918 DR. R. BROOM OX THE 



The presence of this large number of mammalian characters in 

 the Cynodont skull, and the absence of any in either skull or 

 skeleton that might not have been expected in the mammalian 

 ancestor, make the case very strong for the mammalian ancestor 

 havino' been a Cynodont. But the evidence becomes even stronger 

 when we find that most of the peculiai-ities, even minor 

 peculiarities, of the mammalian skull have light thrown on them 

 by the condition of affairs in the Cynodont skull. 



Let us consider some of the more remarkable characters of the 

 mammalian skull in the light of our knowledge of the Cynodont. 



Premaxillarij. 



One of the most striking peculiarities of the mammalian skull 

 is that the nostrils are sepai'ated only by cartilage, so that if the 

 cartilage be removed the nostrils are united. In most reptiles, in 

 birds and amphibians the nostrils are divided by an upward and 

 backward process of the premaxilla, the internasal process. As it 

 is present in Cotylosaurs, Dromasaurs, Pelycosaurs, Therocepha- 

 lians, Anomodonts, and even Cynodonts, one might fancy that 

 here was evidence against the Cynodont ancestry. But thei'e is 

 o'ood reason to believe that the early Mammals retained the 

 internasal process and that it was only lost after the Mammals 

 were well established. 



In both Ornithorhynchus and Echidna the young animal has an 

 internasal process developed on the premaxilla almost exactly as 

 in reptiles. The fact that it is retained as a support to the 

 caruncle or egg-tooth in no way invalidates the conclusion that it 

 is the reptilian internasal process that has been retained. For 

 there cannot have been a time when there was a caruncle without 

 a support, and thus the internasal process must be as old as the 

 caruncle. As we may be pretty certain that the mammalian 

 ancestor was oviparous, we may safely conclude that the internasal 

 process is not a neomorph, but the reptilian structure handed on. 

 In Tritylodon there is an imperfect but distinct little internasal 

 process. The only known specimen is too imperfect to enable me 

 to say whether it formed a complete though slender process which 

 joined with the nasals. Even if it did not in the adult, it is 

 rather probable that it did in the very young animal, since 

 Tritylodon is so much more primitive than the marsupial that not 

 improbably it was oviparous. 



In the skulls of young Diprotodonts (e. g. Mao'ojms) a rudiment 

 of the internasal process is usually present. And in the young 

 TricJiosurns at birth the internasal process, as I recently pointed 

 out can be traced right round in front of the nose. In the very 

 young marsupial, the nostrils are entii'ely lateral and wide apart, 

 and the nasal cartilages pass round in front of each, leaving a 

 sulcus in the middle line between the two. The premaxillaries 

 send up short processes along the sulcus, but from the ends of the 

 processes two strands of condensed but unossified cells can be easily 



