92 
partly misunderstood —, and his theory is not given; evidently it 
is considered as a "quantité negligéable.” 
Paul de Terra's "Vergleichende Anatomie des menschlichen 
Gebisses und der Zåhne der Vertebraten” 1911 which gives an impar- 
tial review of the theories of Oshorn, Kikenthal and Råse 
a. 0,, is here only mentioned because of the immense bibliography 
of everything appertaining to odontology. 
We see in this summary, where moreover only the more im- 
portant theories are referred to, such a multitude of hypotheses 
of the phylogeny of the mammalian teeth that we can be tempted 
with Leche to declare it impossible that any one should be the 
right one. But which is to be choosen in this "embarras de 
richesse” ? 
They can be divided into two main groups: The theories of 
concrescence and the theories of differentiation or cusp-addition. 
We will start with the first group. It is supported by Gaudry, 
Magitot, Dybowski, Kikenthal, Råse, Ameghino, Adloff | 
a. 0.and it supposes every complex mammalian molar to be the represen- 
tative of several reptilian teeth; the more cusps, the more coalesced single 
teeth. From this it may be concluded that the mammals with the 
most multicuspid molars are the most primitive; it would be absurd 
to think that they should first squander a number of reptilian teeth 
away by the shortening of the jaws and then acquire them again 
when it became a necessity to have larger teeth. — Talpa has 44 
teeth and so has Equus; but the number of cusps on the horse- 
teeth is greater than that on the mole-teeth; hence it might follow, 
if ihis theory were correct, that the teeth of Equus were more 
generalized than those of Talpa, but the reverse is the case. Sus 
and Phacochoerus have both $ molars, but må of Phacochoerus has 
far more cusps than that of Sus, though no one would suppose 
Sus to be the more specialised, Several other examples of mistakes 
in systematic relation could be referred to as the consequences of 
this theory; it is supported by no palæontological evidence, no fact 
