230 The Mechanism of Evolution in Leptinotaksa 



sorts are presented which can not be separated into rigorously limited classes of 

 quantitative and qualitative " variations." In no instance in operations of this 

 kind have the w^orkers been dealing with simplest characters, or even with unit- 

 characters in the sense of De Vries, but usually with an undetermined complex, 

 capable of subtending some mathematical value, and the mathematical value has 

 been regarded as a true representation of the character in question. It is largely 

 this sort of thing that has led to the establishment of the idea of quantitative 

 variations by De Vries as distinct from qualitative. 



The fact that the continued " selection " of these characters does not result 

 in the production of unlimited divergence of the character, which is a diflficulty 

 in the minds of many, is not really an element in the present problem. It is 

 quite distinct, and it may well be asked whether any kind of character is 

 capable of unlimited " improvement " by any method of transmutation whatso- 

 ever. It seems that the very nature of natural substance would indicate that 

 there are limits to combinations, to arrangements, and to operations, and the 

 limitation of " selective action " by bounds that are not passable is not surprising. 



The " difficulty " that has entered so largely into the discussion of the demerits 

 of fluctuations as an effective element in the modification of organisms, " that the 

 products of fluctuations can not be fixed in position," does not seem at present to 

 offer any difficulty. Fixity in position is simply stability of composition, 

 method of reaction, or both, and homogeneity of organization ; and it is precisely 

 this that the selectionists have not taken into their considerations and operations. 

 No effort has been made to obtain materials or characters that were single and 

 pure, and a heterogeneous mess extractive like saccharine content, that can be 

 measured in mathematical values, is the limit of their ability to analyze either 

 method or product or character. I shall have occasion to deal with this question 

 further when I come to consider some of the experiments in " selective improve- 

 ment," but the main point that I wish to make at this place is the uncritical use 

 of the biometrical efforts in the last two decades, as an example of an inefficient 

 type of investigating the origin of heterogeneity, in which the differences were 

 assumed to be mere differences in the amount of something present. Without 

 reference to the correctness of some of the conclusions regarding the lack of effec- 

 tive transmutation produced by these differences, and all of the consequences 

 that follow therefrom, it is becoming increasingly clear that diversity is not one 

 of amount present, but that the origin and nature of diversity is best stated as 

 integrated products of associations of productive agencies whose presence is nec- 

 essary for the production of the result and in whose array any break results in 

 nonproduction of the usual manifestation. In some instances it might be pos- 

 sible to express this integration in single terms of measure, but these can 

 never express the really complex nature of the result, in which there are 

 always present relations of position in space, interaction with adjacent materials, 

 the intricate series of antecedent series of events that preceded the end-result, 

 with the constant play of many incident agencies from within and without the 

 mass of the system, and to regard this complex series as a mere quantitative 

 swing of the amount present and as expressing in measure the amount of units 

 subtended on the biometricians' meter-stick, is to blind ourselves to some of the 

 most obvious facts of nature, and to indulge in the entirely useless and false cre- 

 ation of categories and definitions that are not in any way useful in investigation 

 or expressive of certainly determined facts of nature. 



