74 MEMOIRS NATIONAL ACADEMY OF SCIENCES. [vol.xii. 



It was with considerable interest that after I have worked on this assumption for two years 

 I learned that Wetherby ('76, pp. 6-7) had long ago expressed a similar opinion as follows: 



Now on this hypothesis of origin if any ancestral types of the Strepomatidx remain at all, we shall find them among 

 the species inhabiting the upper part of the drainage or that part least affected by changes of level. * * * What 

 the headwaters of the Clinch may produce is yet unknown; and if the suggestion of Dr. Lewis be a just one, that shells 

 are propagated downstream, a part of the key to the solution of this problem lies in the mountain source explorations 

 of Clinch, Holston, and Powells. 



The persistence of these lime-requiring snails near the New River divide, rather than in the 

 correspondingly rapid waters from the Unaka Mountains and the Blue Ridge is at once 

 explained in part if we recall that lime-bearing rocks abound in the north whUe they are 

 conspicuously lacking in the higher mountains to the south. 



The longitudinal changes or variations in lo, progressively downstream, appear to be very 

 similar to what have been called morphologic-geographic chains, which have been described 

 in certain land snaUs by the Sarasins ('99, '01, 'Ola), Plate ('08), and Davenport ('10a). These 

 differences or series appear to be very similar to many geographic variations long ago pointed 

 out by Allen (Cf. Smithsonian Rep. for 1905, pp. 375-402, 1907, for mammals). Unfortu- 

 nately I do not have access to the work of the Sarasins and can not fairly discuss the bearing 

 which their studies may have upon the present subject. Plate, however, claims that the 

 Cerions of the island of New Providence, Bahama Islands, vary from west to east with increasing 

 rainfall and show an increasing number of ribs which give sculpture to the shell; at the same 

 time there is a decrease in size and the shells become progressively thinner. These differences 

 he attributes to the climate and to a responsiveness of the animal. Davenport, however, 

 revises somewhat Plate's observations on their occurrence and considers that the facts may 

 be explained if we assume an invasion of New Providence from the eastern end, accompanied 

 by migrations and diverse crossings, because he finds that Plate's "western" and "eastern" 

 types are more or less intermingled. 



5. CAUSES OF VARIATION. 



Since beginning the study of lo the most frequent question has been concerning the 

 use and cause of spines. Is it an advantage for the headwater shells to be smooth and to 

 lack spines? is a representative question. But aU of the headwater shells are not smooth, 

 as in the South Fork of the Holston, where they are mainly undulate, and in the upper Nolichucky, 

 where they are spinose. It has been suggested that the spines favor entanglement with objects 

 floating downstream and it might be an advantage not to possess them. I have not been 

 impressed by the force of this suggestion and I know of no advantage derived from the lack of 

 spines, nor have I been able to see any advantage derived from the possession of them. 



Physiologists have asked about the composition of the water and of its possible influence, 

 assuming that spines might be a response to its composition. That there are chemical differences 

 in the different parts of the rivers is well known, and further that these influences affect these 

 snaUs is also known. First of all it should be stated that we have no chemical survey of any 

 large river system in North America in sufficient detail to be of much use in such a study, cer- 

 tainly not of the Tennessee and its tributaries. That the sheUs do not abound in waters other 

 than those draining lime-bearing rocks clearly shows that lime is a limiting factor in their 

 range. Industrial refuse in the stream has already exterminated the shells from a part of the 

 North Fork of the Holston and doubtless sewage and mining products wHl continue this kind of 

 work in other localities. The relatively light weight of the sheUs in the upper Powell, when 

 compared with those of the upper Clinch and Holston, shows that when the drainage is largely 

 from nonlime-bearing rocks the shells are not as heavy as in the streams from lime rocks. 

 Originally a small salt stream flowed into the Holston near the locality from which Say's type 

 specimens were taken, but we have no record of its influence upon lo. The erosion of the 

 shells shows that there are considerable differences in the different streams. This is shown in 

 part by the plates of the shells. The shells of the upper Powell are iron stained, much eroded, 

 and relatively free from encrustation. The Clinch shells are much like those of the lower 



