107 
may, on the-other hand, coalesce with the base of 4). Hence it 
may be seen that the same tooth-form may be produced in different 
ways; if one who had not studied the history of the cusps, placed 
the carnassial of Canis, Thylacinus and Hyænodon beside one 
another, he certainly would suppose the single cusps to be homo- 
logous with one another, which is not the case. 
I imagine — although I have not been able to examine .suffi- 
cient material — that from the .facts pointed out here, it will be 
possible to settle the question of the classification of several of 
the carnivorous mammals whose position is still dubious. — Thus 
Palæonictis (fig. 15, diagr. 12), which is often placed among the 
Fig. 15. fresonne occidentalis. p? and m! from above, 
and m! from the outer side. /"/1. 
"Creodonts””, appears as a represeutative of Carnivora vera (cfr. 
Winge, 1. c. pag. 52). Not only is pf the carnassial, but it is 
formed as that of the true carnivores, cusp 1 is wanting, 4 is 
independent. No doubt m? resembles the creodont carnassial, 4 and 
5 being equal; the main difference is after all that 1 is wanting 
and 3 is very low and short; but the whole character of the tooth 
is more that of .a crushing than of a cutting agent, though not in 
as high a degree as that of m! of the true carnivores. But it 
forms the connection between this and the creodont carnassial and 
shows that not the recent carnassial (p?), but the recent tubercular 
tooth (m1 and m?) is derived from the creddont carnassial (m?, 
