1Ö8 



S. IKENO ; 



aud the remaining two of that CCBRBB. Of 6 whites 2 (Table V, F„ No3. 

 2 and 3) have produced besides white offspring 7 lutigenta ones. That each 

 of thes3 two white parents is to be regarded to have tha constitution ccJRRBB 

 will be discussed Later in this paper (s. Table VIII, Nos. 21 and 22, discussions 

 in " Mutations, etc." I). 



From the results in F., above indicated we see that we have in F., 4 

 homo- aud G heterozygous magentas (expscted, .3-3 aud 6-(j), aud that all 

 whites are homozygous, whence we may conclude that the actual and the 

 expected results agi-ee fairly well to each other. Hence it is evident that this 

 cross is based upon one single factor difference. We can also easily under- 

 stand the reason why this cross will produce quite different results fi-om those 

 in the Cross V where white-I is used instead^ of white-II, because the latter 

 contains both R and B like our magenta variety, wliile white-I has none. 



Cross Vn. White-II y. orange and vice versa. (PI. IF, fig. 7 aud 3), 

 ccRRBB X CCfibb F, = CcRrBh 



The cross between white-I and orange has been described before (s. p. 98 

 ff;. That white-II is genotypically different from white-I in spite of their 

 external resemblance is especially clear, wJien we make the cross between the 

 former and the orange, because we get then quite different results : the i^-hybrids, 

 wliether from white-II x orange or its reciprocal, bear always magenta flowfi-s, 

 whose colom* intensity is almost equal to that in our magenta variety (s. the 

 Table of Colom-s, p. 96). As indeed white-II (ceRRBBi may be considered 

 to be a magenta variety which remains colaurless ou account of the absence 

 of C, it is quite natural that its mating with the orange ■wiU produce the 

 magenta, C being introduced fi-om the latter parent. The geuotj'pic constitu- 

 tion of the -Pi-hjbrid in the present case is therefore perfectly equal to that 

 in the Cross YI (p. 106», sa that the compasition of the F,- and -F.-oflspring 

 should be naturally quite th^' same as in those derived from the same cross. 

 This fact could be perfectly confii-med experimentally, as we will see from 

 the F. and F^ oflspring presented in the Table VI, A aud B. 



