302 Bulletin American Museum of Natural History. [\'ol. XXVII, 



already well established in the Triassic Cynodont Galesaurus, and preserved 

 not only in Marsupials and certain Creodonts but also in the early Artio- 

 daetyls {Achoenodon) and in the Edentates. 



In brief, the preceding list of resemblances between Marsupials, Creo- 

 donts and Carnassidents seems to fall roughly into two classes: (1) homo- 

 plastic resemblances and (2) primitive characters inherited by the groups in 

 question from very remote, perhaps Upper Jurassic or earlier, ancestors. 

 To this list of })rimitive characters might be added two others recorded 

 elsewhere by Dr. Wortman namely the small pelvic opening (p. 307) and the 

 Dideli)hid-like sternum of Dromocyon. (For the latter point see ^Matthew, 

 1906, p. 221.) 



On the other hand, deep seated differences between Marsujjial and 

 Placental Carnivores are not lacking. In cases like the one under considera- 

 tion no character or single set of characters can be relied upon invariably as 

 infallible guides to remote genetic connections. But in general the arrange- 

 ment of the cranial foramina (which Dr. ^Yortman especially has elucidated 

 in the Creodonta) and the underlying architecture of the skull seem more 

 likely to retain very ancient features than characters M'hich are more directly 

 related to a particular life habit. 



From Wortman's very careful and full description of the skull of Dro- 

 mocyon (1901, pp. 292-295) it is seen that even in the ]\Iesonychida?, the 

 Creodont family which shows the most numerous resemblances to the car- 

 nivorous Marsupials, the arrangement of the cranial foramina is distinctly 

 of the Placental Carnivore rather than of the Marsupial type. That is, 

 they possess an alisphenoid canal for the ectocarotid, the basisphenoid is 

 not perforated for the entocarotid artery as it is in the Marsupials, and the 

 "optic foramina are distinct from each other and occupy the usual position 

 in the carnivorous skull, just in advance of the sphenoidal fissure" (/. c, 

 p. 294), whereas in Marsupials there are no true optic foramina and the 

 optic nerves issue through the opposite sphenoidal fissures (f. 1. a) which 

 coalesce mesially. 



In regard to the accessory "condyloid" canal in INIarsupials Matthew 

 (1906, p. 214) has cited evidence unfavorable to its supposed homology with 

 the similarly named but somewhat differently placed foramen, in the Creo- 

 donta and Carnassidentia. 



The glenoid region of the squamosal of Dromocyon strongly suggests that 

 of Thylacynus, but the malar in Creodonts never enters the glenoid fossa, as 

 it does invariably in Marsupials. Nor is there any indication that the 

 alisphenoid participated in the articular surface of the glenoid, a very promi- 

 nent feature in all Polyprotodont Marsupials, and one showing clearly in the 

 Santa Cruz "Sparassodonts. " There is never an "alisphenoid bulla" 



