106 
the root itself being relatively weak. Hence we may conclude 
that this  posterior heel-cusp is the true cusp 6, the anterior 
one being only a strongly developed cingulum-cusp on the ridge 
passing from 6 to 5, as Winge also supposed in 1886 (l. c. 
pag. 62) without knowing the position of the roots. All the 
cheek-teeth of Æluropus show a great tendency to form secondary 
cusplets. — Thus in this genus cusp 7 of pt is wanting, the heel 
being unicuspid and placed far back. In. 
this respect it resembles pt of Ursus, 
whose eusp 6 is also displaced backwards 
and is small, its root often being united 
with the root of 5; 4 is wanting, and 
(al the ftooth is in a rather rudimentary 
ae REE condition (fig. 14, diagr. 15). This is 
Fig. 14. Ursus aretos. perhaps the reason why a cingulum-cusp, 
ze mee: Rd is not developed. Thus both tooth-form 
6 and its feeble root united and tooth-formula affirm Winge's view, 
with the root of 8 Kro: that Æluropus is an aberrant member 
ne eå of the Ursidæ, not of the Procyonidæ- 
The other resemblances between Æluropus and Ælurus set forth 
by Lankaster and Lydekker seem to me to be mostly ana- 
logous features due to the adaptation to the same habits. 
It is not my intention to go through all the different modi- 
fications of pt of the Carnivora; but I hope that, from the views 
set forth here, it will be easy to homologise the single cusps im 
the teeth of the various families of that order; Mustelidæ are not 
the least instructive forms, for they show the transformations very 
clearly from the exclusive carnivorous to the herbivorous diet. 
We have here met with four different forms of the upper 
carnassial, which differences are essentially limited to the develop- 
ment of cusp 1-and 4: in Marsupialia (1 and 4 equal), in Carnivora 
primitiva (4 and 5 equal, 1. small), in Carnivora vera, milk-car- 
nassial (1 larger than 4 which may coalesce with the base of 5) 
and in Carnivora vera, permanent carnassial (4 larger than 1 which 
