'470 



BARON FRANCIS NOPCSA ON THE 



Since the fusion of the metatarsals occurs exckisively in terres- 

 trial forms (^?rtCi!a(?a, etc., and Artiodactyla), Proavis must once 

 have been terrestrial. Supposing Proavis acquired this marked 

 specialisation while living in trees, it must have moved about 

 there in a hitherto unknown fashion ; supposing, however, that 

 birds learnt to fly, and only then became bipedal, then we have a 

 quite unique sort of evolution, for all those quadrupeds that learnt 

 to fly in the quadrupedal state became unable to walk. Conse- 

 quently birds must first have been bipedal, and only then have 

 learnt to glide. 



Another argument that can also be brought forward against 

 the hypothesis of an arboreal Proavis can be deduced from the 

 pelvis of the most primitive bird we know {Archceopteryx) . 



While the pes in Arclioeopierxjx shows that this type had 



Text-figure 6. 



Pubes of Dinosaurs and birds. 

 (A) Ccelurus; (B) Arch<Bopteri/x; {C) Struthio; (D) Dromeeus. 



already taken to arboreal life, the pelvis, as shown by the pubes, 

 is of the long and narrow type. This type is onij^ met with 

 in Dinosaurs and one cursorial bird. In the other birds, and 

 especially in the arboreal birds, the pelvis is usually broad and 

 short, and the pubes are wide apart (text-fig. 6). 



In accordance with this structure and ^^ith the abbreviation of 

 the vertebral column, the arboreal birds mostly hop, while the 

 ground birds walk or run. In the best runners the body is 

 the longest, and the pelvis always remarkably na)-row. Since 

 a long body and narrow pelvis are seen in Archceopteryx in a 

 marked manner, its arboreal specialisation can only be called 

 very slight. It was assuredly derived from a terrestrial stock. 



The shoulder-girdle in birds also indicates that their common 

 aiicestor must have once passed through a terrestrial stage of 



