592 MR. R. I. POCOCK ON THE 



Pseudochirus and Petauroides from the Phalangei-in'pe to the 

 Phascolaretina3 because of the crescentic (subselenodont) pattern 

 of the molar teeth in the three genera. Benslej^'s classification 

 was adopted, presumably with approval, by W. K. Gregory 

 (Ball. Amer. Mus. Nat. Hist, xxvii. pp._215-216, 1910), who was 

 acqviainted with but rejected the classification by Winge quoted 

 below. 



Thomas, of course, was quite familiar with the dental character 

 to which Bensley and Gregory attached so much importance. 

 On p. 167 of his Catalogue he pointed out that by the compli- 

 cated subselenodont character of its molars, Pseudochirus, with 

 its close ally Petauroides, stands somewhat apart from most of 

 the other Phalangers, and approaches Phascolarctos, in which a 

 similar but simpler modification is observable. But his reasons 

 for attaching to it subordinate systematic value lay apparently in 

 the circumstance that in young examples of Phalanger a tendency 

 towards the same structure is visible, but the crests on the molars 

 soon wear ofl:', leaving little diflerence between them and the 

 simple quadricuspid molars characteristic of typical genera of 

 Phalangeridae. The obvious, but not on that account necessarily 

 true, inference to be drawn from this fact is that the tendency 

 towards the subselenodont molar pattern exhibited by Phalanger 

 is a primitive character of the family Phalangeridae, which is tem- 

 porarily retained in Phalanger, lost in Trichosurus, Dactylo'psila, 

 and others, and elaborated in Pseudochirus and Phascolarctos. 



"Winge (E Museo Lundii, viii. pt. 1, 1893) held very different 

 views. He adopted two families : (1) Phalangistidaj [=Phalan- 

 o-eridfe] Avith the subfamilies Pseudochirini for Pseudocliirus 

 and Petauroides, and the Phalangistin* [=Phalangerinfe] for 

 the two groups Phalangistse {Phalanger, Trichosurus, Petaurus, 

 Tarsijjes, etc.), and Macropodes (llacropus, Hypsijifrymnodon, 

 etc.); (2) Phascolarctidge with the subfamilies Phascolarctini 

 (Phascolarctos) and Phascolomyini ( Phascolomys) *. 



Winge took as the basis for his classification the degree of 

 ■extension of the tympanic process of the alisphenoid, which in 

 the Phascolarctidae {Phascolomys, Phascolarctos) is small, does not 

 envelope the tympanic cavity, and fails to reach the paroccipital 

 process, whereas in the Phalangeridae, comprising the rest of the 

 genera of Diprotodont Marsupials, the bone in question is large, 

 envelops the tympanic cavity, and reaches the paroccipital process. 

 As accessory characters, the vestigial tail, the presence of a cardiac 

 o-land in the stomach, and the loss of one of the two normal 

 pairs of teats further serve to distinguish the Phascolarctidpe 

 from the Phalangeridae t- 



* To these Winge added Thylacoleontiui {Thylacoled) related to Phascolarctini 

 and Diprotodontini {Biprotoclon, Nototherium) related to Phascolomyini. 



f Winge's valuable paper is inifortunately written in Danish. The statements 

 about the extension of the alisphenoid is taken from Max Weber's work. I have 

 not been able to verify it in the case of the two skulls of PJiascolarctos available 

 for examination. 



