186 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179 



The timbers were fired, and when they had burned down and the rocks had fallen 

 into the pit, the latter were leveled and covered with green tule. The six sacks 

 of camas were placed on the tule and covered with tule and damp grass, then 

 with a layer of earth, and finally with a carefully laid layer of sod. Over the 

 resulting mound was piled wood, which was in turn covered with green willow 

 branches and leaves to prevent rapid combustion. The top fire was kept going 

 nearly forty-eight hours and then the camas bulbs were removed. The remain- 

 ing pile of burned and cracked stones resembled precisely the burned rock areas 

 described above. [Collier, Hudson, and Ford, 1942, p. 38.] 



MATERIAL CULTURE 



In the earliest trasli that accumulated above the volcanic ash, the 

 stone and bone artifacts were practically identical with those recov- 

 ered at the Hat Creek site. These artifacts included cobble hammer- 

 stones, cobble choppers, flake scrapers, and leaf -shaped projectile 

 points. The early trash in this case is considered to be approximately 

 the first 12 inches of midden that accumulated above the ash. This 

 must be approximate, for there was no even distribution of trash over 

 all the site at all times. 



The closest similarities between the artifacts of Hat Creek and Cold 

 Springs are to be found in projectile points and scrapers. Both were 

 made of like materials, jasper and basalt, and both were flaked in 

 the same manner. Five leaf -shaped points were recovered in the first 

 foot of the postash deposit (pi. 35, &, bottom row) . They varied from 

 41 to 48 nmi in length, which compares favorably with the 39 to 51 

 mm. of the projectile points from Hat Creek. The same tendencies 

 toward planoconvexity and serrated edges were present at Cold 

 Springs as had been noticed at Hat Creek. All in all there is so close 

 a similarity between the points of the two sites that a continuity must 

 be considered. 



Some differences must be noted. At Hat Creek 2 of the 12 classifi- 

 able points had concave bases, a variety that was not found in the early 

 midden at Cold Springs. Furthermore, at Cold Springs, associated 

 with the five projectile points just described, was one side-notched 

 point with a concave base. It will be recalled that none of the points 

 from Hat Creek were notched or barbed. This single specimen might 

 be termed the earliest (in our collection) of a type that became more 

 numerous, and soon replaced the leaf -shaped type. 



Not much can be said of the hammerstones and choppers. These 

 crude tools apparently were used and discarded at will. The most 

 frequently used material was a tough fine-grained basalt. Hammers 

 were either unaltered cobblestones from the river gravel or choppers 

 that had been worn and dulled. The tools classified as choppers were 

 also river cobbles, but had been roughly flaked to a jagged cutting 

 edge. There is little chance that either the hammerstones or the chop- 

 pers had been hafted — their shapes would have made it most diffi- 



