Analysis of Heterogeneity in Some Simplest Chakactees 229 



in any way changed or even deformed by the removal from the organisms of the 

 supposed forces productive of the range in the character. These results I have 

 shown in concentrated form in tables 30 and 31. 



Table 31. 



No. of mating. 



Parents, 

 index. 



Progeny. 



Fraternity No. 



Range of index. 



Totals. 



M. 2. g II 



M. 14. g II.... 

 M. 16. g II.... 

 M. 2. g III.... 

 M. 14. g III... 

 M. 14. g III... 



r2.761 



2.894 

 ^2.668 



2.708 

 2.692 

 2.923 



2.778 

 2.810 

 2.746 



2.751 

 2.863 

 2.641 



f2.776 



2.781 

 12.894 



f2.777 



2.731 



12.618 



2.750^ 

 2.692 > 

 2. 681 J 



2.750") 

 2.774 y 

 2.661J 



2.764T 

 2.689 y 

 2. 891 J 



2.8111 



2.751 y 



2.554J 



2.846"! 

 2.651 y 

 2. 758 J 



2.58n 

 2.774 Y 

 2. 841 J 



M. 2. g III.. 

 M. 14. g III.. 

 M. 16. g III. 

 M. 2. g IV... 

 M. 14. g IV.. 

 M. 16. g IV.. 



fM. 



If. 



{?: 



fM. 



If. 



fM. 



If. 



fM. 



{f. 



16 



177 



I 



251 67 

 421 



29 



Total. 



All of the stocks went into hibernation in June and remained quiet until 

 October of the same year. I then decided that the series was not worth the space 

 and attention that it required, and therefore drew lots for three to be continued 

 for two more generations in the autumn of 1908. The choice fell to pairs 

 8 g IV (AA), 10 g IV (BA), and mass culture 2 g IV, and these were continued 

 for two generations more, completing the series early in January 1909. The 

 result was in no wise changed, as is shown in table 32. 



In the literature, ranges in characters of the kind used are commonly 

 attributed to the action of incident environmental conditions of temperature, 

 moisture and the like, and to conditions of nutrition, all of which were con- 

 trolled in the series, and were uniform and optimum, and should have given 

 uniformity and decreased ranges in the attribute under consideration if the 

 differences were due to environment as is commonly asserted. The entire series 

 in this direction was killed off at this point, nothing of any value being expected 

 from its continuation. 



I do not know of any other tests of this question that have been made with 

 the same refinement, or even any tests of any sort. If I am not entirely mis- 

 taken, the instances of range in fluctuations of characters have been of reason- 

 ing from effect to cause, and instances where observed ranges were not other- 

 vrise accounted for have been attributed to this " environmental action," which 



