No. 2.) SNAn.S OF THE GENUS 10— ADAMS. 75 



Powell. The Saltville shells are iron stained; the South Fork shells are large, heavy, with 

 moderate erosion, and suggest an abundance of lime. The lower Nohchucky shells show con- 

 siderable erosion on old shells and they are not encrusted. The shells from near the mouth of 

 Little River are stained and eroded as in waters containing little lime. The shells from the 

 Upper Tennessee proper are in general not hon stained or encrusted, while farther downstream 

 (lot 151) they are eroded and with filamentous algse gi'owing on them. While not much weight 

 can be attached to these observations, they are clearly indicative of variable local chemical 

 conditions. From the standpoint of the chemical composition of the water the presence of 

 mixed communities of shells as near Rogersville, with great extremes from smooth to spinose 

 shell on the same shoal, is particularly confusmg, because we can not believe that in such a 

 situation the agitated waters can show a corresponding chemical diversification. 



Another characteristic closely associated with smooth shells, but not confined solely to 

 them (in group 3), is the relative degree of globosity of the aperture in terms of the diameter 

 of the shell, the shell index. The relatively globular shells are headwater shells in the Powell, 

 Clinch, and Holston. It is possible that this feature is influenced by the size of the foot, because 

 we may infer that a larger foot is an advantage over a smaller one to an animal living in rapidly 

 jflowing water. Cooke ('95, pp. 89-90) has shown that in the marine shell. Purpurea lapillus, 

 the exposed shells were stunted with a short spire and have a large mouth or aperture, while 

 the shells from sheltered places were larger, often had a spire and a smaller aperture. E-ussell 

 ('08) found that the limpet. Patella vulgata, upon an exposed shore was dwarfed somewhat, the 

 shells were thick and heavy and thus in contrast with shells from sheltered situations. Walker 

 ('09, p. 289, fig. 63) has shown that the large pond snail Lymnaea stagnalis produces local 

 or habitat variations. In exposed situations a smaller low-spired shell with a large body whorl 

 aperture is produced (variety B), while in quieter waters the shells are larger, have a higher spire, 

 and when the aperture is very large the shell is thinner. Semper ('81, p. 440) states that Nerita 

 mortoniana lives in salt, brackish, and fresh water. The salt-water form is smooth, but the 

 brackish and fresh-water forms often develop spines, and a related subgenus Cliton, which is 

 spinose, is characteristic of streams. 



In turrita and its allies the shells tend to elongate and the aperture remains small, but 

 lovdonensis has a shorter spire and a larger aperture and lives in the same situations as turrita. 



The variations of the size of the shell (small lo in small streams, and large in large streams) 

 appear to be a general rule, although there are some exceptions. Thus the headwater shells 

 (group 6) of the Clinch are large shells, and the angitremoides from the French Broad and 

 upper Tennessee are relatively small shells which live in a large stream. The size of the shells 

 is probably influenced to some degree by the abundance of lime available. 



The climatic differences between the conditions where the headwater shells live and those 

 farther downstream deserve mention as a possible cause of variations. The differences in the 

 size of the shells from headwaters downstream may not be due solely to the size of the stream 

 but to the combined influence of the size of the stream and to the climatic difference, because 

 the streams become larger southward. To the south there is not only the higher average tem- 

 perature but there is also the longer growing season. A comparison of the mean annual tem- 

 peratures (BuU. Q, U. S. Weather Bureau, 1906) of Big Stone Gap, Va., and Wytheville, Va., 

 which may be taken to approximate the climate of the headwater shells is 54° and 53° F., 

 respectively, while in Knoxville and Chattanooga it is 57° and 60° F., respectively. The 

 temperature of the air thus averages a few degrees higher. If the average latest date of killing 

 frost (for vegetation) in the spring and average latest dates of killing frost in the fall are used 

 as rough estimates of the warm growing season for snails, the southern localities have a spring 

 about a month longer, and a fall about two weeks longer, thus in all a growing season from six 

 to eight weeks longer than have the headwater shells. It thus seems reasonable to expect that 

 the difference in climate is to some degree responsible for the larger shells to the soath. 



From the preceding discussion it will be seen that the causes for the differences observed 

 in these shells may be due to several factors. The ones just mentioned appear to be due pri- 

 marily to the character of the medium in which the animal lives. Another set of influences 



