128 THE OLDEST MAMMALS. 



suet teetli iu wliicli tlie grooves are oblique. Moreover, 

 whereas the recent form was provided with four molar 

 t«eth (m 1 — Jit 4), the fossil had but two such teeth ; while 

 the form of these teeth was quite unlike in the two. 

 Hence, when we add that there are other important differ- 

 ences between them (into the consideration of which it 

 would be difficult to enter here), it will be apparent that 

 the view of regarding Plagiaulax as a near ally of the rat- 

 kangaroo was the result of attaching too much impiortance 

 to resemblances, and overlooking differences. Indeed, such 

 resemblances as do exist between the two may be regarded 

 merely as a well-marked instance of the phenomenon of 

 parallelism, treated of in an earlier chapter. 



Taking it, then, as proved that Plagiaulax is not a near 

 ally of the rat-kangaroo, we have to consider whether it 

 can be affiliated to any other group of existing mammals. 

 Before doing so, we have, however, to mention that there 

 are certain other Seeondai'y mammals allied to Plagiaulax, 

 in which the whole of the cheek-teeth are like the true 

 molars of the latter. We have already stated that in 

 Plagiaulax the lower molars have a median longitudinal 

 groove, and it may be added that the ridges bordering 

 such grooves are surmounted by a iLumber of small 

 tubercles. In the upjper jaw, if we may judge by some 

 allied genera, the molars had three such tubercular ridges, 

 separated by two grooves. Similar molars occur in the 

 skull repjresented in Fig. 43, which is that of a mammal 

 discovered a few years ago in the Secondary rocks of 

 South Africa, and named by Sir K. OAven Trilylodon ; 

 but it will be noticed that there is no trace of the 

 cutting and obUcj^uely-grooved premolar teeth of Pla- 

 giaulax, the premolars being like the molars. Detached 

 molars of similar type have been found in the Trias of 

 Stuttgart, and others occur in the Stonesfield slate. 

 I\Ioreover, in Dorsetshire and North America, there are 

 certain nearly allied mammals (Bolodon) in which the 

 ujijjer molars have only two, in place of three, longitudinal 

 ridges of tubercles. These forms, if other proofs were 

 wanting, clearly show, indeed, that the resemblance 

 between Plagiaulax and the rat-kangaroo is not a g-enetic 



