Analysis of Heteeogeneity in the Population 325 



1906 as to the permanence of this sort of " variation " when determined in this 

 manner. Attempts since 1906 have been as positive failures in this respect as 

 were the earlier ones. 



Biometrically the assumption is made that the material is homogeneous; in 

 reality it is not so, and eliminating by proper methods the larger portion of the 

 somatic variations, the proposition is still true that any instance of statistically 

 determined amount is not permanent and does not become fixed by continued 

 selection, due to the fact that the statistical determination does not choose 

 identical conditions in the character. No better demonstration of this could 

 be asked for than that presented in the history and improvement of the sugar 

 beet. The area of the pigment exposed on the pronota of my material may well 

 be equal in area or proportion of surface exposed, but the pattern conditions 

 presented, the array of active agents, factors, and determiners present and that 

 are brought into this array gives so extensive opportunity for heterogeneous con- 

 ditions that the real nature of the series is constantly shifting and is hopelessly 

 confused, so that the result statistically is to end with the " conclusion " that 

 the " variations " are somatic and so not inheritable. They may be so or not ; 

 the method has not been able to test the point. I have seen series in which the 

 conditions present were known to be accurately inherited and recognizable by 

 the arrangements in the pattern, but the biometric method resulted in a hopeless 

 array of ranging areas of pigment, which could not be fixed. It is this recogni- 

 tion of the results of the confusion of real conditions under the blanket of bio- 

 metrics that formed the transition from the former situation to the present one. 

 The results attained in this first set are correct as far as they can be with the 

 method employed, and were carried as far as it was possible to go in nature and 

 in the experimental analysis of the problem of place variation. 



As has been shown in the preceding pages, this pattern is not a haphazard 

 array of colored areas, but an accurately constructed system, in which the color 

 is the end-result of many agents, and from this point of view quite different 

 conditions are conceived of in the population as the cause of the heterogeneity 

 found, and so the phenomena of place and of geographic variation take on a new 

 aspect, becoming capable of analysis along new and profitable lines. This is not 

 the place to deal with the gametic constitution of the materials used or of 

 the factors that have been recognized as present and acting in the population in 

 this complex system. 



It is from this point of view that the phenomena of place and geographical 

 " variation " deserve analysis, so that the result will be a determination in 

 terms of effective agents, internal and external, in the production of the results 

 observed. Further, the examination will answer accurately questions as to the 

 actual, permanent, germinal differences that may or may not exist in the sepa- 

 rate portions of a widely distributed race, and in experiment these differences 

 can be tied to combinations of environmental conditions that exist in nature, so 

 that in the aggregate a more complete understanding and picture of the condi- 

 tions in nature will be produced. Out of investigations of this sort must come 

 the rational explanation of the origin and meaning of the facts presented by 

 closely allied, locally placed isolated races, as for example the Achatinellidae in 

 the Hawaiian Islands, as described by Gulick, and many other similar instances 

 in nature. The explanation of the origin and the experimental duplication of 



