106 s. IKENO : 



and B were always absolutely linked to each other, we will have naturally 

 no means of discerniug the composite nature of this factor-complex, but some- 

 times the linkage is broken down, at least partially, and we are then enabled 

 to disclose its real nature (s. the Ci-oss YIII, p. 112 fi'.) 



Cross YI. Wiltc-IIx magenta. (PI. IT. fig. 7 and 1). 

 ccBBBB X CCBBBB F, = CcBBBB. 



Though white-II is externally very similar to white-I, it may differ 

 sometimes from the latter in certain respscts. The white-IT bears white 

 petals like the white-I ; but sometimes (not always) they have few magenta 

 stripes or spots ; filaments are white, but often some few ones are magenta (s. 

 p. 120). The i^i-hybrids (1918) bear magenta flowers, whose colour intensity 

 is almost perfectly similar to that of the magenta parent (s. the Table of 

 Colom'S, p. 96). As the white-I and white-II are genotypically different fi-om 

 each other, the composition of the i^i-ofi^pring produced ex white-II x magenta 

 .is quite different fi-om that of those ex white-I X magenta (Ci-oss V). Thus 

 we have the results shown in the Table Y (s. p. 107.) 



In tliis Table we see lx)th in F, and i^, the production of a certain 

 number of unexpected individuals. Thus 8 out of 10 i^i-plants have segregated 

 in F, into magentas and whites, as was just expscted, whereas each of the 

 remaining two has produced besides these two classes of the segregates 1 

 orange which is quite imespected. The formation of these two oranges is 

 very dUBcult to be accounted for, and it might be due to the contamination 

 from other famUies, though utmost care was taken for avoiding such. Till 

 the contrary to the latter assumption will be definitely established these two 

 oranges Anil not be taken into accoimt, and then we see clearly that each of 

 these Fy parents is of the constitution CcBBBB. 



In /'; generation three magentas have produced among others 1 flesh- 

 coloured, 1 orange and 1 pseudo-white (Table Y, Fj, Nos. 2, 3, 4), and two other 

 magentas 3 flesh-coloured and 5 red plants (Table Y, Fj, Nos. 6 and 7). Since 

 the geuotypic constitution of flesh-coloured and pseudo-white races is not yet 

 exactly known it is naturally impossible to make any surmise about the mode 

 of theii" production (though probably by mutation), and these imexpeoted 

 plants must here be left out of account. The production of the orange and 



