3<J2 ANNALS NEW YORK ACADEMY OF SCIENCES 



though the physical environiiiGiit were similar. If tiie rising of the 

 Mexican tableland conditioned the dispersal of the genus from that 

 center, we can see in this different biotic environment the reason why 

 the marginal species in North America should be primitive, while the 

 marginal species in South America are highly specialized. In general, 

 it would be true that the species of the dispersal center (or those nearest 

 to it, where, as in this case, it has become ill adapted for the habitat of the 

 race) will ho the most progressive and those of tlie marginal areas nearest 

 the primitive stock. But where the scattering primitive forms, in fol- 

 lowing the primitive climatic conditions, arc brought into a new floral 

 and faunal environment, this may profoundly modify tliom and cause a 

 rapid divergence and specialization. 



Dispersal of Birds 



As a class, birds are extremely difficult in their taxonomy. They are 

 held closely to type in comparison with mammals, and the differences 

 between them are mostly directly and obviously due to adaptation. 

 Adaptive parallelism obscures the true affinities to such an extent that 

 even at the present day the major classification is somewhat uncertain. 

 This difficulty is the greater on account of their rarity as fossils. There 

 is no reason to interpret this rarity as indicating any lack of abundance 

 of birds in the faunae of Tertiary and later Mesozoic time ; it is presum- 

 ably to be accounted for by their generally upland habit, small size and 

 the lightness and fragility of the skeleton. The small minority of fossil 

 birds which are known from anything more than a few fragments are, 

 with two or three notable exceptions, aberrant types — ground-birds, 

 marine or lacustrine types, whose habitat facilitated their preservation 

 as fossils. By far the most notable and instructive of these exceptions 

 is Archceopteryx. 



It has been customary to class the greater number of the ground-birds 

 (RatitEe) as a more primitive sub-class. On a priori grounds, this may 

 be correct enough, since it would appear theoretically that feathers must 

 have preceded flight, the ability to fly being conditioned by high organ- 

 ization plus small size, and this would involve a rapid circulation and 

 high temperature, which could hardly be attained without a nonconduct- 

 ing coating over the body. But it appears certain that most, and possible 

 that all of the existing ground-birds are rcadaptations to terrestrial 

 habitat from flying ancestors, and their resemblances are due almost 

 wholly to adaptive parallelism. 



Owing to their powers of flight, the dispersal of birds is much less 



