OF THE AUSTRALIAN MAESUPIALIA. 165 



For descriptive purposes the foot of Fhascocjcde fiavipes may be taken as representini,' 

 the type from which the modifications of the remaining forms have been derived. In 

 this animal the foot is relatively short and broad. In the specimen here figured (PL 7. 

 fig. 2) the ratio of the breadth, measured directly behind the hallux, to the length, 

 measured from the tip of the hallux to that of the third digit, excluding the claw, was 

 found to be in the proportion of 1 : 3-5. The hallux is greatly reduced as compared 

 with that of the Didelphyidje and Phalangeridse. As in the latter, however, it is set at 

 a considerable angle to the remaining digits, and is, further, clawless and provided witii 

 a slightly swollen terminal pad. The remaining digits are approximately equal in size 

 and provided with strong curved claws. Their most conspicuous feature is their parallel 

 arrangement and consequent restriction of lateral motion. The plantar surface is naked 

 and granulated, and bears five prominent pads, all of which are elongated and trans- 

 versely striated. One of these is placed at the base of the hallux, and extends from the 

 notch separating that digit from the second backwards towards the heel ; its posterior 

 two-thirds is separated from its anterior tliird by a sharp constriction. Three other pads 

 are situated at the bases of the outer digits ; they are completely separated from one 

 another, and the middle one is placed somewhat in advance of the others. A fifth pad 

 is placed on the outer margin of the sole opposite the middle of the hallucal pad. It is 

 probable that even in arboreal progression these structures are not so functional as in 

 those arboreal forms m which the hallux is well developed and the outer digits are 

 radially arranged. The original function of the hallucal pad is to oppose the outer 

 digital pads anteriorly and the marginal pad posteriorly, just as the hallux opposes the 

 outer digits. 



In the remaining species of Phascogale and in Dasxjurus hallucatus we find much the 

 same condition as in T. fiavipes, such differences as do exist, apart from the progressive 

 increase in size, being of a minor character. In P. Sicainsoni and its Tasmanian repre- 

 sentative P. minima the pads are shorter and more oval than in P. fiavipes. The 

 proportion of breadth to length in P. minima is approximately 1 : 3. In P. dorsalis 

 these characters are repeated, except that the constriction seen in the hallucal pad is 

 here replaced by a distinct gap. The proportion of breadth to length is 1 : 31. In 

 P. Wallacei the proportion is 1 : 38, and in P. ThorbecJciana 1:4. In the latter the 

 hallucal pad is undivided. P. apicalis, which was seen to be aberrant in its premolar 

 relations as compared with its size, is also aberrant in its foot-structure, the lieel being 

 elongated, so that the ratio of breadth to length becomes 1 : 4-2. The hallucal pad is 

 undivided, and the marginal pad is very thin posteriorly. In P. calura and its larger 

 representative P. penicillata the pads are elongated as in P. fiavipes ; the proportions 

 are respectively 1 : 5 and 1 : 375. 



The form described by Spencer as Phascogale macdonnellensis shows a partial 

 departure in foot-structure from the normal members of that genus. The foot is 

 relatively short, the proportion of breadth to length being 1 : 3'2(). It presents a peculiar 

 swollen appearance. The hallux is reduced to scarcely more than a tubercle. AU of the 

 pads are comparatively well developed, and their surfaces are transversely striated as in 

 Phascogale. The hallucal pad is here subdivided. According to Spencer's account, the 



23* 



