IXTRODUCTIOX. 31 



theoretically possible to extract different types which breed 

 true from it. In practice, however, errors of measurement and 

 observation or environmental influence may be too great for such 

 types and their heterozygotes to be distinguished. From these 

 considerations as well as the experimental evidence of his own 

 researches, Nilsson-Ehle concludes that there is no inherent 

 difference in the mode of inheritance between quantitative and 

 qualitative characters and that all variations may be placed 

 in two groups : — 



1. Variations which are inherited. 



2. Variations which are not inherited and which are prob- 

 ably entirely due to changes in the environment. 



It is obvious that the larger the number of factors in 

 which the parents differ and the greater the effect of environ- 

 ment, the more difficult it becomes to separate the factors or to 

 determine the exact mode of inheritance. If we consider the 

 case of two parents, which differ from one another in three 

 factors, there will be in the F2 generation eight homozygotic 

 combinations and nineteen heterozygotic, Avhich, in general, 

 aWII lie between the homozygotic forms. Thus we obtain a series 

 containing twenty-seven stages betAveen the two parent, forms. 

 If the original difference between the parents is not very large, 

 these forms will lie very close to one another. If, in addition, 

 environmental differences supervene, the limits of variation of 

 one form will very soon overlap those of the next or even of 

 several others, in fact, in many cases, the limits of the tAvo 

 parents themselves overlap. 



Taking these facts into consideration, it is not surprising 

 that up to the present in no case has the inheritance of the size 

 of an organ been entirely elucidated and the various factors 

 determined. All that has been possible has been to show that 

 segregation undoubtedly occurs and that the facts are in accor- 

 dance with the Mendelian interpretation and with the existence 

 of many factors, all capable of being inherited independently. 

 In the investigations described in this paper it has been possible 



