146 MAINE AGRICULTURAL EXPERIMENT STATION. I9II. 



The nature of Johannsen's results on this point have been 

 very cleverly set forth in the accompanying diagram w^hich is 

 taken from a paper by Wood and Punnett.* 



2. That a population of beans, no matter from how sup- 

 posedly "pure" a commercial variety it is taken, is really not a 

 homogeneous unitary aggregation, but instead is made up of a 

 varying number of lines or strains, each of which breeds true 

 to itself when propagated in isolation. In other words the popu- 



FiG. 78. Diagram to illustrate Johannsen's results with beans. (From 

 Wood and Punnett). 



lation in question is a mixture of several component lines. The 

 individuals in each line produce offspring true to the type of 

 the line, rather than to the type of the population as a whole, 

 except in cases where by chance the population type and the type 

 of one or more lines happen to be the same. 



3. That when mass selection alters the population type it does 

 so by a process of isolating from the mixture certain strains 

 whose own types are different from the original general popula- 

 tion type, and which differ in the direction towards which selec- 

 tion was made. Thus if one begins in a general mixed popula- 

 tion of beans to select for planting the largest beans, and by 

 so doing increases the average size of bean in the crop, what he 

 really does is gradually to throw away all beans except those 



*Wood, T. B., and Punnett, R. C, Heredity in Plants and Animals. 

 Mendel's Principles and their Bearing on Agricultural Problems. Trans. 

 Highland and Agr. Soc. Scotland. Ser. 5, Vol. XX. pp. 36-86, igo8. 



