Page:The American Indian.djvu/82

54 Almost all of the twine and coil basket weavers are stone boilers, that is, they cook in baskets by dropping hot stones into their contents. Close fine twining and coiling is thus a necessity, for cooking baskets must be water-tight. In the pottery region of the Pueblos and southward, basketry is open and coarser. This undoubtedly accounts for the very high development of the art in California and northward.

Returning to the concentric distribution of the coil and twine technique, one must wonder which is the older. Some of the Déné coiled globular baskets are almost identical with a Chinese style, but this is more likely due to similar materials, for the intervening Siberian styles are more like those of the Eskimo. Thus coil seems to center in California and twine on the coast of the north, thus indicating their most probable centers of dispersion. We must, however, allow for more complex conditions, since archæological remains in certain cliff-houses indicate a high development of coil in prehistoric times. The studies of Kroeber and Barrett, as to the direction of the spiral coil in making the basket, suggest that central California and Arizona are of one type, while southern California (the Shoshoni-speaking tribes) and the interior as well as on northward, are of another. The meaning of this is not quite clear, but can be most readily explained as due to differentiation from two centers of influence. Hence, the chronological relation of coil and twine basketry remains a problem for the future.

The central portion of the bison area marks a hiatus between the basketry of the east and the west. Down the Mississippi and south from the Great Lakes, across the Antilles and on into the manioc or Amazon area of South America, basketry has one common characteristic in that it is made of flat or splint-like materials. Basketry of this sort is also found in the Andean region, but seems not to have been the prevailing style. The material is usually cane, which is probably responsible for the observed distribution. North of the Ohio River and in New England where suitable cane was not to be had, we find wood splints in use. These are made from the easily separated annual layers of certain trees. It seems a