198 The American Naturalist. [April, 
the ridge formula changes in so gradual a manner as to render 
it unavailable as a basis of exact divisions, as has been re- 
marked already by Lydekker. It is evident that the primi- 
tive Proboscidia had incisor teeth in both jaws, and that these 
had more or less of the usual enamel investment. The grad- 
ual modification of these features is therefore another indica- 
tion of the line of descent of these animals. The primitive 
Proboscidia had likewise four premolars, as is now seen in 
Dinotherium. The successive loss of these teeth is no less an 
index of the evolution of the modern types of the order, than 
the other modifications referred to. In general, then, the 
phylogeny of the order may be represented thus : 
Primitive Proboscidia. 
Within each genus certain parallel modifications of the 
composition of the crowns of the molar teeth may be ob- 
served. The cross-crests may be single, or they may be di- 
vided up into tubercles. The valleys between them may be 
open (i) or they may be blocked by (2) a system of single in- 
termediate tubercles ; (3) by numerous intermediate tubercles ; 
or (4) by the thickening of the primary tubercles. I arrange 
the species according to these characters. 
