230 PROCEEDINGS OF THE ACADEMY OF [1889. 



to be considerably longer and more slender tban those of Dinictis 

 (Filhol No. 4, PI. 4, figs. 1 and 5). 



The phalanges of Dinictis are in general very much like those of 

 Cryptoproeta but are proportionately stouter. Those of the prox- 

 imal row are quite long, broad and strongly arched towards the 

 dorsal side of the foot. Those of the second row are flatter and quite 

 deeply excavated upon the fibular side of the distal end, in order to 

 allow the retraction of the claws. This excavation is much more 

 marked than in Cryptoproeta or in any of the Viverridce which I have 

 been able to examine. According to Mivart (No. 12, p. 157) the 

 claws of Prionodon are almost as completely retractile as in the. cats, 

 as certainly seems to have been the case in Dinictis. The ungual 

 phalanges are less strongly arched than in Cryptoproeta ; they are 

 short, very much compressed and not far from straight. In shape 

 they are quite like the unguals of Paradoxums, but the articular 

 surface is of a different shape and the backwardly projecting pro- 

 cess beneath this surface is much better developed, though not to 

 the same degree as in the true cats. There is no trace of the bony 

 lamina or hood which in the last-mentioned family is reflected over 

 the base of the horny claw. This hood occurs in Hoplophoneus and 

 probably also in the later Nimravidce from the John Day, though we 

 may infer that in this respect Archcelurus agrees with Dinictis. 



It will be seen from the foregoing description that the hind-leg 

 and foot of Dinictis presents a very curious assemblage of charac- 

 ters. In some respects these characters are very primitive, as for 

 example the large third trochanter on the femur, the greatly flat- 

 tened ankle-joint, the articulation of the astragalus with the cuboid 

 and the shape of the calcaneum. Indeed, I was at first misled into 

 the belief that this specimen belonged to a creodont and accord- 

 ingly referred it to Hycenodon (quoted in Cope, No. 3, p. 344) until 

 material subsequently brought to light showed its true nature. Such 

 a creodont-like tarsus associated with the highly differentiated denti- 

 tion of Dinictis is certainly very remarkable. In general, the speci- 

 men' before us has much resemblance to the corresponding parts of 

 Cryptoproeta and other viverrines, as well as to Cynodictis, but at 

 the same time there are points of approximation to the felines, of 

 which the most important is perhaps the degree of retractility 

 shown in the ungual phalanges, which seems to be equalled among 

 the Viverridce only by Prionodon. 



