Of THE WOMBATj koala, AND PHALAKGEES. 887 



the result of convergence must be excluded as they are not 

 evidence of affinitj'. Thirdly, those features which are common 

 to all the animals under considei'ation, to many other Marsupials, 

 and to other Mammals cannot be employed in making a classifi- 

 cation of the Wombat, Koala, and Phalangers. We are then left 

 with a series of useful characters, and those which are sheltered 

 deep down in the body should be of value, for they are less liable 

 to be inflvienced by climate and habits than are the more super- 

 ficial characters. 



The anatomical characters must be aiTanged in four groups : — 



A. Those in which Phascolarctos and Phascolomys differ from 

 the Phalangeridse. They support the systems of Weber (12) and 

 Winge (13), and Weber's family of Phascolarctid^ is employed 

 here for the Koala and Wombat. 



B. Those in which Phascolarctos and the Phalangeridfe differ 

 from Phascolomys. Special attention must also be paid to the 

 comparison between Phascolarctos and Psetidochirus . These 

 support the systems of Thomas (11), Bensley(l), and Gregory (4). 



C. Characters in which Phascolarctos differs from Phascolomys 

 and the Phalangeridse. 



D. Characteis differing in all forms. 



Groups and D are evidence in favour of Pocock's system. 



In estimating the value of the anatomical data one must always 

 bear in mind the habits and the character of the diet of these 

 animals. Phascolarctos and the Phalangeridse are arboreal 

 animals, but lead different kinds of life. The latter are active, 

 but the former is a clumsy, sluggish creature, clinging to branches 

 by its modified feet. Phascolomys, on the other hand, is an active 

 animal. Its hind-limbs are fossorial and its fore-limbs are for 

 active progression. A study of the myology shows that these 

 modes of life are accompanied by numerous differences in the 

 muscles. As regards the diet, that of Phascolarctos and Phasco- 

 lomys is bulky, whereas that of the Phalangerida3 is not ; and the 

 PhascolarctidsB have specialisations in the structure of their 

 alimentary canal to meet the demands imposed by it. Finally, 

 although the life habits differ, we find the arboreal sluggish 

 Koala and the active fossorial Wombat possess numerous 

 characters of great importance in their skeletons and soft parts, in 

 which they differ from the active arboreal Phalangeridfe. We 

 must always aim at discovering the part played by each structure 

 in the animal economy before we can assess its true systematic 

 value. Some characters are well understood, but there are others 

 of whose general adaptive purpose we are totally ignorant. 



A. Characters in which the Phascolarciidce differ from the 

 Phalangeridce. 



External Characters : — Pocock (9) showed that the supratragus 

 in the Phalangers and many other Mammals is a prominent ridge 

 with a well- developed lobe ; but it is low, inconspicuous, and 



60* 



