Origin of Mammalian Carpus and Tarsus*.* • !■<• 91 



■ ' ', ■ ')',■''' 



centralia, we may conclude that two centralia exrsied'in'4b,e : ,(3®(ylo- 

 saurian carpus. 



In Procolophon (7, 8, 9), the Diaptosaurian, most nearly related to 

 the Cotylosaurs, the carpus is fortunately well known. Proximally 

 there are a well-developed pisiform, a large ulnare, and a fair-sized 

 narrow intermedium. The radiale appears to have been cartilaginous. 

 In the centre of the carpus are two centralia, and distally there are 

 four carpalia. If the radiale is rightly regarded as having been 

 cartilaginous, then the carpus would appear to have agreed with that 

 of Eryops as regards the elements except that in Procolophon the 

 carpale 5 is absent or cartilaginous. 



The structure of the skull renders it probable that the Thero- 

 cephalians are the direct descendants of the Cotylosaurs in the line 

 which gave rise to the Mammalia. Of the Therocephalians the 

 only carpus known is that of Theriodesmus (10). Fortunately the 

 specimen is well preserved and the degree of displacement of the 

 bones so slight that there can be little doubt as to the interpretation 

 of the various elements. Bardeleben (11) has recently redescribed 

 and figured the carpus. He has shown that there are four bones in 

 the proximal row — radiale, intermedium, ulnare, and pisiform — two 

 centralia and four carpalia. The figure of the carpus which I give 

 is only a slight modification of that given by Bardeleben. 



Among the Dicynodonts the carpus is pretty well known. In 

 most genera it is well ossified, but in the aquatic Lystrosaurus it is 

 mainly cartilaginous. The only Endothiodont carpus known is that 

 of Opisthoctenoclon agilis, but fortunately this is known by a fairly 

 satisfactory specimen. As in most primitive reptiles, the proximal 

 row consists of four bones — a broad short radiale, a small inter- 

 medium, a large ulnare, and a rather small pisiform. In the centre 

 of the carpus are two broad centralia. The distal row is formed of a 

 ]arge carpale 1, small carpalia 2 and 3, and a large carpale 4. On 

 the radial side of the carpus as preserved are three small bones, 

 which may belong to the hand, but as the specimen is of very small 

 size and crushed in on the base of the skull, it is impossible to be 

 quite certain of the nature of the fragments. I have figured them as 

 they occur. The two bones adjoining the first carpale and first 

 metacarpal are probably the bones of a prepollex, as in Theriodesmus. 

 The third little bone possibly does not belong to the manus. 



The carpus of Oudenodon (12) I have already elsewhere described. 

 It closely resembles that of Opisthoctenodon, but there appears to 

 be no trace of a prepollex. The proximal row consists of .radiale, 

 intermedium, ulnare, and pisiform. There are two centralia, and 

 apparently five carpalia. It is doubtful if the fifth carpale is really 



