Of these five may be regarded as principal villages and hence large; the others may have been small. The average for all together is 127 persons per village. If we allow 175 persons for each of the larger ones, we must assume 50 for the smaller. These figures seem of the correct order of magnitude. _Willits Valley._--The tribe inhabiting Willits Valley extended from the inland valleys clear to the coast. Stewart makes it clear, however, in contradistinction to Barrett, that they had no permanent villages actually on the coast before they moved in that direction ahead of the American advance to the north. The Northern Pomo thus, unlike the Coast Yuki, lived a long distance inland and traveled to the seashore only as occasion demanded from time to time. Stewart lists 9 village sites: Mitom, the principal village; Tsamonda, a small village; Nabo; Talel, with 8 dwelling pits; Tsaka, with 8 pits; Bakau; Cotsiu; Kacebal; and one of unknown name. He says that there is no evidence that all these were occupied at the same time, but "several must have been occupied simultaneously because five were occupied by the parents of Indians still living." Much light is thrown on the situation in Willits Valley and adjacent areas by the work of Merriam (in the manuscript "Northern Pomo"). Merriam splits the natives into three dialectic subgroups: the But-kow-hah-po-mah of upper Outlet Creek, the Sho-jul-po-ma of eastern Little Lake Valley, and the Met-tum-mah of Willits Valley proper. The But-kow-hah-po-mah had a principal village But-kow-hah-chut-te, corresponding to Stewart's Bakau plus "3-4 rancherias." If we allow 150 for the main village and a possible 25 each for the outlying rancherias, we get 250 for the group. This seems quite reasonable for a small, somewhat isolated subtribe. The Sho-mul-po-mah had for a principal village Sho-tse-yu-chut-te, which is mentioned by Barrett and corresponds to Stewart's Cotsiu. In addition, Merriam cites from Barrett 6 other villages, 4 of which he confirms as villages. One of these, Tah-nah-kum-chut-te, he says contained a sweathouse having a capacity of 200 people. According to a principle enunciated by Powers (p. 168), but which is of somewhat doubtful validity, the capacity of a sweathouse or assembly house is equivalent to one-third of the population. Thus the Sho-mul-po-mah might have had 600 people. However, if we allow that the principal village, by analogy with Kasha, had 175 and that each of the villages of Barrett which were confirmed by Merriam had 4 dwelling houses each (i.e., 56 people) then the population would be computed at 400.