The Theory of Polymery 343 



resembling the light parent, and 20 will be intermediate. Six 

 of the 64 will be somewhat lighter than the darkest, 15 will be 

 still lighter but not so light as the intermediate type, 15 will be 

 somewhat paler than the intermediates, and 6 will be still paler 

 but not colorless. When both three and two pairs of genes are 

 interacting the Fi is intermediate, but with three pairs there 

 are more classes in the F2 and the percentage of parental types 

 to be expected is smaller. In each case, however, some parental 

 types are expected, but no plants deeper than the dark parent 

 nor paler than the light parent would be recovered in the Fo. 



The Theory of Polymery 



Duplicate, cumulative, nondominant genes might be found 

 which determine other characters than color. In fact, differences 

 in gross size, in weight, in yield per acre, and in many other 

 measurable characteristics in a number of plants and animals 

 have been explained by assuming the presence of a certain 

 number of such genes. Let us set up some hypothetical cases 

 and see how they might be explained on the basis of this scheme. 



Let us assume that two plants differ in height by 24 cm. The 

 smaller is 30 cm tall and the larger 54. Let us assume, also, 

 that the difference in height is produced by four interacting, 

 cumulative, duplicate, nondominant genes, and that each con- 

 tributing gene adds 6 cm to the height of the plant. The geno- 

 type of the smallest parent would be titi toto XX, where XX 

 represents 30 cm common to both parents, whereas the larger 

 parent would be TiTi T2T2 XX. This parent would have the 

 30 cm contributed by A"A^ and 6 cm for each of the four con- 

 tributing genes, and would be 54 cm tall. The Fi would be 

 Titi T2t2 XX, would have only two contributing genes in addi- 

 tion to the residual heredity indicated by XX, and would be 42 

 cm tall, exactly intermediate between the two parents. The F2 

 would segregate as in Fig. 96a. More F2 plants would be inter- 

 mediate than would be found in any other class, the parental 

 types would be the least common, and no plants would be found 

 more extreme than either parent. Since all the Fi plants are 

 genotypically identical, different Fo families from selfing dif- 

 ferent Fi plants should theoretically all be alike. If there is 

 any difference among them it should be due to the inability to 

 grow all the plants of a family that are theoretically possible 



