432 Bulletin American Museum of Natural History. [Vol. XXVII, 



in new positions take on the functions and appearance of the vertebrae whose 

 2:)Osition they have usurped. 



(5) The perforation of the transverse process of the seventh cervical 

 by the vertebral artery in IMarsupials. This very rarely (e. g., Lepus) 

 happens in Placentals. 



(6) The large size of the proximal caudal vertebrae and of the root of 

 the tail in many primitive mammals. 



(7) The preservation of lumbar parapophyses in arboreal and cursorial 

 forms and the loss of them in fossorial forms. 



The " preaxial" and " postaxial" horders of mammalian limbs. 



The conception of what constitutes the "preaxial" and "postaxial" 

 borders of the parts of the limbs, as worked out by Huxley and Flower (1885, 

 pp. 361-373) appears in certain respects to be artificial and incorrect. It was 

 founded in part on erroneous notions of the i)hylogenetic relations of the 

 Chelonia, "Enaliosauria" and Cetacea to the lower Mammalia. Flower's 

 diagrams imply that the ancestral mammals had slender limbs resembling 

 the highly modified type realized in the Primates. His theory starts w^ith the 

 limbs "extended at right angles to the axis of the trunk" in a horizontal 

 plane. It assumes that the limbs were "parallel to each other" and more 

 or less completely homodynamous. Concurrent evidence from many lines 

 shows that these assumptions are both unnecessary and incorrect. In regard 

 to the ultimate derivation of the primitive cheiropterygium from the })tery- 

 gium of unknown Crossopterygians or Proto-Amphibians it is not necessary 

 at jiresent to make any assumptions. Neither do we know the actual genetic 

 series which led from Carboniferous Amphibians to early Permian Cotylo- 

 saurs and thence from late Permian Therocephalians to early Triassic 

 Cynodonts, Triassic Monotremes and Metatheria and perhaps Jurassic and 

 Cretaceous Eutheria. But so many collateral offshoots of the direct series, 

 representing different stages in the evolution of the limbs are known that 

 the interpretation of "pre-" and " i)ost-axial " above referred to requires 

 considerable modification. 



The subject may be introduced conveniently by a consideration of the 

 shoulder girdle in primitive Placentals. The shoulder girdle in certain 

 embryonic and adult Rodents and Insectivores retains in a more or less 

 reduced condition certain bilaterally paired structures called by Gegenbaur 

 "prseclavia (cf. Weber, 1904, p. 94), which are doubtfully homologized by 

 some authors with the procoracoids of Monotremes. In the adult Micro- 

 gale (p. 237), as recorded by Leche, these structures appear as two large 

 plates which dorsally overlap the sternum in the middle line. This, joined 



