4 88 



THE AMERICAN NATURALIST. [Vol. XXXVII. 



number of specimens, 20 or 30, might show an apparent break 

 in the numerical variation and seem to warrant the separation of 

 some individuals as species, but a larger number of specimens 

 shows that the number of ribs cannot be used in the separation 

 of species without other more important characters. 



It would be interesting to know the results obtained by an 

 examination of a large number of specimens from different 

 localities, to ascertain the stability of the data recorded in this 

 paper. 



My thanks are due to Professor C. B. Davenport of the Uni- 

 versity of Chicago for valuable suggestions in carrying on this 



