NO. 2.] SNAILS OF THE GENUS 10— ADAMS. 37 



Group 26 also shows great variabiKty and great distance between the spines, as is shown 

 by the maxima at 15.5 mm. and 16.5 mm. This is the maximum modal condition of signifi- 

 cance, and this, as a group, contains the spmes with the greatest distance between them from 

 any locahty. They are of the loudonensis type of shell. 



Group 27 is of the turrita type, and, as might be expected, has a much narrower distance 

 between the spines. The mode is at 11.5 mm., and the curve is remarkably symmetrical, ex- 

 cepting for the skewness on the upper side at 14.5 mm. These shells, from Indian camps, should 

 be compared with the same kind in group 25 from an Indian shell heap on account of the simi- 

 larity in their range of variation. In both cases the number of individuals is limited. 



Taking this series as a whole, it is noticeable that on account of the lack of smooth shells 

 there is, of course, no discontinuity between smooth and spinose shells, as in the case of the 

 other streams. Contrasting the headwater shells, of turrita affinities (gTOup 20), with those 

 below Chattanooga (group 27) there is seen to be a marked difference, a break which is com- 

 pletely bridged by an intergrading mode (group 22). Groups 21, 24, and 26 show correspond- 

 ing changes in the loudonensis shells. 



c. Spine Indez. 

 A. POWELL EIVER. 



Group 1, The spine index is the average of j^— ^ , , — -. and gives a percentage 



■^ ^ ^ distance between spmes, & r *= 



ratio in terms of spine height. In group 1 this index is modal at 12 per cent, plate 22, No. 1. 



Group 2 has the same mode. 



Group 3 shows a progressive increase to a mode at 17 per cent, in harmony with a tendency 

 also shown in groups 1 and 2. 



Group 4 shows even greater advance with a mode at 22 per cent and with a decided skewness 

 toward the preceding groups. 



Group 5 is modal at 22 per cent. Considering the river as a unit there is a downstream 

 progressive increase of index from 12 to 22 per cent. The gradual character of the change 

 from one extreme to another is quite remarkable. 



B. CLINCH EIVER. 



Group 6. Smooth shells. Group 7 is modal as groups 1 and 2 of the Powell, at 12 per cent, 

 plate 23, No. 7. Group 8 has progressed to 17 per cent, Group 9 to 22 per cent, the extreme 

 for this river. Group 10 has declined to 17 per cent, and Group 11 advances to 22 per cent. 



The stream, as a whole, shows greater irregularity than the Powell. If group 10 is excepted, 

 there is a progressive downstream increase from 12 to 22 per cent, as in the Powell. 



C. HOLSTON EIVER SYSTEM. 



Group 12 is modal at 12 per cent, plate 24, No. 12, as is also Group 13, but the latter in- 

 clines toward a higher degree. Groups IJf, 16, and 16 are all modal at 17 per cent. Groups 17 

 and 18 continue to a higher mode at 22 per cent. Group 17 is transitional between the lower 

 ndodes and group 18. This gradual progressive downstream increase in shell index from 12 to 

 22 per cent recalls that of the Powell and indicates a considerable local degree of stabifity. 



D. NOLICHUCKY, FRENCH BROAD, AND TENNESSEE SYSTEMS. 



Group 19 is modal at 17 per cent, plate 25, No. 19, a higher mode than in other headwater 

 groups. Groups 20, 21, 22, 23, and 25 are clustered at 22 per cent, and Groups 24, 26, and 27 

 are at 27 per cent. If the turrita types are assembled, groups 19, 20, 22, and 27, their respec- 

 tive modes are at 17, 22, 22, and 27 per cent, a progressive series. By a similar arrangement 

 of groups 21, 24, and 26 the modes — 22, 27, and 27 per cent — are also progressive. Thus the 

 approximately homogeneous series of shells show a progressive downstream increase in spine 

 index. 



Spine index is dependent upon two variables, the height of spines and the distance between 

 them. If progressively downstream both varied directly, or inversely, at the same rate, the 

 index would not change, and if spine height increased more slowly than the distance between 



