ix] Spitrious Allelomorphism 155 



From these results it appears that the erect standard is 

 dominant to the hooded. Next we have the remarkable 

 feature that whereas the purples and the whites are both 

 represented in the two classes erect and hooded, the reds 

 are all erect. This fact indicates that those gametes which 

 bear the factor for erect standard do not bear the factor 

 B which causes the purple or blue colour ; and conversely 

 the gametes which do not bear E, the erectness-factor, bear 

 B. The gametes in respect of these two allelomorphic 

 pairs have thus the composition Be, or bE, and as regards 

 transmission of characters the effect is that which would be 

 produced if B were allelomorphic to E. In view of this 

 curious fact it seems not impossible that we may be obliged 

 hereafter to extend the conception of allelomorphism, and 

 to recognize that factors concerned with features of organisa- 

 tion which seem to have no special physiological association, 

 may be allelomorphic to each other. 



Another curious result follows from the existence of 

 this repulsion or spurious allelomorphism. If the hooded 

 purples in F^ be bred from, the /% families will be either 

 all hooded purples again or they will be such plants together 

 with hooded whites. Similarly the red bicolours will give 

 jF 3 consisting either of red bicolours, or of red bicolours and 

 erect whites. But the purple bicolours, must all by their 

 constitution be heterozygous in both the factor for erectness 

 and for blueness, and consequently their F 3 families contain 

 hooded purples, erect bicolour purples, and erect bicolour 

 reds in the ratio 1:2:1 (together with whites if the par- 

 ticular /% happens to be heterozygous in either C or R, the 

 fundamental colour-factors). 



Such bicolour purples thus present the anomaly of being 

 permanent heterozygotes, though in appearance they are 

 actually the original type of Sweet Pea. Families may thus 

 consist entirely of 



Hooded Erect Erect 



unicolorous purples bicolour purples bicolour reds 



i 12:1 



and the bicolour purples in such families will give similar 

 families again indefinitely. 



On determining the pollen-characters of those families 



