yiS THE ORIGIN OF SPECIES 



species suggests that there are no real Hmits ; the less common 

 difficulty of deciding on the hmits of an order or a class, as with 

 Tritylodon mentioned above (p. 713), points to the same conclusion. 

 Perhaps more important is the fact that classifications based on 

 superficial examination aie often contradicted by a knowledge 



Fig. 552. — Diagrams to show the difference in construction between the wing of 

 a bat and that of a bird. In the bat all the digits are present, four of them 

 are very long, and the wing surface is provided by skin stretched from linger 

 to finger and between the arm and the flank. In the bird there are only three 

 digits, these are vestigial, and the wing surface is mainly provided by feathers 

 in rows parallel with the axis of the limb. — From Romanes. 



of internal structure. Whales, for instance, were for long regarded 

 as fishes, but dissection shows them to be mammals ; even more 

 obviously, bats are not birds, as their wing skeleton is different 

 and easily derived from that of mammals (Fig. 552). The theory 

 of evolution has, however, had more influence on classification 

 than has classification on evolution, since most taxonomists 

 now attempt to make their systems show not only similarity 

 in structure but also community of descent. This is, however, 

 not always possible and is always of necessity conjectural, so 



