01' THE AUSTKALIAX MAESUPIALIA. 107 



minute, and in b the corresponding tooth of the opposite jaw is only slightly larger. 

 In d, which is the specimen described and figured by Thomas {op. cit., ])\. 27. tig. 5j, 

 the postei'ior premolar is almost fully formed, but is still beneath the bone. The 

 first members of the molariforra series are considerably larger than in the younger 

 specimen, approximating more nearly in size to the second members of the latter. A 

 somewhat similar size relation is seen in c, in which the posterior premolars are fully in 

 place. The minute size of the first molariform teeth in speciQ\en a ajipears to point out 

 these teeth as milk-premolars, which are afterwards replaced as in normal forms. The 

 dental variability is, however, so great in Myrmecohius that the evidence of a single 

 specimen must be regarded as suggestive rather than conclusive. 



Thylactnin^. 



In a former paper (190 1 «) the molar patterns of the Tasmanian Wolf {Thylacinus 

 cynocephalus) were described as representing the final stage in the carnivorous evolution 

 of the Dasyuridae. To this conclusion, which turns out to be erroneous, the writer was 

 led partly by a comparison of the molars of the animal with those of Dasyurus, the 

 only other member of the family at that time available, and partly by the view expressed 

 by Thomas (1888), that the relations of Sarcopliilus are with Thylaciims rather than with 

 Dasyurus. In the present paper Thylaciims is assigned to a separate division, and even 

 suggested as a foreign or unrelated element in the Australian family. 



Reference has already been made to the fact that the various genera of the Dasyiirinae 

 present successive phases of an extremely homogeneous dental evolution, the cliief 

 features of the transformation being : {a) reduction of the protocone in the upper molars 

 and of the talonid in tlie lower ; {b) approximation of the external styles of the upper 

 molars to their respective cusps ; and (<?) reduction with final obliteration of the posterior 

 premolars and of the paraconid in the first lower molars. In all of these characters 

 Thylacinus stands apart from the typical Dasyuridae, so that while it presents a degree 

 of carnivorous dental specialization only slightly inferior to that presented by Sarco- 

 philus, its evolution must have proceeded independently of that of the Dasyurinae, at 

 any rate in so far as the carnivorous stages of the latter are concerned. 



The resemblances between Thylaciims and the Sparassodonta of the South- American 

 Miocene have been noted by different writers and regarded as indicating affinity between 

 the latter group and the Dasyuridae. Lydekker (1899) has suggLsted the Sparassodonta 

 as the ancestral forms of the Australian family. Tlie cliaracters in which Thylacinus 

 resembles the Sparassodonta, however, prove to be exactly those in which it difl'ers from 

 the advanced Dasyurinaj, whose evolution can be directly traced to minute insectivorous 

 forms such as Sminthopsis and Phascogale. If the resemblances between Thylacinus 

 and the Sparassodonta represent affinity rather than parallel development we can 

 assume no closer relation between both of them and the Dasyuridse than is implied by 

 a possibly common origin of the South- American and Australian faunas. 



It must not be supposed that Thylacinus presents any dental characters vvhicli would 

 prevent it from being theoretically derived from one of the smaller dasyurine forms. 



SECOND SERIES. — ZOOLOGY, VOL. IX. 16 



