1910.] Summary of the Relations of Cynodontia to Mammalia. 141 



id) The probable position of the tympanic cavity in Cynodonts (Fig. 1, 

 B), immediately below and internal to the posterior end of the jaw (p. 127). 



{e) The position and relations of the su])i)Osed tympanic bone of ^fherap- 

 sids (p. 121). 



Summary. 



The foregoing review of the mammal-like reptiles of the Permian and 

 Triassic appears to show that it is quite unnecessary to have recourse to 

 unknown Amphibian-like creatures of the Devonian in the search for the 

 immediate ancestors of the jMammalia. Nor does it favor the view that the 

 mammals are descended from streptostylic reptiles or from any reptiles of 

 the subclass Diapsida. There seems, in fact to be a large body of evidence 

 in favor of the view advocated by Osborn and Broom that the mammals have 

 been derived from the Triassic reptilian order Cynodontia, although not 

 from any known member of it. As is well known the Cynodontia foreshadow 

 the mammalia: especially in the constitution of the temporal arch, in the 

 development of the secondary palate and of paired occipital condyles, in the 

 possession of incisors, canines premolars and molars, and in the enlarge- 

 ment and functional importance of the dentary. Perhaps even more sug- 

 gestive of mammalian affinities are the resemblances in the pterygoids, 

 in the phalangeal formula (so far as known) in the humerus, carpus, tarsus, 

 shoulder girdle and pelvis. Structurally if not genetically the Cynodontia 

 certainly bridge over the gap between mammals and reptiles; because in 

 combination with the above mentioned mammalian features they possess 

 many reptilian characters, inherited from the orders below them. For 

 example they retain a bar connecting the jugal with the postorbito-frontal 

 (absent in Baiiria), and the pterygoids have anterior flanges (although lack- 

 ing the i^terygoquadrate portion) ; a quadrate, articular, angular, surangular 

 and splenial are present, and the vertebrtie have intercentra. All these are 

 characters which may reasonably be sought in the remote ancestors of the 

 mammalia. 



As regards the mammalian auditory ossicles, the theory which homolo- 

 gizes the mammalian incus with the reptilian quadrate, the body of the 

 malleus with the articular, rests upon considerable evidence, which has been 

 developed, especially by Reichert, Kingsley and Gaupp. The strength of 

 this theory proceeds from the fact that it is founded upon morphological 

 relations common to all embryonic mammals on the one hand and all modern 

 reptiles on the other. Gaupp's corollary that the Promammals once had 

 two joints between the jaw and the skull was shown to be open to serious 

 objections, but only in so far as it assumed the mammals to have been 



