278 ORTMANN— CORRELATION OF SHAPE AND 



are connected by intergrades, and any line drawn between them must 

 necessarily be artificial and arbitrary. It is advisable, however, to 

 retain the old names in a varietal sense, for the extremes are rather 

 strikingly different. 



Group of Fusconaia pilaris in the Upper Tennessee- 

 drainage. 



Fiisconaia pilaris (Lea) is the representative of F. subrotunda 

 in the upper Tennessee system, and it is extremely hard to distin- 

 guish the two. All I can say is that pilaris is a smaller shell than 

 subrotunda, a character mentioned also by Simpson. For the rest, 

 there is no difference, and it is actually impossible to tell younger 

 subrotunda from pilaris, as is shown by the fact that pilaris re- 

 peatedly has been reported from the Ohio River. In the Tennessee 

 River in northern Alabama, typical subrotunda is found, and I have 

 specimens from this region. But since I have no material from this 

 river between these parts and the vicinity of Knoxville, Tenn., I 

 cannot discuss the relation of these two forms. They may pass 

 into each other. 



The pilaris-gvoup of the upper Tennessee contains a number of 

 nominal species which have been distinguished on entirely insuffi- 

 cient grounds. I have revised them^^ in the following way. 



I. Fusconaia pilaris (Lea). — D. 55 per cent, or over. 

 U)iio pilaris Lea, '40 (French Broad and Holston). — Quadrula 



pilaris Simpson, '14, p. 893. 



Dia. (Lea) : 63 per cent. Simpson gives two measurements, 

 47 and 54 per cent., which, however, belong to the next form, which 

 he unites with this. 

 Unio globatus Lea, '71 (Holston). — Quadrula globata Simpson, '14, 



p. 899. 



Dia. (Lea) : 68 per cent. ; (Simpson) : 67 per cent. 

 Quadrula andrctvscc Marsh, '02 (Holston). — Quadrula andrcwsi 

 Simpson, '14, p. 895. 



Dia. (Simpson) : 55 per cent. 



^^ Proc. Amcr. Philos. Soc, 57, '18, pp. 527-529. 



