tion of landshell distribution is inoperative except in a broad 

 sense in the study of the aquatic molhisks and. since free dis- 

 persal is much Less dissected while the barriers which remain 

 are comparatively insurmountable, not to be overcome by fluc- 

 tuations of temperature and humidity during periods of years, 

 fauna I areas become of prime importance. 



Much lias been written by writers of more philosophical 

 ability than field experience on the natural means of dispersal 

 of fresh-water Mollusca and several theories such as the (ding- 

 ing of gelatinous masses of gasteropod eggs to the feet of 

 aquatic birds have been advanced and passed so frequently in 

 the literature as to have been accepted as an established fact by 

 one or tAvo recent writers. The similar dispersal of the I nionoid 

 glochidia by fish is well known but the dispersal of gasteropod 

 eggs is yet to be proven by an authentic instance. The fre- 

 quent occurrence of totally unlike faunas in different drain 

 age systems within a comparatively few miles and the fact 

 that there are no instances on record of discontinuous distri- 

 bution of species which may not be explained more easily 

 otherwise, is proof enough that this means is far less important 

 than these writers would have us believe. 



Aquatic mollusks frequently cross barriers never-the-less. 

 In several years' held experience the writer has found per- 

 haps half a do/en undoubted instances which have taken 

 place through natural agencies yet have not gone so far as to 

 have obliterated their clearness. In all cases the means were 

 explainable by a single simple cause, the geological phenom- 

 enon of stream-capture — the headwaters of a stream Mowing 

 into one drainage system by faulting, tilting of the strata, or 

 erosion becoming transferred to another drainage. The mol- 

 lusks in these headwaters may then pass freely into the sec- 

 ond system providing other agencies do not interfere. 



The San Bernardino Mountains afford a typical instance 

 at our very doors. , Near the north borders of this Range lies 

 Bear Valley which contains two considerable though shallow 

 bodies of water, Bear Lake and Baldwin Lake. According to 

 the San Gorgonio Topographic Sheet of the U. S. Geological 

 Survey the elevation of Bear Lake which occupies the west 

 and upper end of the valley is about 6750' above sea level. 

 Toward the east the elevation drops down gradually to Bald- 

 win Lake which has no present outlet but whose surface is 

 6674' elevation. This is separated on the east by a ridge 

 130' high (according to aneroid determinations made by the 

 writer) from Arrastre Creek which Mows to the Mojave Des- 

 ert. The ridge appears to be the result of a fault running in 

 a north-west, south-east direction between Cashenbury Springs 

 and the old Rose Mine cutting off what seems to have been 

 at one time an extension of the valley opening to the north- 

 east onto the Mojave Desert. As already noted Baldwin 

 Lake has no outlet, it is alkaline. Bear Lake which is little 



ii 



