156 Half-hoariness hi Matthiola 



(4) That when the JK pair is combined with H, the colour pair 

 CR being also present, the full degree of hoariness due to the combina- 

 tion CRHK masks the presence of ./. 



Whether the combination JK produces the half-hoary condition in 

 all cases when H is absent, i.e. whether ./, unlike H, reacts with 

 K quite independently of C and R, has yet to be determined. The 

 half-hoary type used in the present experiments being azure in colour 

 obviously contained CR. The form originally used in the earliest 

 experiments had indeed white flowers, but the absence of colour must 

 be attributed to inhibition not to the absence of either C or R. For 

 this half-hoary white gave all hoary in F^ when crossed with either 

 GRH or RHK whites and creams. Hence we must conclude that it 

 contained both C and R and belonged to the class of inhibition whites 

 referred to above (p. 147). Thus in neither case can we tell whether ./ 

 would have been effective if either G or R had been lacking. Until 

 a half-hoary white has been found or made which can be shown to 

 belong to the class of defective whites this question must be left 

 undecided. So far such a small number of extracted half-hoary plants 

 have been flowered in matings which could yield the required com- 

 binations that the fact that those that were recorded were all coloured 

 leaves the point still in doubt. Where a half-hoary tyjae is crossed 

 with a glabrous type, whether coloured or not, which lacks both H and 

 K, we may however venture to predict that the JK pair brought in by 

 the one parent will be effective and that the F^ will be all half-hoary. 

 With a half-hoary type therefore we are not able to predict as with a 

 fully hoary type the result in F, when the mating is with a glabrous 

 type of unknown composition. The fully hoary gives akuai/s fully hoary 

 in F^ whatever the composition of the glabrous parent as regards the 

 factors CRHJK. The half-hoary, whether coloured or a white of the 

 inhibition class, gives (1) fully hoary, (2) half-hoary or (3) quarter- 

 hoary according as the glabrous type (1) contains H either with or 

 without / and K, (2) lacks all the three factors H, J, K, or (3) contains 

 K but not H or J. To these three cases we may even need to add 

 a fourth. For if the glabrous parent contain J without either H or 

 K, it is quite probable that the F^ will be a form (three-quarter- 

 hoary) recognizably distinct from both the half-hoary and the fully 

 hoary. 



The results recorded above being capable of interpretation on the 

 scheme of inter-relations between the factors as here stated, we may 

 turn back for a moment to a consideration of the results recorded in 



