TOOTH-GENESIS IN THE CAVIIDA. 283 
the incisors are reduced in number, there is an absence of 
canines with the presence of a diastema. The number of cheek- 
teeth in some of the Multituberculata is in excess of that found 
in most Rodents, while in others it is not in excess of that present 
in the Lagomorpha. In any case, this is only what might be ex- 
pected, as itts well known that a progressive reduction in the 
number of cheek-teeth is, and has been, taking place throughout 
almost the whole mammalian series. 
There still remain other points to be considered in the same 
connection. Within the limits of the existing Hystricomorpha 
very different patterns of cheek-teeth are to be found. The 
crown-surfaces of the so-called anterior permanent premolar of 
Aystrix leucura (Camb. Zool. Mus. 861 D) is shown in fig. 7, 
Fig. 7. 
Crown of First Upper Permanent Premolar of Hystrix leucura before eruption. 
Enlarged. (Camb. Univ. Zool. Mus.) 
and of Cavia cobaya in Pl. 26. fig. 7B. The former is decidedly 
multituberculate, the latter is not. In all the Hystricide the 
tooth-change is now known to occur and comparatively late in 
life, whereas in Cavia the deciduous tooth is shed zm utero. It 
may, therefore, justly be inferred that the Hystricide are more 
primitive in this respect than are the Caviide. Though this may 
not altogether justify any conclusions as to the pattern of the 
molar crowns, nevertheless, taken in conjunction with what has 
been said above, I think it affords some additional evidence in 
favour of the conclusion that the multituberculate is the primi- 
tive pattern of the Rodent molars. 
Lastly, there is a large amount of evidence collected by Forsyth 
Major in favour of this view, which he was the first to set forth 
in his paper on the Miocene Squirrels [6]. 
This opinion, however, was not shared by Cope, and is not by 
Osborn. Cope derived the Rodentia from the Tuillodontia, a 
suborder of the Bunotheria, from a type closely allied to Esthonyx, 
