256 BUKEAU OF AMERICAN ETHNOLOGY [bull. 52 



East of the river the superficial formation is of different character. 

 It is a Hght-gray to buff or pinkish sandy loess, composed of fine 

 rounded sand grains, mica scales, and very fine clay. It is slightly 

 cemented by lime but readily crushes in the hand to a fine dust. 

 Sections 3 m. to 4 m. deep show great uniformity and an absence of 

 horizontal structure. On the contrary there is the incomplete verti- 

 cal structure, characteristic of loess. The formation is thus distinctly 

 eolian in texture and structure. The sand and mica scales are derived 

 from the river bed and the clay from the pink loess of the underlying 

 formation. 



The prevailing winds of the district blow from the north or north- 

 west down and across the river valley, where these materials are 

 deposited and exposed by the stream. The origin of the superficial 

 formation is thus sufficiently obvious. It is an eolian loess, blown 

 out of the river bed. 



The original surface of this superficial loess was a slope or plain, 

 very gently inclined toward the river. It is a surface which implies 

 somewhat uniform and gradual accumulation of the deposit. Dunes 

 are absent. A«*the present time the plain is cut by branching gullies, 

 2 m. to 5 m. deep, which have extended from the river bank back into 

 the loess to a distance of one-half a kilometer. The gullies are of 

 recent origin and are growing. Since the observation of the writer 

 and his colleague was limited to the immediate vicinity of Sotelo, 

 these ma}^ be of local character and occasioned by a special condition, 

 which gave storm rains an opportunity to initiate and develop the 

 single system of branching channels that were observed. 



Restricting the inference to this particular locality, there may be 

 distinguished the earlier episode of loess accumulation, which was 

 general and may be still in action, and the immediate incident of 

 erosion. The latter may have been in progress a score of years or half 

 a century, scarcely more, for in this soft loess the growth of channels 

 is very rapid, even where rains are occasional. 



Any fossils indigenous to the loess must be older than the gullies. 

 This would be true also of any human burials exposed in the side of a 

 gully, such as were found, for no burial would intentionally be made 

 where it would be liable to exposure. Accidental burials might occur, 

 however, in holes in the loess near the temporary head of a growing 

 gully, as is explained in a subsequent paragraph in connection with 

 the finding of a horse's jawbone. The writer would thus distinguish 

 tlu-ee conditions of occurrence, any one of which might apply to any 

 remains found in the loess: (1) Occurrence as a fossil indigenous to 

 the loess and contemporaneous with the stratum in which the remains 

 occur; (2) intentional burial, younger than the formation, but older 

 than the gullies where the bones are exposed by erosion; (3) acci- 

 dental burial, which may coincide with the growth of the gulHes. 



