1200 DK. C. A. BEXSLET OX THE EVOLUTION 



offer a slight suggestion of special relationship between Biirramys, the Bettongiinae, and 

 I'ltdlanger. 



Spencev has recently described under the designation Wj/niardia bassiana a fossil 

 Marsupial from the Table Capo Beds of Tasmania, in which he identifies characters of 

 resemblance to both joolyprotodont and diprotodont sections, and which he states " may be 

 regarded as intermediate between the former and the latter, and as indicative of a stage 

 in the development of the Australian Marsupials when the ancestors of the recent 

 Diprotodontia were beginning to diverge from the original polyprotodont stock, from 

 Avhicli they have been developed within the limits of the Australian region." This view 

 is based on a detailed study of the parts preserved, which unfortunately include neither 

 tlie dentition nor the front or hind feet. While it would be difficult to add to the 

 excellent compai'isons presented by Spencer, it is probable that the reference to the 

 animal as an intermediate form must be excluded, if for no otlier reason, on account of 

 its large size. Throughout the present paper tlie effort has been made to show that the 

 actual prototypal or central forms must have been comparatively minute insectivorous 

 animals. In the Didelphyidoe, which occupy an ancestral position to all of the Australian 

 Marsupials, we find that the most primitive forms are the minute animals belonging 

 to the genera Jlat'iJiosa and Tei'umys. Tlie species of the Oligoceue Perather'mm, whicii 

 is closely related to the preceding genera, and probably the actual ancestor of the 

 Australian series, were animals of small size. The most primitive members of the 

 Dasyuridse, and even of the Pbalangerida;, are of like proportions. Thus all the 

 forms which approach the liypothetical intermediate type are of small size. Not only 

 this, but the diprotodont modification itself may, as already explained, be shown to 

 represent an insectivorous adaptation wliich could only have taken place in comparatively 

 small animals. The relations of Wijnlardia are more jirobably with one of the advanced 

 genera, such as Tseudochirus, Fliascolarctus, or I'Jialaiiyer. 



MACEOPODID.^. 



With regard to the relationships of the PotorointE and Bettongiinae it has been 

 mentioned above that the modifications of the pes appear to contradict the general 

 subfamily division and generic sequence as determined by reference to the dentition, 

 Ilypsipi'ijmnodon showing a closer approach to Fotorous than to Bcttongia, M'hile 

 Calopri/mnns shows a closer approach to Betiongla than to Fotorous. There is no doubt, 

 however, that the correct plan of division is according to dentition, because the divergent 

 characters of the sectorial premolars, on which the dental division mostly depends, are 

 of a very fundamental kind. It is impossible to derive the straight and comparatively 

 unelaborated sectorials of the PotorointB and Macropodinai from the multigrooved 

 rotated sectorials of Hypsiprymnodoii. On the other hand, while Beltongia is widely 

 separated from II>jpsiprymnodo)i in foot-structure, it must have passed through a 

 Ui/psiprymnodon-stage in arriving at its present condition. The Potorointe have also 

 pasijcd through a Hi/j^sipri/mnodoii-stago in foot-structvirc, but this does not prove affinity 



