92 INHERITANCB IN POULTRY. 



wholly pure, since the characteristics in second generation hybrids are rarely 

 exactly like those of their grandparents ; consequently various additional 

 hypotheses have been offered accounting for this feature. Hacker (1904) 

 points out that chromosomes do not pass from cell to cell unchanged except 

 for their growth and division. During the resting stage "of the nucleus it 

 disappears. The new nucleus which arises in the position of the old is at 

 first small ; it arises inside of the old chromosome as a spore arises in the 

 mother cell ; its material has been derived from a part only of that of the 

 mother chromosome ; the remainder goes to form part of the cytoplasm. 

 Though chromosomes from different parents tend to separate to distinct 

 gametes, still all gametes are infected by each kind of characteristic." 

 McClung (1905, p. 329) assumes, more vaguely, a mutual influence of synap- 

 tically paired chromosomes in the prophase of the first spermatocyte. 



A different suggestion is offered by Ziegler (1905). He assumes that each 

 chromosome of maternal or of paternal origin carries determinants of all 

 characteristics. After maturation all gametes contain the same number of 

 chromosomes, but the proportion in them of chromosomes of paternal and 

 of maternal origin varies. Gametes rarely contain exclusively maternal or 

 paternal chromosomes, but whenever the proportion from one parent is high 

 the gamete acts as though it contains exclusively the gametes of that one 

 ancestor. If two gametes that are prevailingly paternal unite in a zygote 

 the resulting hybrids (of the second generation) show all the grand-paternal 

 characteristics. The difficulty with this hypothesis is that, like Cannon's, 

 it does not account (any better than the first hypothesis) for the diverse 

 combinations of characteristics shown in the second hybrid generation. 



Still another suggestion has been made by Morgan (1905). It is that the 

 gametes are not pure, but contain determinants of both allelomorphs a and 

 a' , and that one of these dominates in half of the gametes and the other in 

 the remaining half. The advantage of this hypothesis is that it accounts 

 for latent dominant characters in recessive individuals. This hypothesis 

 assumes that the gametes of hybrids are always impure, and that this im- 

 purity can not be got rid of. This seems to me to be contrary to experience. 

 Moreover, except for the explanation that it offers of latency — which has 

 "been accounted for on other grounds by Cuenot— it offers no practical advan- 

 tage over the theory of pure gametes. 



From the foregoing diversity of hypotheses it is evident that we lack a fully 

 satisfactory cytological explanation of the facts other than that of purity — 

 the fact of imperfect dominance and the fact of particulate inheritance com- 

 bined with purity in the second hybrid generation. Perhaps it will suffice 

 to suppose a restricted purity of gametes such that the determinant of a 

 characteristic may become infected to a slight degree by the presence of its 

 allelomorph. 



