202 DR. B. A. BENSLEY OX THE EYOLUTIOX 



Dorcojisis. The evolution of the series indicates a migration from Australia to New 

 Guinea. 



The derivation of Dovcojjsis is interesting from a general hiological standpoint, 

 inasmuch as there is direct evidence that in the whole course of its evolution it has been 

 three times terrestrial aud twice arboreal. The alternation of arboreal and terrestria 

 phases may be expressed as follows : — 



Arboreal forms Didelphvidie — ->- Phalangeridse Dendrolaqxts. 



Terrestrial forms. . . . Marsupial ancestors Primitive Ifacropodiiia: -»■ Macropus. 



Setomjx does not appear to be nearly related to Dendrolagits and Dorcofsis, although 

 it agrees wdth those forms in the peculiar characters of the sectorials and the small 

 degree of hypsodoutism of the molar crown. It is removed partly by its small size and 

 complete canine reduction, and partly by the jiurely terrestrial character of the pes ; all 

 the secondarily terrestrial species of the brachyodont line, namely those of Dorcopsis, 

 are Papuan, while Setonyx is Australian. It is probable that Setonyx is a member of 

 the Small Wallaby section of the genus JIacrojius which has assumed feeding-habits 

 somewhat similar to those of the tree-living Deiidrolagus. The fact that Dorcopsis, 

 although terrestria], continues the sectorial evolution begun in Dendrolagus, shows that 

 such developments do not necessarily begin and end with arboreal life. 



The members of the hypsodont series are best arranged as representing a minor 

 radiation, the divergent characters oi the various genera being, for the most part, 

 associated with differences of environment, and not affecting to any appreciable extent 

 the dentition or foot-structure. Ouychoyale, however, appears to be more nearly related 

 to Macrojyus than to the remaining genera, and Lagorchestes, of which Lagostrophus is a 

 modified form, aj)pears to be similarly related to Fetrogale. The typical and predominant 

 forms of this radiation are the species of Macropiis. It is an interesting fact that this 

 genus, in -n hich the m.ost important stages of the grazing evolution are to be found in 

 point of numbers of species, stands out as the successful type of the family, just as the 

 genus Fhalavge7- does in the Phalangerinse, and JPsendochiriis in the Phascolarctinae. 

 Within the genus the Small Wallaby, Large Wallaby, and Kangaroo sections rejiresent 

 successive stages of the same graminivorous and saltatorial evolution. The predecessors 

 of the Small Wallabies appear to have undergone a distributional differentiation with 

 the production of the various species comprising the section. The latter appear to have 

 given rise locally to Large Wallabies, aud these in turn to Kangaroos. This sequence is 

 indicated, first, by the obvious affinities and wide distribution of the Small Wallabies; 

 secondly, by the fact that it is possible to differentiate between small and large premolared 

 species, especially in the two lower sections ; and, thirdly, that there are indications of 

 relationships between certain members of the Small and Large Wallaby sections having 

 similar geographical ranges. Thus the West-Australian Small Wallaby {Macropns 

 JEngenii) is prototypal to the Large Wallaby {M. inna) of the same region, and also to 



