1907,] OF MAMMAL-LIKE REPTILES. 1051 



Galechirihs scholtzi (11). The facial region is not unlike that of 

 Palcnoh alter ia, there being no specialised canine. The lower jaw 

 in general structure is essentially similar to that of the Anomo- 

 clonts and differs from that of the Therocephalians in the absence 

 of the large coronoid process of the dentary. The shoulder- 

 girdle is like that of the Therocephalians and differs from that of 

 the Anomodonts in the absence of the cleithrum. The humerus, 

 ulna, and radius are like those of the Therocephalians, except 

 that the ulna has no olecranon process. The carpus is very like 

 that of the Anomodonts, and the digital formula is 2, 3, 3, 3, 3. 

 There ai-e intercentra in the cervical region and the ribs are 

 single-headed. There are large numbers of slender abdominal 

 ribs. The pelvis is plate-like, with the ilium small and passing 

 upwards. 



It will be seen that we have here a form which agrees with the 

 mammal-like reptiles in the digital formula, in the structure of 

 the shoulder-gii'dle, and for the most part in the limbs, but with 

 a somewhat more primitive mandible, but which differs from 

 them and agrees with the Diapsidan reptiles in having abdominal 

 ribs and a plate-like pelvis. Had the manus not been preserved, 

 I should have placed the form somewhere near the Pelycosauria, 

 and the Procolophonia among the primitive Diaptosaurians ; and 

 had the mandible not been so essentially Anomodont, I should 

 still have done so in spite of the digital formula. But the 

 combination of characters shows that we have a form on the 

 mammalian line, but not far removed from the Diaptosaurian or 

 Cotylosaurian origin. Exactly where the point of origin has been 

 is not clearly indicated, but the descent has most probably been 

 either from a generalised Cotylosaurian or from a primitive 

 Diaptosaurian. How the formula 2, 3, 4, 5, 3 was changed into 

 2, 3, 3, 3, 3 is not known, but in Galechirus we see some indication 

 of the change. The metacarpals increase in size from the first to 

 the fourth, just as is usually the case in Diapsidans, and this 

 seems to show that the limbs were directed outwards from the 

 body considera,bly, as in lizards, and that the reduction had but 

 recently taken place. In the Therocephalian Theriodesiinus (12) 

 there seems from the figures to be some indication of a transition 

 from the Diapsidan to the Synapsidan formula, but in an un- 

 doubted Therocephalian pes I have examined there is not the 

 slightest indication of the larger formula, the numbers being 

 the typical 2, 3, 3, 3, 3, and I feel inclined to believe that the 

 change has taken place in two rapid stages, 2, 3, 3, 4, 3 and then 

 ^, o, o, o, o. 



In the accompanying figures the development of a number 

 of important structures is traced through the various groujDS that 

 lie between the Ootylosaurs and the Mammals. The types in no 

 case lie in the direct line of descent and are only taken as the 

 best-known examples of the different stages of the development. 



I shall not take into consideration the skull generally, as 

 though it is well known in all the Therapsidan suborders it is 

 Proc. Zool. Soc— 1907, No. LXX. 70 



