606 MR. R. I. POCOCK ONT THE 



The deep, backwardly directed pouch of Trichosurus seems so 

 well adapted for the safe retention and carrying of the young up 

 in the trees that its replacement by a laterally directed poucli 

 with the orifice opening towards the tail, such as is seen in 

 Phascolarctos, seems most unlikely. The use of the lateral 

 extension for lodging the young carried by a mother who doubt- 

 less climbs at times head-downwards is easy to imagine, but the 

 reason for the reversion of the orifice demanded by the theory of 

 the descent of Phascolarctos from the Phalangerine stock is 

 puzzling. The structure of the pouch alone supplies evidence 

 that Phascolarctos is not closely related to any genus of the 

 Phalangeridse. 



The characters of the three types of pouches above described 

 may be summarized as follows : — 



a. The orifice, not encircled beliind, opening backwards and 

 downwards; cavity of pouch extended laterally on each 

 side along the depression between the thigh and the 

 abdomen Fhascolarctos. 



h. The orifice snbcircular, surrounded by a flap of skin all 

 round; cavity subcircular, a little deeper laterally than 

 elsewhere Fhascolomys. 



c. The orifice opening forwards, at the anterior end of the 



cavity, which is longer than wide Fsevdochirus, 



Fhalanger, Trichosurus. 



The Classification of existing Diprotodonts. 



My opinion on the classification of the Diprotodonts is as 

 follows : — 



1. Winge's removal of Phascolarctos from the Phalangeridse 

 must be accepted. But although that genus shows points of 

 resemblance to Phascolomys not shared by other existing Dipro- 

 todonts, the diflferences between them are too many and too 

 important to admit of their ascription to the same family. 

 Moreover, if we adopt as criteria of family rank such characters 

 as those distinguishing the Kangaroos from the Phalangers, the 

 characters separating the Koala and the Wombat should, I think, 

 be given superfamily rank indicated by the titles Phascolarc- 

 toidea and Phascolomyoidea. Assuming Winge to have coirectly 

 placed the extinct forms above referred to, the Phascolarctoidea 

 will contain the two families Phascolarctidee {Phascolarctos) and 

 Thylacoleonidee (TA?/^acoZeo) ; and the Phascolomyoidea the two 

 families Phascolomyidte [Phascolomys, Lasiorhinus, Phascolonus) 

 and Diprotodontida) {Diprotodon, Nototherium). 



2. The two above-mentioned superfamily groups are individu- 

 ally equivalent to a group of that rank, containing the rest of 

 the Diprotodonts, for which two names are available, Hypsi- 

 prymnoidea and Phalangeroidea, 1 prefer the latter on account 

 of its greater familarity and its derivation from the name of a 

 more primitive genus. Accepting for this group Thomas's families 

 Macropodidse and Phalangeridfe, it seems clear in the case of the 

 former that if the characters of Potorous and its allies entitle 



