March, 19./1.J K. FUjn ASJJ Y. KUW ADA:— ON THE COMPOSITION 85 



Thus in the case AA.a the dominant factor A is duplicated 

 and the seeds of darker colour arc produced, while in the case 

 aa.A it is single, and the light blue colour is the result. 



According to the old formula, Fi will be Aa in both reci- 

 procals, and no difference of the intensity of colour should be 

 expected. 



It is also to be noted that, if the maternal plant with black 

 seeds is homozj'gotic, the dominant character in seeds, is re- 

 presented in triploid condition ns 3vl, so that the /^i-seeds 

 above referred to, where it is represented either as 2A or A, 

 may be coiisiderabl}' lighter in colour when compared with the 

 seeds of the maternal plant. 



One may object that i^i-plant is always diploid and not 

 triploid, notwithstanding the triploid nature of endosperm of 

 the seed ; for the endosperm does not develope to be a plant 

 of normal statue and produce pollen grains and embryo-sacs, 

 and the gametes are always derived from the plant developed 

 from the legitimate diploid embryo. That is quite true, we never 

 maintain that /^i-normal plant is triploid ; but the seed of P- 

 plant or jFj-plant is triploid so far as the endosperm, which is 

 one of the two embryos or young F^- or P^-plants and presents 

 the chief seed characters, is concerned. Thus the factorial 

 formula ought to be put down as triplicated, whenever the 

 seed -characters belonging to the endosperm, and neither to the 

 definitive embryo nor to other tissues as seed coat and pericarp, 

 are considered. 



We find that this point w^itli regard to the explanation 

 of the results of reciprocal crossings of maize has been fully 

 recognized by Correns^-' and still he seems not to have realized 

 that the effect of the triploid nature of endosperm tissue 

 and the application of corresponding compositions of fac- 

 torial formulae in the Mendelian researches is more than 

 that. We will soon return to this point ; but let us now 

 compare i^^'gcnerations as represented in the old and the 

 new formulae, and take a simple case of a monohybrid between 



1) CoRREXS, C, Die nciK'n Vererbungsgeselze. lt'12. p. 54. 



