I 



IX] and Spurious Allelomorphism i6i 



possible that such a system as the 7:1:1:7 may be the 

 result of a complex series of repulsions exerted among a 

 number of factors. At present this suggestion is quite 

 unfounded. It could however be tested if breeding on a 

 really large scale could be undertaken, and supposing it to 

 be true, the evidence for its truth would appear in the 

 relative infrequency with which some types appeared. From 

 that evidence the missing gametic combinations could be 

 identified. As yet however it is quite premature to pursue 

 such an analysis, and we must be content to note that 

 when, as in these Sweet Peas, there is heterozygosis be- 

 tween a number of distinct allelomorphic pairs, the numerical 

 proportions in which the various combinations occur may 

 certainly be affected by the interactions exerted by allelo- 

 morphs of different pairs upon each other. 



The Possibility of Selective Mating between Gametes. 



It has naturally occurred to many minds that as gametes 

 are now known to possess differentiating qualities, these 

 differentiations may affect the readiness with which various 

 classes of gametes may unite. We recognize that the simple 

 Mendelian numbers are produced when every kind of female 

 gamete has an equal probability of uniting with every kind 

 of gamete produced by the male. Conversely, when irregular 

 and unexpected numbers appear as the result of experi- 

 ment, the question may have to be considered whether the 

 irregularity is not due to a selective assortment taking 

 place among the gametes, such that certain types of unions 

 occur in fertilisation with greater readiness than others. 

 Hitherto it is doubtful whether any instance has been dis- 

 covered in which abnormal numbers can be proved to 

 occur with such regularity as to warrant a recourse to this 

 hypothesis. Correns observed several families of maize 

 where F^ from F^ round seed x wrinkled seed, self-fertilised, 

 contained a great excess of round seeds. The totals were 

 8975 round : 1711 wrinkled where the expectation is 8014 

 round : 2671 wrinkled. Thirty-five F^ plants contributed 

 to this total and the discrepancy between observed result 

 and expectation was fairly constant throughout. 



B. H. II 



