1910.] Interrelations of the MarsiipinI Si/borders. 227 



became greatly pronounced: possibly in i)roportion as the ancient pterygoid 

 muscles shifted their attachment from the (hvindling pterygoids to the 

 uprising pterygoid wings of the alisphenoid. The "alisphenoid bulla," 

 at first a mere depression in the hinder border of the glenoid region, for the 

 reception of the anterior wall of the membranous tympanic cavity, grew 

 backward, finally embracing the tympanic annulus and bridging over the 

 petrosal to unite with the paroccipital process and mastoid. 



Within the order Marsupialia the genetic relations of the suborders 

 Multituberculata, Diprototlontia, Ctenolestoidea Polyprotodontia appear to 

 be somewhat as follows: The Prototherian ancestors of the Monotreme- 

 Marsupial-Placental stem (p. 229) probably had incisors canines, premolars 

 and molars. Perhaps as far back as the Ui)per Triassic this type, by 

 dental reduction gave rise to the Multituberculata which paralleled the true 

 Diprotodonts and retained certain primitive features in common with them 

 but were nevertheless not ancestral to them (p. 170). 



The remaining stock, preserving their heritage of four kinds of teeth, 

 gave rise at different times to the Diprotodonts, Ctenolestoids and existing 

 Polyprotodonts. 



There can be little doubt that these three suborders are intimately related 

 to each other, as shown by the fundamental agreement in the foramina 

 (p. 222), in the general architecture of the skull (p. 221) and in the foot- 

 structure (p. 201). This inference is supported l)y the existence of several 

 groups (Peramelidte (p. 208), Csenolestoids (p. 209), Wijnijardia (p. 214)) 

 which combine structural features otherwise peculiar either to the Polypro- 

 todontia or to the Diprotodontia. ^Moreover, with regard to every detail of 

 the dentition, cranial foramina, alis{)henoi(l bulla, lower jaw, etc. all well- 

 founded analogies sustain the inference that, morphologically, the Kangaroos 

 have been derived from primitive Didelphids and not vice-versa. 



On the other hand in several features the Diprotodonts appear to be 

 more primitive than the Didelphids. It is only among the Diprotodontia, 

 for exam])le, that we find a structure which appears to be homologous, and 

 is certainly structurally identical with, the pecten of the eye in Apteryx 

 (Johnson, 1901). 



Again the malleus of Diprotodonts ajiproaches the Monotreme type, 

 in the length and breadth of the anterior (Folian) process and in the union 

 of the latter with the tympanic (Dobson, 1879); while the stapes is often 

 columelliform; and there is some evidence to show that these are very 

 primitive mammalian characters (p. 151); whereas in the Polyprotodont 

 Didelphis the stapes and malleus approximate to the Placental type (Doran, 

 1879, pi. Ixiv, fig. 15). The Diprotodonts, moreover, exhibit the tendency 

 for the persistence of both azygous veins to a greater degree than do the 

 Polyprotodonts (Beddard, 1907, p. 219). 



