consequence most of the remains of the former inhabitants are found there also, though they are by no means confined to that side. These remains are quite abundant in the vicinity of Verde, and less so between that point and the mouth of the river. The causes which have induced American settlement in the large area of bottom land about Verde doubtless also induced the aboriginal settlement of the same region, although, owing to the different systems of agriculture pursued by the two peoples, the American settlements are always made on the bottom lands themselves, while the aboriginal settlements are almost always located on high ground overlooking the bottoms. Perched on the hills overlooking these bottoms, and sometimes located on the lower levels, there was once a number of large and important villages, while in the regions on the south, where the tillable areas are as a rule very much smaller, the settlements were, with one exception, small and generally insignificant. The region treated in these pages is that portion of the valley of Rio Verde comprised between its mouth and Verde, or Beaver creek, on the north. It was entered by the writer from the south; it is not proposed, however, to follow a strict geographic order of treatment, but, on the contrary, so far as practicable, to follow an arrangement by types.

The domiciliary ruins of this region fall easily into three general classes, to which may be added a fourth, comprising irrigating ditches and works, the first class having two subclasses. They are as follows:

Stone villages.

a. Villages on bottom lands.

b. Villages on defensive sites.

Cavate lodges.

Bowlder-marked sites.

Irrigating ditches and works.

The ruins of the first group, or stone villages located on bottom lands without reference to defense, represent in size and in degree of skill attained by the builders the highest type in this region, although they are not so numerous as those of the other groups. They are of the same type as, although sometimes smaller in size than, the great valley pueblos of the regions on the north and south, wherein reliance for defense was placed in massive and well-planned structures and not on natural advantages of location. In the north this class of ruin has been shown to be the last stage in along course of evolution, and there is a suggestion that it occupies the same relation to the other ruins in the Verde region; this question, however, will later be discussed at some length. The best example of this type on the lower Verde is a large ruin, located in a considerable bottom on the eastern side of the river, about a mile above the mouth of Limestone creek. This is said to be the largest ruin on the Verde; it is certainly the largest in the region here treated, and it should be noted that it marks practically the southern limit of the Rio Verde group.

The ruins of the second subclass, or stone villages located on defensive sites, are found throughout the whole of this region, although the type reaches its best development in the northern portion, in the vicinity of Verde. The separation of this type from the preceding one is to a certain extent arbitrary, as the location of a ruin is sometimes determined solely by convenience, and convenience may dictate the selection of a high and defensible site, when the tillable land on which the village depends is of small area, or when it is divided into a number of small and scattered areas; for it was a principle of the ancient village-builders that the parent village should overlook as large an extent as possible of the fields cultivated by its inhabitants. A good illustration of this type of ruin is found a little way northeast of Verde, on the opposite side of the river. Here a cluster of ruins ranging from small groups of domiciles to medium-sized villages is found located on knobs and hills, high up in the foothills and overlooking large areas of the Verde bottom lands. These are illustrated later. Another example, also illustrated later, occurs on the eastern side of the river about 8 miles north of the mouth of Fossil creek. The village, which is very small, occupies the whole summit of a large rock which projects into the stream, and which is connected with the mainland by a natural causeway or dike. This is one of the best sites for defense seen by the writer in an experience of many years.

Cavate lodges are distributed generally over the whole northern portion of the region here treated. At many points throughout this region there are outcrops of a calcareous sandstone, very soft and strongly laminated and therefore easily excavated. This formation often appears in the cliffs and small canyons bordering on the streams, and in it are found the cavate lodges. The best examples are found some 8 miles south of Verde, in a small canyon on the eastern side of the river, and it

is noteworthy that in this case stone villages occur in conjunction with and subordinate to the cavate lodges, while elsewhere within this region and in other regions the cavate lodges are found either alone or in conjunction with and subordinate to stone villages. To this latter type belong a number of cavate lodges on the northern side of Clear creek, about 4 miles above its mouth. The cavate lodges of the Verde differ in some particulars from those found in other regions; they are not excavated in tufa or volcanic ash, nor are the fronts of the chambers generally walled up. Front walls are found here, but they are the exception and not the rule.

Plate XI. Map of the Valley of the Rio Verde.
larger view

Bowlder-marked sites are scattered over the whole region here treated although they are more abundant in the southern part than in the northern. They are so abundant that their locations could not be indicated on the accompanying map (plate XI). These constitute a peculiar type, not found elsewhere in the experience of the writer, and present some points of interest. They vary in size from one room to considerable settlements, but the average size is two or three rooms. They are always located with reference to some area, generally a small one, of tillable land which they overlook, and all the data now available support the inference that they mark the sites of small farming or temporary shelters, occupied only during the farming season and abandoned each winter by the inhabitants, who then return to the main pueblo—a custom prevalent today among the pueblos. These sites are found on the flat bottom lands of the river, on the upper terraces overlooking the bottoms, on points of the foothills, in fact everywhere where there is an area of tillable land large enough to grow a few hills of corn. They often occur in conjunction with irrigating ditches and other horticultural works; sometimes they are located on small hillocks in the beds of streams, locations which must be covered with water during the annual floods; sometimes they are found at the bases of promontories bordering on drainage channels and on the banks of arroyas, where they might be washed away at any time. In short, these sites seem to have been selected without any thought of their permanency.

Irrigating ditches and horticultural works were found in this region, but not in great abundance; perhaps a more careful and detailed examination would reveal a much larger number than are now known. Fine examples