72 DARWINISM TO-DAY. 



ence seems, therefore, to teach us that selection of fluctuating 

 variations leads us to only a certain point, and then stops in 

 this direction. We get no evidence from the facts in favour 

 of the view that the process, if carried on for a long time, 

 could ever produce such great changes, or the kind of 

 changes, as those seen in wild animals and plants." 



There is something inherent in the make-up of the organ- 

 ism and something inevitably incident to the phenomena of 

 variation which prohibit, even in the most favourable cases, 

 the indefinite movement of variation. Johannsen, 6 pro- 

 fessor of plant physiology in the University of Copenhagen, 

 finds that beans bred in pure lines, /. e., not crossed, conform 

 perfectly with Galton's law of regression. And Johannsen 

 holds that this regression must be a serious brake on the 

 species-modifying, i. e., species-forming, activity of natural 

 selection. That is, while the species mode can be moved in 

 one direction or another by pure line breeding it can be so 

 moved only very slowly. And this same law of regression 

 will tend to break up a "mixed population" resulting from 

 crossed and miscellaneous breeding into distinct pure lines ; 

 that is, each independent form-type tends to be constant, not 

 constantly moving, i. e., transforming. New types must 

 arise chiefly then through (a) the crossing of races or 

 species (= hybridisation) or (b) through mutations. 



Delage 7 in his criticism of selection makes the point that 



because the causes of variation are more feeble than the 



P. . , . . causes of fixity (as evidenced by the massing of 



JJ 6.1. 1^6 S Critil" 



cism of Delboeuf s variations around and close to the mean or mode, 

 and their increasing scarcity as they recede in 

 any direction from the mode (species type)), the species 

 tends always to stand still rather than to change. If in the 

 first generation a thousandth of the individuals vary in the 

 same way, in the second generation only i-iooo of the 

 thousandth part will show the same variation, reasons 

 Delage. But, as pointed out in chapter vi ("Darwinism De- 



