In this account the author has mentioned a few of the more prominent mushroom rock ruins, confining himself to those which can be observed in a hurried visit to the canyon. It is undoubtedly true, as reported by several cowboys, that the side canyons, difficult of access, concealed many others which a longer visit would bring to light. The characteristics of the ruin crowned pinnacles, or leaning buttresses of rock in Hill Canyon are shown in [plate 13].
CONCLUSIONS
As artifacts were not found in or near the buildings on the Hill Canyon cliffs, and as the ruins show no evidence of former habitation, it is evident that they were not dwellings. Their use and the kinship of the people who built them can be judged only by what is left of their walls and the character of their masonry. As has been pointed out, the most prominent of these ruins are circular rooms or towers, arranged in clusters, for an interpretation of which we may look to similar architectural forms found elsewhere in the Southwest.
Their commanding position suggests that these towers were constructed for lookouts and for defense, but the questions might very pertinently be asked, Why should either of these uses necessitate three or four almost identical buildings grouped together, when one would be sufficient? Why are some of them in places where there is no broad outlook?
The massive character of the walls suggests a fortification, but why if defense were the only explanation of their use would not one large building be preferable to many, especially as it would be more easily constructed. It might be urged that they were granaries; but if so, why were they placed in such a conspicuous situation?
In searching for an explanation for the construction of these buildings, an examination was made of aboriginal towers in the valley of the San Juan and its tributaries, especially the Yellow Jacket Canyon and those tributaries entering it on the northern side. In the Mesa Verde National Park the author has also discovered several towers which are in a comparatively good state of preservation. Some of these are situated on high cliffs, others stand in valleys hidden by dense forests of cedar.
Towers are, roughly speaking, scattered sporadically in numbers over a wide extent of country, bounded on the east by Dolores River and on the south by the Mancos River and the San Juan. They extend as far west as Montezuma Creek, following it up north as far as exploration has gone and occurring as far south as Zuñi. Rarely, if ever, however, do we find towers in the dry, sandy, wastes south of the San Juan, and they are unrepresented in the great ruins of the Chaco Canyon. Although there seemed to be certain minor differences in the construction of towers found at different places in this area of distribution, all are identical in essential features.
The towers of Hill Canyon bear a close likeness to those in the region mentioned, except that their masonry is poorer and their walls are more dilapidated. This can be ascribed in part to the material out of which they are built, for whereas the stone in the southern part of the area is soft and easily worked, that in the Hill Canyon region is hard but can readily be split into slabs which did not require much manipulation to bring them into desired shapes for use. The tall and better built towers of the San Juan ([pl. 14, a]) and its tributaries are sometimes single rooms without connections with other buildings, but are more often surrounded at their bases by rooms not unlike those of pueblo ruins. Thus at Cannon Ball ruin the towers rise from the midst of secular rooms and the same is true of the tower in Cliff Palace and elsewhere. This leads to the supposition that these buildings were constructed for some purpose other than as lookouts: they bear all the outward appearance of sacred rooms called kivas of pueblos and cliff dwellers. If we accept this explanation[20] that the McElmo towers are round kivas, as suggested by Holmes, Morgan, and others we can explain why several are united in a cluster, for it would seem that each room in such a cluster belonged to a family or clan. The use of these towers as here suggested can not, however, be proven until excavations of them are made and the signification of the banquette constantly found annexed to their inner wall is determined.
Several structural remains in Ruin Canyon ([pl. 14, b]), a tributary of the Yellow Jacket, especially those at the head of the South Fork, give a good idea of the relation of the tower to surrounding rooms. Here we find towers constructed of fine, well preserved, masonry rising to almost their original height, but crowded into the midst of rectangular rooms imparting to the whole ruin a compact rectangular form. Several towers in this canyon are without surrounding rooms, others have rectangular, square or D-shaped ground plans, but the author studied none with two or three concentric surrounding walls.
The form of one of the largest ruins in Ruin Canyon situated near the fork of the canyon, closely resembles Far View House, in the Mesa Verde National Park. It has a central tower around which are rooms with straight walls, the intervals between which and the circular wall of the tower having a roughly triangular shape. While there is but one tower in this ruin, its similarity in form and position to the large central kiva of Far View House indicates that towers in the McElmo are practically ceremonial rooms, as has been long suspected.