1910.] Carpus of PhciKirodus, Mrtn'scothrritnn, Tdlif/rada. 447 



alternating carpus of Euprotogonia as well as that of the IJasal Eocene 

 Pantolamhda was also an inheritance from the Creodonts and that therefore 

 tlu^ serial carious of Phenacodus was probably secondary. This inij)ortant 

 generalization receives strong additional supjjort from the fact already noted 

 that the most essential features of the interlocking carjins were foreshatlowed 

 as far back as the Triassic and are still preserved in varying degrees in the 

 more primitive pentadactyl unguiculates of every order. ISIatthew's con- 

 clusion opened the way for a recognition of the ]>ossibility that the Perisso- 

 dactyla and Artiodactyla may have been derived from forms with a carpus 

 resembling that of the oldest Creodonts, and at the same time it imi)lied that 

 Phenacodufi in respect to its carpus was very far from being the " atavus " 

 of all the Ungulates. Matthew also concluded that the carpal pattern of 

 Phenacodus was probal)ly derived from that of Euprofogonin: (1) by the 

 flattening of all the elements, (2) by the reduction of the centrale and its 

 fusion with the antero-external angle of the sca{)hoid, (3) by the broaden- 

 ing and flattening of the magnum; this caused the reduction of the lunar- 

 unciform facet which however persists in a reduced condition in some 

 specimens. 



Meniscotherium. This Lower Eocene C^ondylarth is relatively highly 

 specialized in the degree of complication of its cheek teeth, and belonging 

 as it does to an order in which the carpals showed great variation, it A\()uld 

 not be surprising if its carpus also had undergone considerable modification. 

 As provisionally reconstructed by Matthew, the five toed manus appears to 

 lack the free centrale, and all the carpals are considerably flattened. The 

 trapezium is large. The lunar is suggestive of that of Ilyra.x. The mag- 

 num is flattened but is much smaller than in Hyracops socialis, as figured by 

 Marsh (Matthew doubts the accuracy of this figure). A small lunar-unci- 

 form contact is retained. In brief, the carpus of Meniscotheriur/i is becom- 

 ing serial like that of its contemporary Phenacodus, perhaps in adaptation to 

 exclusively ambulatory, as opposed to grasping, functions (cf. Hyrax). 



Taligmda. The carpus of Pantolambda (figured by Osborn, 1898, j). 

 187), a genus which in many characters is structurally ancestral to the 

 Amblypoda, is strongly interlocking and directly derivable from the Creodont 

 type. As noted by JNIatthew, the centrale occupies its primiti\c position 

 which is also the position of the internal hook of the scaphoid in Perisso- 

 dactyls, i. e. it articulates with the magnum, trapezoid, limar and scaphoid. 

 The trapezium is very large (as in Creodonts) and the pollex rather diver- 

 gent. The lunar-unciform contact is broad. This may be partly secondary 

 and a result of the progressive widening of the unciform which is character- 

 istic of the order. The manus of Peripiychus is not known, but judging from 

 the close resemblance of its pes to that of Pantolamhda, its manus was prob- 

 ably of the interlocking type. 



