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



related, although theu* range is discontinuous. These two bear much the relation to one another 

 that nolichuckyensis and spinosa do to each other. 



Turrita in the French Broad (lot 195) grades into nolichuckyensis. As recta approaches 

 both spinosa and unalcensis it appears that it may have been derived from the region to the 

 south of its present location where the other spinose shells flourish. Brevis, on the other hand, 

 appears to be most closely related to the complex mixtures of spinose shells of the lower Powell 

 and Clinch, and is not directly related to the other spinose shells. It may be a relic or repre- 

 sentative of the ancient form and a degree of spinosity developed as a local race in a transvefse 

 stream. The development of the turrita series in the Nolichucky and Tennessee Rivers is 

 strong evidence favoring the view that they have originated in this vicinity, and that spinosity 

 has spread elsewhere from here. 



Among the spinose shells the amount of differentiation in the forms indicates that not 

 only are some different degrees of spinosity inherited, but also the size and thickness of the 

 shell (angitremoides) , degi'ee of elongation of the shell, size of aperture, relative number of 

 spines on a whorl (turrita), and possibly other characteristics. 



The mixed shell populations not considered in the preceding discussion must now be 

 considered. The shells of the Powell River below lyttonensis appear to be a very mixed and 

 variable population, due to the intermingling of different degrees of relatively smooth and 

 spinose shells. The relatively smooth shells are possibly relics of an ancient colony antedating 

 the invasion of the spinose shells, which lagged behmd while the other migrated farther up- 

 stream. In the Clinch the degree of spinosity in the Clinchport shells is puzzling, particularly 

 when we consider that they are bounded upstream by paulensis and downstream by the small, 

 low-spined brevis. One wonders if some of the relatives of recta had not at one time, when the 

 Holston was tributary to the Chnch, had a colony of these shells which have since been absorbed. 

 Below brevis (lot 17) in the Clinch the population is again a mixed one, comparable to that of 

 the lower Powell. In the Holston confusion begun in the vicinity of Rogersville, with the 

 mixed population of smooth (fluvialis) and spinose shells of the loudonensis, spinosa, and recta 

 types, and a series which also includes the shells of inverse development. It may seem strange 

 that the dominance of the smooth shell does not spread over the lo population so that smooth 

 and not the spinose shells could be the most abundant kind. A similar condition is found in 

 cattle, in which the hornless condition is dominant when crossed with the horned, and yet horned 

 cattle and not the hornless are the predominant kinds. 



It may be possible to interpret the variations of these cross-fertihzed snails in terms of 

 pure strains, which are the nearest approach to the pure Hnes of the self-fertiUzed plants studied 

 by Johannsen, and Lang's experiments on Helix appear to harmonize with this view. There 

 may be pure strains of the smooth shells, of the undulated shells, of shells developing spines, 

 not only in different degrees but also at progressively earher ages until the entire postembryonic 

 life is spinose. Obviously it would be a large undertaking to prove this. Yet the individuality 

 of the 14 recognized forms readily lends itself to such an interpretation, particularly if we 

 assume, and with considerable reason, that almost any degree of a character may be inherited. 

 This gives this view wonderful adaptability and it gives another standpoint from which to 

 view the intergradations. The idea of pure strains harmonizes well with the form of inheri- 

 tance, Mendelian, which is so characteristic of snails and which probably applies to lo. The 

 variation which we see in dominance with age (inverse development) suggests the dominance 

 of smooth shells over undulations, and some degrees of spinosity. The different degrees of 

 inversion may be considered as due to the different amounts of the characters present in the 

 strains crossed, combined with normal fluctuations. In the Rogersville shells, which show the 

 greatest amount of inversion, the smooth shells appear to be crossed with shells of greater 

 spinosity than in the other locahties, and it looks as if the crosses are in a greater degree of 

 instability than when less contrasted forms are crossed. 



This change of dominance with age calls to mind the condition which we see in these ani- 

 mals which have marked secondary sexual characters. Often in such animals the young male 

 resembles the female, but with maturity it develops the secondary sexual characters of the male. 



