24 APPLICATION OF PKINCIPLES OF HEREDITY TO BREEDING. 



a plant which is completely homozygote in all its characters, then it 

 will transmit the same form of every character to all its offspring, and 

 we have eliminated all variations due to recombination of characters. 

 There will still be differences between the plants grown from the same 

 seed, but these differences will be due to environmental influences, 

 such as differences in available moisture, plant food, sunlight, and the 

 like. It is highly important to make this distinction between indi- 

 vidual variations which are due wholly to environment and those 

 which are due to recombinations of hereditary characters. It will be 

 seen later that so far as experimental evidence goes there is much 

 reason to believe that selection of those fluctuations which are due 

 wholly to environment as a rule has no effect whatever in changing 

 the hereditary characters of the plant. On the other hand, in those 

 plants which are not homozygote in all their characters, as is. practi- 

 cally always the case in a species that regularly cross-fertilizes, there 

 will be variations due to recombinations of different characters, and 

 selection will have a marked effect in species of this kind. The effect 

 of selection on fluctuating variations has been much confused because 

 of the effect produced by mass selection in mixed populations of fixed 

 forms like wheat, which effect will be further discussed later. In 

 wheat, and other self-fertilized species, individual selection^that is, 

 selection in which we keep the progeny of each mother plant sepa- 

 rate — soon proves that we can not modify these fixed strains by 

 selection; that is, generally speaking. The facts have further been 

 confused because of the effect which either mass or individual selec- 

 tion has in gradually changing the character of cross-fertilized crops 

 like corn. In these cross-fertilized crops either mass selection or 

 selection annually to a single mother plant causes a gradual change 

 in the direction of the selection. But when we eliminate the effect 

 of the law of recombination, which occurs continually in cross-fer- 

 tilized forms, and practice selection annually to a single mother 

 plant, we find that it is apparently impossible to produce modifica- 

 tion by selection, except in rare instances. The investigations on 

 which this reasoning is based will be given later in these pages. 



The amount of investigation which this subject has received can 

 hardly be said to be sufficient to settle it for all cases, for there are a 

 few exceptional cases which do not behave in the usual way and 

 which are not understood. In the main, however, the investigations 

 all agree. The first work bearing strictly on the effect of selection 

 on forms from which all variation other than fluctuating variation 

 due to environment has been eliminated was done by Prof. W. 

 Johannsen, of Copenhagen, on beans and barley. This work will 

 be referred to more in detail when we come to consider the effect of 

 selection on self-fertilized species, as will also the remarkably accu- 



165 



