VERTEBRATE PALEONTOLOGY IN 1909 655 



In regard to the relationships of Carnivora, Dr. Matthew 

 disputes the alleged origin of the whale-like zeuglodonts from 

 the creodont division of the order, arguing that the creodont 

 resemblances of the zeuglodont Protocetus are restricted to 

 certain features shared with Hycenodon, a specialised Oligocene 

 cursorial type of the Creodontia ; such features being absent 

 in the primitive Eocene members of the latter group, such as 

 Sinopa. The presence of these features " in a marine mammal 

 of Middle Eocene time cannot be due to relationship, for if it 

 were it would be shown more or less clearly in Sinopa, the 

 early Eocene ancestor of Hycenodon. On the other hand, 

 the resemblance between Protocetus and the more generalised 

 creodonts does not indicate any very near relationship, except 

 such as we should expect a primitive cetacean to show to the 

 primitive members of other orders." 



The last sentence is noteworthy on account of its containing 

 an admission that, in the author's opinion, the zeuglodonts 

 are members of the Cetacea; this view not being shared by 

 all American palaeontologists. 



After discussing the classification of the order, Dr. Matthew 

 proceeds to observe that the Cretaceous ancestors of the Carnivora 

 appear to have been small arboreal mammals resembling the 

 North American opossum in size and habits, but more nearly 

 related to the primitive Insectivora. The angle of the jaw 

 was not inflected and the cheek-teeth were of the opossum-type, 

 although their cusps were at first conical, and did not acquire 

 the notched shear-like form characteristic of Insectivora and 

 opossums till a later date. 



In discussing a suggested relationship between the Creodont 

 Carnivora and the Marsupialia, the author remarks that 

 although the carnivorous representatives of the latter are in 

 many respects of a very archaic type, yet they differ in so many 

 important characters from creodonts, and indeed from all primitive 

 placentals, that any true kinship seems impossible. " It would 

 appear, therefore," he adds, "that the distinctive characters of 

 the marsupials, whether primitive or secondary, were at all 

 events fixed and constant as far back as the early Tertiary, at a 

 time when the various placental orders show a marked degree 

 of approximation to one another." 



Dr. Matthew divides creodonts into three main groups. 

 First, an " adaptive " group, as represented by the families 



