168 



MUTATION. 



produced this aberration. Now we know what breeding true 

 is worth in this material as a criterion for geno-typic purity, but 

 even in other material, the sudden production of a novelty 

 after four generations, comprising only very few plants in 

 every generation could hardly be called mutation without 

 breeding tests. 



We know, that in several instances the difference between 

 _ _. individuals having altern- 



Ab £ ABAB At *B 

 "0 







Fig. 20. 



ative characters is due 

 to the presence or absence 

 of not one definite gene, 

 but of at least one of a set 

 of two or mo re genes which 

 have a similar action on 

 the development. In our 

 comparison of the chain 

 of processes leading to 

 the development of any 

 quality of an organism to 

 a material chain of metal 

 links, we can imagine how 

 in some place two links 

 are not held together by 

 one single link, but by two 

 or three or more links, all 

 passing through the same 

 upper link and through 

 the same lower one, lying 

 side 037 side. Ordinarily, 



The effect of "duplicate" genes genes h u k b k th 



which affect the same stage in a chain of \ 



processes in a similar way, and are there- chain breaks at that point, 

 ^e.m respect to the end-result, of equal Jf m any place Qn the 



Only when both A and B are missing chain however a dupli- 

 the following links are unsupported. cate ^ j g put in a Knk 



can be taken out without interrupting the chain. (Fig. 20). 

 The chain remains intact until at this point both links are 



