Compound Characters by Cross-breeding. 51 



more fully studied, we can scarcely appreciate the relation of these 

 methods to each other. In one instance, however, that of de Vries' 

 Antirrhinum, the evidence as to this generation (F 3 of my nota- 

 tion) is more complete, and we can attempt to realise what has 

 taken place in that case. 



In most of the instances hitherto studied the compound 

 character is a colour-character 1 , which is resolved after a cross 

 with a white variety, and such an example is that of de Vries' 

 Antirrhinum, which I propose to consider here. In this case a red 

 Antirrhinum was crossed with a white, giving a first cross (genera- 

 tion Fi) red like the red parent, white being recessive. F 1 on 

 self-fertilisation gave a somewhat miscellaneous offspring (genera- 

 tion F 2 ), which could nevertheless be classified with fair accuracy 

 into four forms, which de Vries calls Red, Flesh colour, Delila 2 , 

 White. These four forms existed in numbers approaching the 

 ratio 9 : 3 : 3 : 1, if the minor differences are neglected. 



This ratio 9:3:3:1 suggests to de Vries that there is here 

 a phenomenon of " Di-hybridisation," a term used to denote the 

 case of a cross between parents differing in two pairs of allelo- 

 morphic characters, as, for example, between a yellow round pea 

 and a green wrinkled pea. In this latter case, if we call the 

 heterozygote (F ± ) YRGW, then, as we know, the offspring of self- 

 fertilisation will be in appearance 



9 YR:SGR:3YW:1 GW, 



and in reality 



The 9 YR =1 YRYR + 2 YRYW + 2 YRGR + 4, YRGW 

 The 3 GR = 1 GRGR + 2 GRGW 

 The 3 YW = 1 YWYW + 2 YWGW 

 Thel GW = 1 GWGW. 



In interpreting his case of resolution on the analogy of di- 

 hybridisation, de Vries regards the red parent as consisting of 

 a union of flesh colour, F, and Delila, D, corresponding to the two 

 dominant characters, yellow and round, while he conceives the 

 white parent as representing the two recessive characters, which 

 he calls W and W', though it is not suggested that W differs 

 from W. 



The first cross he represents as 



White x Red 

 W + W'xF+D 

 giving W + W ' + F + D, the red hybrid. 



1 Indications however are not wanting that similar phenomena of resolution 

 may occur in regard to structural characters also. 



2 Pale tube with red lips. 



4—2 



