DISCONTINUOUS DISTRIBUTION OF MAMMALS. 683 



to a graminivorous existence the argument has no cogency, since 

 we know that similar types may be evolved separately under 

 similar conditions from similar ancestors, or even from distinct 

 ancestors, a southern bridge being thus unnecessary. 



A second theory to account for the present localisation of 

 the marsupials is the polar theory, according to which all forms 

 originate in the Northern Arctic region, from which the older 

 forms are constantly pressed southwards by the newly-evolved 

 forms. Thus the oldest mammals, the marsupials, have been 

 pressed back to the two Southern points, where they are found 

 at present. This theory, however, leaves unexplained the 

 absence of marsupials from Africa, for it has been shown that 

 the data on which the occurrence of a sea in the region of the 

 present Sahara was postulated was erroneous. 



Beddard* discusses both views, and they are so little con- 

 vincing to him that he establishes a third theory, viz., that mar- 

 supials originated in the region of their greatest density — Aus- 

 tralia ; and from here wandered to Europe and America. He 

 mentions the difficulty presented by the absence of marsupials 

 from Africa, but ends up by saying that this absence may even 

 be used to support his theory ; for the emigrants from Australia 

 were so comparatively weak in numbers in Europe that colonies 

 were not pressed down into Africa. To this the obvious answer 

 would, of course, be the question : How is it, then, that they 

 were numerous enough to exert pressure in the direction of 

 America? 



From the above it will be seen that our conception of the 

 course followed by nature in leaving the marsupials stranded 

 in the two distant Southern regions is far from clear or unani- 

 mous. . 



The two forms of Australian marsupials which have the 

 greatest claims to being the ancestral form are probably the 

 Cuscuses and the Pouched Mice. Their claims I base on their 

 full dentition, long tails, general resemblance to living Creodont- 

 like Eutheria, and to the American marsupials, as also on their 

 distribution. The Cuscuses are on the borders of, and even in, 

 the Oriental region, and the Pouched Mice are widely distributed 

 over the Australian region, even to the Aru Islands. The 

 Cuscuses have retained the archaic prehensile tail, but one species 

 of Australian mice has also done so. The Pouched Mice have 

 retained the full dentition, whereas in the Cuscuses some of the 

 premolars have become functionless and vestigeal. The Pouched 

 Mice are more archaic in having a rudimentary marsupium, and 

 it is probable that both are somewhat specialised descendants of 

 the ancestral Australian marsupial. But if this view of the 

 origin of Australian marsupials be correct, i.e., that they became 

 diflferentiated in the Australasian region from a generalised 

 Creodont ancestor, and if. further, the American marsupials 

 originated in a similar manner from a like Creodont, we need 

 not be surprised that Africa has no marsupials, since we see 



* " Zoogeography." 



