OTHER THEORIES OF SPECIES-FORMING. 281 



hen with a parrot's beak. And we can declare with con- 

 fidence that a chorda dorsalis can never appear in a beetle. 

 Through the very fact that an animal belongs 



Constitutional 



limitations on to a group the possibilities of variation are 

 variation. distinctly delimited and in many special cases 



these possibilities may indeed be very narrow." Plate 

 does not mention in this connection the fact that some 

 biologists have seen in this restriction of the range of varia- 

 tion which inevitably accompanies specialisation in the 

 development of animal groups an important factor in the 

 determination of lines of descent. Cope gave much import- 

 ance to this factor, and very recently Rosa, 16 in a most in- 

 genious and suggestive paper, attributes to this "progressive 

 reduction of variability'' a large importance in the dying 

 out of old species and the origin of new ones. 



4. "By the correlations which bind each organ to others 

 the range of variation is also restricted." 



5. Many facts of palaeontology seem to prove the existence 

 of orthogenetic evolution. "Wherever a large supply of ma- 



ractsfrom terial permits the working out of a phyletic 

 palaeontology se ries, we always see a limited number of lines 



supporting or- 

 thogenesis, of development, which despite occasional side- 

 branching run essentially in straight lines, in steps which 

 lead gradually one to another." 



6. The phyletic series (chains of forms) of recent species 

 Single phy- a ^ so snow > where we are able to trace them, 



letic lines. distinctive single lines of development. 



Eimer's particular theory of orthogenesis, which we 

 have chosen as a representative of the orthogenetic doc- 

 trine in general (although few biologists who believe in 

 the principle of orthogenesis accept this theory in detail), 

 may be briefly stated as follows : 



Modifications of organisms, that is, lines of evolution, are 

 not miscellaneous, but occur according to control along a 

 few definite directions, these lines of change being deter- 



