236 NATURAL SCIENCE [April 



variety was definite, only in varying degree, the more or less 

 unadaptable being increasingly rigorously weeded out in each succeed- 

 ing generation, and this would continue to be so as long as only a 

 very gradual migration or no migration at all recurred, provided 

 conditions remained more or less the same. Directly, however, any 

 sudden change occurred, the balance of adaptable variations would 

 have to change, and indefinite variations would become once more 

 apparent. Now this change might perfectly explain the fact that 

 variations are more definite in nature than under domestication, 

 because changes of environment are less extreme in nature than when 

 under the selective power of man, also this very change in the 

 balance of variations might itself be the cause of atavism, etc., 

 which has frequently been noticed under some of these extreme 

 changes in environment or when crosses between allied species have 

 occurred. 



Thus, so far from definite variations being a difficulty in the 

 acceptance of the theory of Natural Selection, it would be precisely 

 what on a ^priori grounds would be expected. 



Finally, there are those classes of variations which are said to 

 be directly modified in response to environment and in which the 

 variations have possibly been inherited after the third or fourth 

 generation of exposure to the modifying cause. To cite one instance, 

 the fact of " chlorophyllous tissue being much more developed in sun- 

 light than in shade." i 



It is not enough to show that sunlight causes an increased 

 amount of chlorophyll to be formed, for the fact that increased 

 function brings about a compensatory increase in the tissue so 

 exercised is admitted by all biologists : it is not enough even to 

 demonstrate that this increase is detectable when the variety thus 

 formed is again bred under the normal amount of sunlight. It must 

 further be shown : (1) How many seeds or cuttings of the parti- 

 cular plant were set ; (2) How many of those that were set died ; 

 (3) What was the result of a control experiment in which the same 

 conditions were observed in everything but the excessive light ; (4) 

 That in all the experiments the cuttings were so situated that the 

 growth of each was, as far as possible, entirely free from interference 

 by the others ; (5) That the chlorophyll does not develop more 

 quickly, when, with otherwise similar conditions, rigorous Natural 

 Selection is allowed to occur ; (6) It must also be shown, as far as 

 possible, that no immediate ancestor of the selected variety experi- 

 mented upon has had a greater supply of chlorophyll, and conse- 

 quently that the action that the sunlight had apparently induced 

 was not in reality only a favourable reversion. 



In short, to disprove the action of Natural Selection, it is neces- 



1 P. 72, "Origin of Plant Strnctures." 



