pS Natural History Bulletin. 



the streams, would form similar deposits, each pond or swamp 

 forming in this way a bed of Loess. The changes in the level 

 of water in the ponds would produce a change in the extent of 

 the muddy flats along their shores, thus facilitating the distri- 

 bution of the terrestrial shells. These ponds and streams, by 

 shifting about through the combined influence of floods and 

 drouths, extended the distribution of the sediment, and subse- 

 quent erosion completed the work necessary to produce the 

 present topography. 



In this way we can best account for the remarkable differ- 

 ences in the level of the base of the Loess deposits, which has 

 always been a stumbling-block to those who maintain that the 

 deposit was formed in large lakes. The ponds, not 

 being at the same level, and each being a center of formation, 

 would naturally produce a combination of deposits, the bases 

 of which would be at different levels. Even near the tops of 

 ridges ponds could bodily lift themselves upward by slowly 

 filling up, each successive deposit during flood-time in rainy 

 seasons bringing both the bottom and the shores higher. 



It is quite probable that a deposition of material is going on 

 in the same manner to-day in the ponds of the Thousand-lakes 

 region of Minnesota and Iowa, and the Sand Hills region of 

 Nebraska. 



Of these conclusions the sixth is the least clearly 

 defined, and is offered here merely as a suggestion. The 

 presence of Loess on the very tops of hills would of course be 

 a serious objection to a portion of it, unless we make a great 

 allowance for erosion. It does seem clear however that all of 

 the Loess in any given region could not have been deposited 

 by the same agency (?'. c. one stream or a single large lake) 

 but that it is rather composite in its origin. 



