388 W. Bateson. 



In several earlier writings I have nsed the expression "compound" 

 in referring to such characters as the grey of mice, which are liable 

 to give various other colours in Fç,. It may probably be desirable 

 to retain some such term to denote characters of this kind. But 

 whereas in the earlier stages of the inquiry I regarded a compound 

 character as capable of being tiansmitted in its entirety as allelo- 

 morphic to the absence of the whole compound, the further work that 

 has been described shows as has been seen that this conception was 

 erroneous. The process of ''resolution" by which the breaking-up of 

 compound characters was supposed to be eftected, similarly becomes 

 capable of description in terms of simple recombination. 



The F, ratio 1) : 7. 



AVe have seen how a character may be imperceptible unless it 

 meets in the zygote another character belonging to another allelo- 

 morphic pair. The next complication arises from the existence of 

 characters which depend for their appearance on the presence of two 

 factors neither of which can be perceived in the absence of the other. 

 It is to this phenomenon that is due the appearance of coloured 

 flowers in Fj from the cross of two white Sweet Peas (Lathyrus 

 odoratus). It was in the attempt to trace the inheritance of pollen- 

 shapes that this case was found. Certain white Sweet Peas have 

 roundish pollen-grains instead of the normal elongated grains. Crossed 

 together these two types gave coloured F,, usually purple, rarely red. 

 F.J from such a cross contains 9 coloured: 7 white. Inspection of 

 such a table as that given in Fig. 13 shows that the 9 are due to 

 the meeting of say, C and R, while in the 7 squares either one or 

 both of these factors are wanting. If the Fj is puri)le, F.^ has always 

 liitherto contained red members, the ratio being (Fig. 15) 



27 purple : 9 red : 28 white 

 36 28 



9 7 



The distinction between the purples and tlie reds is obviously due to the 

 0])eration of a third factor, which we have called B, blue, invisible 

 in the absence of and R, or either of them. In certain families 

 also both the purples and the reds show a number of subordinate 

 types, distinguished by various intensities of colours, and by peculi- 

 arities in the distribution of the colours on the flowers. All these 

 can be shown to be due to the presence or absence of several other 

 distinct factors whose effects are superposed on those of the primary 

 factors we have already named. 



