OF THE AUSTRALIAN MAESUPIALIA. 



87 



deposits *, the extinct fauna is composed of large specialized terrestrial types showing no 

 more approximation to theoretically prototypal forms than, if ^Ye may cite an analogous 

 case, the memhers of the giant fauna of the mid-Tertiary portion of the placental 

 radiation do in comparison with the actually ancestral Creodonta. It will be apparent, 

 therefore, that beyond the limited although valuable evidence affoi'ded by geographical 

 distribution, or the occurrence of allied faunas living at different periods in other 

 countries, we are dependent for the history of the Australian series on the sequence of 

 adaptive characters as presented by linng forms. 



The results recorded in the present paper represent an attempt to construct a plan of 

 the Australian radiation by tracing tlie sequence of adaptive modifications in the dentition 

 and foot-sti'ucture of tlie various genera, in some cases of the species. The possibility 

 of making such an arrangement with advantage was suggested to the writer some time 

 ago by the perusal of the opinions expressed by Huxley and Dollo with reference to 

 an arboreal ancestry of the Marsupials, as determined by the characters of the pes. The 

 extensive observations of the latter writer in particular, although primarily designed to 

 show the presence throughout the group of a prehensile type of pes, appeared to indicate 

 further that most, if not all, of the progressive changes in foot-structure throiigh which 

 the Australian Marsupials have passed from the time of their origin are actually repre- 

 sented either in the existing Austrahan forms or in the American Didelphyidae, and that 

 it would accordingly be possible to explain in a measure the relationships of the various 

 genera by simply arranging them on a basis of the sequence of adaptive modifications 

 of the pes. The fact, however, that phylogenetic conceptions based on the adaptive 

 characters of single structures have frequently been found to be misleading when applied 

 to others, partly on account of the occurrence of convergent developments, and the 

 presence of a conspicuous example of convergence in foot-structure between the 

 Peramelidae and Macropodidte, indicated clearly the advisability of cliecking such an 

 arrangement by reference to the adaptive changes of other organs. For this purpose 

 the characters of the molar teeth were selected, the latter offering the advantage that 

 the main features of their evolution have already been demonstrated in the parallel case 

 of tlie Placentalia. 



The results of this study were published in a preliminary paper (1901 a), but on 

 account of many limitations both of material and literature they included only a general 

 or family arrangement. The opportunity of examining the case in detail was afterwards 

 afforded by a visit to London, where, through the kindness of the officials of the 

 Geological and Zoological Departments of the British Museum, tlie writer was able 

 to study the splendid collections both of fossil and existing Marsupials there preserved. 

 The latter form the basis of the present j^aper. 



Reference has been made above to the necessity of depending in the present case on 



* Cf. Broom (1S96). The presence of these forms in cavc-doposits will be noted as in all probability due to no 

 efforts of their own, and therefore as throwiug no light on the couditioua affecting their ancestors before the 

 differentiation of definite carnivorous forms. 



