the greatest reliance on their impregnable site, and freely adopted such arrangement as convenience dictated. [Footnote 4: See Millstone for April, 1884, Indianapolis, Indiana.]