108 REDUCTION OF VARIABILITY. 



It is rather obvious that it is not selection itself which causes 

 purity, but some process which accompanies selection. There- 

 fore, if we discover what there is in selection to cause purity, 

 decrease of variability, we can pass on to the question whether 

 this process acts independently of selection, and therefore 

 independently of usefulness. 



It is here that the proposed term Total Potential Variability 

 comes in useful. We have used this term to express the number 

 of genes for which a group of organisms, or one single individ- 

 ual is not pure. 



The simplest instance of a group of organisms exhibiting a 

 small potential variability is the case of two or three plants 

 which all have the same set of genes, which we will together 

 call X, with the exception that each has an additional one, 

 which is lacking from the others. The constitution of the three 

 will therefore be X — A, X — B and X— C. To simplify matters, 

 let us suppose that there is no crossing; we are dealing with 

 self-fertilized plants, or even with asexually propagated plants. 

 The total potential variability in our group of three plants is 

 three. If we multiply these three plants, the potential variabil- 

 ity remains three, that is, it remains three just as long as in 

 every generation at least one plant of each type is reproduced. 

 And if for any reason any one of the three types, X — A, X — B 

 orX — C dies out, or is not included in a new group, a new colony, 

 that new group or new generation will have a potential 

 variability which has dropped from three to two. Any new 

 colony which originates from our mixture, out of a few seeds or 

 a few tubers, will have the original potential variability if all 

 the three types are included. But if they are not, it is lowered 

 to two, or one, or zero. Any little group starting from seeds 

 or tubers of one individual will in our example be without any 

 potential variability, it will be in every respect pure. And 

 right here the real nature of selection as a factor in purity 

 comes to light. If from a mixture of three pure lines we select 

 one plant with desirable characters, the resulting group will be 

 wholly without genetic variability. But it is clear, that it is not 



