NO. 2.] SNAILS OF THE GENUS 10— ADAMS. 83 



more weight in estimating the genetic relations than the degree of spinosity, although it must 

 be borne in mind that the time of life at which spines develop may be only an expression of the 

 degree of intensity of the spinose character, a lesser degree of spinosity may be said to be domi- 

 nant over a greater degree, because relatively smooth shells appear to be dominant in inversion. 



Turning now to the evolution of the different forms, it seems that on account of the lack of 

 spines, few undulations, and their occurrence in the headwaters, the smooth shells are the most 

 primitive. On examination of the forms of shells which appear to stand more or less inter- 

 mediate between the relatively smooth and those of the turrita type, as paulensis, lyttonensis, 

 lovdonensis, and possibly verrucosa, it is readily seen that they are not a homogeneous series of 

 transitional forms. All are geographically located between smooth and spinose shells, with 

 which they intergrade more or less, except loudonensis. Verrucosa intergrades perfectly upstream 

 wiih. fiuvialis and downstream with recta. Undulations are also present in the Upper Clinch, 

 in clincTiensis, and particularly in paulensis, which develops spines late in life. Paulensis 

 might be described as a combination of a smooth and undulate shell contaminated by spines 

 late in its development. Such a combination has the appearance of a product of relatively local 

 conditions which has permitted these three characters to be combined. Lyttonensis also appears 

 to be a combination of smooth shells and those with small spines, and the amount (relatively 

 large) of inversion associated with it suggests that different degrees of spinosity may be expressed 

 in such a combination, and it does not follow that all which is inherited is shown. Lovdonensis 

 appears to be a shell very similar to lyttonensis, and to a less degree to paulensis, in the ontogeny 

 of its spines, but in these the spines are of course much smaller, as they are from much smaller 

 streams than the Tennessee. Lovdonensis is a very spinose shell which develops from a smooth 

 young shell. The similarity of development raises the question as to whether or not lyttonensis 

 and paulensis are not simply geographic or habitat variations of the same form, which with the 

 diversion of the drainage early migrated far up each of these rivers, or on the other hand, were 

 they independently, or convergently, formed by the combination of the spinose shells invading 

 the streams and crossing with them in approximately the places where they now live? The 

 history of the streams seems favorable to either view. The lack of inversion in lovdonensis 

 might be interpreted to mean that it is not a combination, as in the other forms, but that it 

 has the characteristic to develop spines at a certain age. 



Verrv^osa is not intermediate between the smooth and spinose kind of shells, except in the 

 North Fork of the Holston. It is not found intermediate between the smooth and spinose shells 

 in either the Clinch and the Powell, where as a distinct type it is completely lacking. In the 

 Holston it forms a perfect transition between the sm.ooth. fluvialis and the spinose recta. Ver- 

 rucosa has invaded the headwaters of the South Fork as fluvialis has the North Fork, and 

 develops with such distinctness as to indicate its independent character. With departure 

 from the forks of the Holston undulations rapidly diminish. They are shown only to a small 

 degree in the upper Clinch, and to a much less degree in the PoweU, and what is even more 

 remarkable they are absent from that part of the upper Holston proper where the smooth and 

 spinose shells are so confusedly intermingled (near Rogersville). In view of these relations 

 it appears that these apparently intermediate forms of shells are not a homogeneous series, 

 because their homogeneous character is much more apparent than real. They appear to have 

 been formed by combinations of several different degrees of intergradation rather than as a 

 single series between the smooth and spinose shells. The presence of inversion in verrucosa 

 is significant and suggests that undulations as weU as spines are capable of suppression by the 

 dominant smooth shells. 



Among the shells which are spinose throughout post-embryonic development there is little 

 that can be used to measure values. As these shells are spinose from the start and the degree 

 of spinosity varies so much with the size of the shell and the stream the characters useful 

 in the other shells are not applicable here. There is remarkable unity and continuity in this 

 series: unaTcensis, angitremoides, nolicTiuckyensis, spinosa, and turrita. Brevis and recta are the 

 odd members of the series. The turrita series seem to converge toward the headwaters of the 

 Tennessee proper, as here turr\ta and spinosa converge, unaTcensis and angitremoides are closely 



