other divisions leaving the parent village for other sites, and the ebb and flow continuing until at some period in its history the population of a village sometimes became so reduced that the remainder, as a matter of precaution, or for some trifling reason, abandoned it en masse.  This phase of pueblo life, more prominent in the olden days than at present, but still extant, has not received the prominence it deserves in the study of southwestern remains.  Its effects can be seen in almost every ruin; not all the villages of a group, nor even all the parts of a village, were inhabited at the same time, and estimates of population based on the number of ruins within a given region, and even those based on the size of a given ruin, must be materially revised.  As this subject has been elsewhere[1] discussed, it can be dismissed here with the statement that the Casa Grande group seems to have formed no exception to the general rule, but that its population changed from time to time, and that the extent of the remains is no criterion of the former population.

    [Footnote 1:  See pp. 179-261 of this Report, “Aboriginal Remains in
    Verde Valley.”]

It will be noticed that in some of the mounds, noticeably those in the immediate vicinity of the Casa Grande ruin, the surface is very irregular.  In this instance the irregularity indicates a recent formation of surface; for at this point many walls now marked only by mounds were standing within the historical period.  External contour is of course a product of erosion, yet similarity of contour does not necessarily indicate either equal erosion or equal antiquity.  Surface erosion does not become a prominent factor until after the walls have fallen, and one wall may easily last for a century or two centuries longer than another similarly situated.  The surface erosion of a standing wall of grout, such as these under discussion, is very slight; photographs of the Casa Grande ruin, extending over a period of sixteen years, and made from practically the same point of view, show that the skyline or silhouette remained essentially unchanged during that period, every little knob and projection remaining the same.  It is through sapping or undermining at the ground surface that walls are destroyed.  An inspection of the illustrations accompanying this paper will show what is meant by sapping:  the external walls are cut away at the ground surface to a depth varying from a few inches to nearly 2 feet.  After a rain the ground, and that portion of the walls at present below its surface, retains moisture much longer than the part of the walls which stands clear; the moisture rises by capillary attraction a foot or two above the ground surface, rendering the walls at this level softer than elsewhere, and as this portion is more exposed to the flying sand which the wind sweeps over the ground it is here that erosion attains its maximum.  The wall is gradually cut away at and just above the ground surface until finally the