336 The Mechanism of Evolution in Lbptinotaksa 



reasons: (1) It is all too commonly assumed that natural species and 

 materials in nature are " pure," by purity being meant uniformity of action 

 and homogeneity of constitution, which in my experience in using materials that 

 are wild in nature is not true in any instance thus far found. There is taxo- 

 nomic purity, or limitation to the range of the characters present, and this is a 

 real, permanent condition in these natural species, but within these limits the 

 population shows all degrees of diversity in its composition and in the factors 

 and agents present, so that for accurate investigation it is necessary that all 

 materials from nature be reduced to homogeneous conditions before they are 

 used in the conduct of refined experiments. Because of the many lines of 

 investigation in different places that these materials have and are being used in, 

 it was necessary that this point be driven home, so that repetition of these 

 experiments or the orientation of others of similar character must have always 

 this point in mind, to either employ the same or identical materials in the 

 operation, or else endless confusion will surely result. (2) The condition 

 determined has considerable value in the insight that it gives into the nature 

 of heterogeneity in nature, of a sort that might well be productive of the initial 

 stages of wide separation of portions of the original stock, and the origin in 

 nature of localized groups that might become the basis of " varieties," " spe- 

 cies," or other taxonomic groupings of whose reality there is not the least doubt, 

 and of which the problems of their origin are well worth careful experimental 

 investigation. 



In the time following the publication of Darwin's work, the existence of 

 groups in the population of this sort were frequently postulated and rather 

 generally believed in, but I think were not accurately demonstrated in any 

 instance. With the rise of the biometricians the effort to recognize and deter- 

 mine these groups was taken up anew, and with what hopefulness is fully indi- 

 cated in the writings of the most eminent of the biometricians — Galton, 

 Pierson, Davenport, Dunker and others — ^but the methods and the tests pro- 

 duced no real progress, save keeping alive the question. The investigations of 

 De Vries and his postulate of " elementary species " presented the old question 

 in a new light, and the pure-line conception of Johannsen added but a new 

 method of analytically isolating some of the possible population conditions Tjpon 

 which divergence and heterogeneity in nature might be based. I believe that 

 the investigation of conditions in the population by the combined methods of 

 biometrics and genetic analysis, as is employed here, may lead directly to the 

 formation in terms of modem conceptions of hypothesis of the methods of species 

 formation that are capable of experimental proof. I shall have occasion to 

 consider these conditions in further experiment and analysis as a basis for the 

 development of conceptions of species formation in nature, without the aid of 

 abrupt and cataclysmic happenings or the long-time accumulation of undeter- 

 minate utilitarian " variations." 



In this analysis of heterogeneity diverse aspects of many problems have been 

 presented and discussed. I have utilized mainly one set of characters, for the 

 reasons given, and combined in this investigation I used statistical methods, the 

 tests by breeding, and modem analyses of any sort that woujd elucidate the 

 problems considered. 



