OF THE AUSTEALIAX MAESUPIALIA, 199 



The foot-structure is so liomogeueous throughout the family tliat it may practically ])e 

 neglected in making a phylogenctic arrangement of the genera. Evolution has proceeded 

 along two main lines leading respectively to tlie Phalangerinae and Phascolarctinge, the 

 chief distinctive features of these being the development of selenodont and huiiodont J R A R Y 



modifications of the molars. A third subsidiary line, represented by Tarsipes, is 

 characterized by dental retrogression and the development of a more advanced stage of 

 arboreal elaboration in the pes than is found in the remaining membei-s of the family. 



The facts of the dentition show that the phascolarctine line cannot be derived from 

 the phalangerine one, or vice versa. In the former the most primitive form is 

 Pseudochirus. Petcmroides, which agrees with Pseudochirim in dentilion, must be 

 placed as a derivative of that genus on account of its volant character. Phascolarctus, 

 the terminal form, represents, in all its dental characters, a direct advance on 

 Pseudooldrus*. 



In the Phalangerinse the most primitive members are Acrohates, Bistcechitrus, and 

 Drmnicia, these forms showing in some respects an interesting approximation to an 

 insectivorous prototype, yicrobafes and D/sfcec/mrus are more closely related to one 

 another than either of them is to JJromicia. They are derivatives of a common stem- 

 form, from which they have become specialized by the reduction of the posterior 

 j)remolars ; they have diverged through the development of Acrobates into volant 

 animals. In the extent of reduction of the posterior premolars, and in the development 

 of special insectivorous characters in the lower median premolars and first lower m.olars, 

 DlstcecJmrus represents a more advanced form. Dromicia shows a more marked tendency 

 towards the omnivorous modifications characteristic of the larger specialized genera. 

 The volant form Petaurus occupies a derived position with reference to Dromicia, so 

 that its non-volant relative Gijmnohelidexis, which is said l)y Thomas to be identical with 

 it in dentition, is probably a derivative of that genus. Buctylopsila represents a du-ect 

 advance in dentition on Petaurus, and is to be regarded as another derivative of 

 Oymnobelideus. The derivation of the genus Phalaiiyer is a somewhat difficult matter. 

 As noted above, it represents a direct advance on Dacli/lopsila in the reduction of the 

 vestigial teeth, especially the median upper premolars, and also in the development of 

 incipient herbivorous characters in the molars. On the other hand, it departs from botli 

 Dadi/lopsila and Petaurus in the abrupt substitution of herbivorous for insectivorous 

 characters in the incisors, and of well-developed sectorial premolars, somewhat like those 

 of Dromicia nana, for vestigial ones. Tricfiosurus is plainly a slightly more specialized 

 form of Phalanger. The fossil form Burramys (Broom, 1896) should probably bo 

 included in the Phalangerinit>, on account of the lack of development of tlie median pre- 

 molars, which removes it from the Bettongiina?. The characters of the sectorial premolars 



* In connection ■with the present interpretation of the afMnitics of Phnscolarctus it is interesting to note that in 

 the structure of the manus the animal agrees with Psewlochirus, and differs from all the remaining members of 

 the family, the first and second digits being arranged so as to oppose the tliird, fourth, and fifth. The peculiar 

 straightening out of the normally intlccted angle of the jaw, which is so noteworthy a feature of I'/iascolarcUis, is also 

 indicated in Psexidocldrus CooJci, and to a certain extent in some other species. Other species of Pseudochirus 

 such as P. Albertisi and P. Corinncf, have the angle fully inflected as in normal phalangerine forms. 



