36 Transactions of the South African Philosophical Society. 



greatly reduced in the Cetaceans. The discovery of Galechirus 

 seems to show that there were primitive Therocephalian-Iike reptiles 

 in which the teeth had not become specialised and the coronoid 

 process' not yet developed, and that possibly the Anomodonts are 

 descended from these early types. Though the presence of the 

 large canine in Dicynodon seems to point, on the other hand, to an 

 ancestor with specialised dentition. The presence of well-ossified 

 sclerotic plates is a Diapsidan character, but sclerotic plates have 

 been found in Lystrosaurtis by Owen, and myself in Oudenodon. 

 They are unknown in any Therocephalian or Cynodont. It is to be 

 regretted that the temporal region is unknown. Probably there will 

 be found a single temporal fossa, and a descending quadrate. 



The shoulder girdle belongs to the type which is common to both 

 the early Synapsidan and early Diapsidan reptiles. If the cleithrum 

 is absent it is a very important specialisation, as it is still retained 

 in the Anomodonts and probably also in Theroceph'alians. The 

 absence of a distinct acromion process is in agreement with the 

 Therocephalians and a point of difference from the Anomodonts. 



The humerus agrees with that of the Therocephalians in general 

 characters, but it is more primitive in retaining the ectepicondylar 

 foramen. It differs greatly from the Anomodont humerus. The 

 absence of the olecranon process from the ulna is a point of 

 difference from all known Therapsidans. Even Pareiasaurus has a 

 well-formed olecranon. The ulna of Galechirus is almost typically 

 Diapsidan in type. 



The carpus is primitive and very little specialised. The number 

 of elements and the arrangement is not very dissimilar to that in 

 Sphenodon, but the resemblance is still greater to that of the 

 Anomodont as exemplified by Oudenodon. The most noteworthy 

 features of the structure are the presence of a 5th carpale, apparently 

 absent from Theriodesnius and the Anomodont Opisthoctenodon, but 

 present in Oudenodon, and the interposition of the inner centrale 

 between the radiale and the lst_, carpale. In this latter character 

 the afi&nity is again with the Anomodonts rather than with the 

 Therocephalians. 



In its digital formula Galechirus is a typical Therapsidan, and 

 shows no trace of the Diapsidan formula as seems to be present in 

 Theriodes7nus. There seems, however, to be some suggestion of the 

 Diapsidan in the lengths of the metacarpals. Galechirus most 

 probably walked with the digits directed somewhat outwards. 



There is too little of the vertebrae preserved to be of much service 

 in the tracing of affinities. The presence of intercentra in the 

 cervical region is a primitive character met with in most early 



