[Footnote 4: The kiva is the assembly chamber, termed estufa in some of the older writings, particularly those of the early Spanish explorers. A full description of these peculiar structures has already been published in an article on Pueblo architecture; Eighth Ann. Rep. Bureau of Ethnology, 1886-’87, Wash., 1891, pp. 1-228.] The smaller courts are but little larger than the largest rooms, but it will be noticed that while some of the rooms are quite large they are always oblong. This requirement was dictated by the length of available roofing timbers. The cottonwood groves on the river bank would provide timber of fair size but of very poor quality, and, aside from this, roofing timbers longer than 15 feet could be obtained only at points many miles distant. In either case the hauling of these timbers to the site of the village would be a work of great labor and considerable difficulty. The width of the rooms was, therefore, limited to about 20 feet, most of them being under 15 feet; but this limitation did not apply to the courts, which, though sometimes surrounded on all sides by buildings, were always open to the sky. [Illustration: Plate XVI. RUIN AT MOUTH OF FOSSIL CREEK.] It is probable that the central and northern portion of the southwestern cluster comprised the first rooms built in this village. This is the portion which commands the best outlook over the bottom, and it is also on the highest ground. Following this the southern cluster was probably built; afterwards the northern cluster was added, and finally the northwestern cluster. Subsequently rooms connecting these clusters and the eastern end of the village were built up, and probably last of all were added the rooms which occupied what was originally the eastern end of the main court. This hypothetic order of building the clusters composing the village is supported by the character of the site and the peculiarities of the ground plan. Most of the rooms in the northwestern cluster and in the eastern part of the village were but one story in height, while the crowding in the interior of the village, direct evidence of which is seen on the ground plan, could take place only after the rooms surrounding that area had been located, and when hostile pressure from outside made it undesirable to extend the bounds of the village; in other words, at the latest stage in the growth of the village.