being of a uniform thickness and easily removed, impressed themselves fully upon the exterior surface of the jar, the plain portions being the impress of the smoothed sides Fig. 5. of the mold hole. Of the wicker-marked ware, however, only the prominent projections of the form made an impression, the plain surfaces corresponding to the sandy filling that was resorted to for preventing the soft clay from squeezing into and through the interstices. In some kinds of basketry more filling was necessary than in others, which explains the frequent greater separation and irregularity of the markings. For example, Fig. 5 shows far less of the wicker impressions than Fig. 3, and Fig. 2 gives only the irregular salients of an exceedingly coarse support. It seems probable that the wicker-marked pottery is the most primitive, and an extended study of it might lead to a clearer understanding of the beginnings of pottery-making. The next distinct advance was apparently the use of a fabric as a base, supported by some smooth surface, and then as a further development the coil ware, a process still in use among the Moki, and the simplest and easiest way of modeling a clay vessel without the aid of the wheel; progress in pottery, as in other arts, having been in the direction of simplicity of construction combined with skill in execution.

The probable line of development in pottery-making was then about like this:

1. Made on the outside of a wicker form. Confined chiefly to bowls. 2. Made on the inside of a wicker form. 3. Made on netting in mold hole. 4. Coil-made. 5. Wheel-made, which Indians seem never to have attained.

There was doubtless no sharp line of separation anywhere between these several stages, but they merged into each other as the dawn merges into the day.