OF THE AUSTRALIAN MARSUPIALIA. 109 



like. In this character Thylacinus differs markedly from the Dasyiirinte, more especially 

 the carnivorous forms, but resembles Prolhylacinus and Amphiproviverra. The talonitl 

 shows indications of tiie tliree original cusps. Of the latter the hypoconid is best 

 developed and tends to be slightly trenchant. In the fourth molar the talonid is 

 reduced to a small spur, as in Avqihiprovicenri. Aa interesting character is seen in the 

 first molar, Avhere the paraconid is comparatively well developed, as again in the last- 

 named form. As already pointed out, this structure is absent in the carnivorous 

 Uasyvu'inoe. 



Anlenwlar Teeth. — The incisors are very similar to those of Sarcophiliis. The median 

 upper teeth show signs of having been formei'ly procumbent, as in the insectivorous 

 Dasyurinoe, in being more rounded in section, or, in otlier words, less completely 

 spatulate than the lateral teeth, and also in being separated at their l)ases and approxi- 

 mated at their tips. The incisor-rows, unlike the premolar- and molar-rows, show the 

 same shortening exhibited by Sarcophilus. In this case the lower canines are almost 

 apical in position. 



The canines show no essential difference from those of Sarcophilus, except that they 

 are rather more slender and more evenly curved. 



The chief feature of the premolars is that they increase in size from before backwards, 

 the posterior tooth, as in the Protliylacinidse, showing no indications of the reduction 

 w hicli is so marked a feature of the Dasyurinse. 



PERAMELID^. 



Reference has already been made to the fact that the dental characters found in the 

 smaller insectivorous members of the Diisyuridfe are prototypal not only to those of 

 the larger carnivorous forms of that family but also to a considerable extent to thosi- 

 of the omnivorous Peramelida? and Phalangeridte. The close sequence of modifications 

 which is observable in the dental evolution of the Dasyurinse, and wliich is of so perfect 

 a kind that it is impossible to distinguish w^here the insectivorous evolution ceases and 

 the carnivorous evolution begins, shows at once that one is a direct natural continuation 

 of the other. Nevertheless, on the omnivorous side we find, in certain of tlie Peramelidie 

 and Pha]angerida3, modifications which are almost as inseparable from those of the 

 insectivorous DasyuridjE as are those of their carnivorous relatives, but which at the 

 same time usher in advanced modifications of a widely divergent kind. While, therefore, 

 it may be shown that the differentiation of the Phalangeridee in some respects represents 

 an insectivorous specialization, there is abundant evidence that as regards dentition the 

 primary division of the Australian radiation has been the differentiation of a carnivorous 

 and an omnivorous line from insectivorous prototypes. 



It is interesting to note that the Peramelidie, or including with them the Phalangeridse, 

 occupy much the same position as intermediate omnivorous types with reference to the 

 Dasyuridse on the one hand, and the members of the herbivorous section on the other. 



16* 



