236 Atkinson. 



position. The chromosomes from pijcnocarpa representing; furrowedness 

 and repandness of rosette leaves would occupy a favorable position, 

 while those representing narrowness and outness would be in an un- 

 favorable position. The chromosomes for green stems and green tubercles 

 from pycnocarpa would enjoy a favorable position, those for red stems 

 and red tubercles from nutans an unfavorable one. The chromosomes 

 representing the petals of pycnocarpa are in a favorable position, 

 while those from nutans are unfavorably situated. In this case the 

 favorably situated chromosomes dominate, the unfavorably situated ones 

 have no influence. "Synapsis" is unimpaired, it is neither "imperfect" 

 nor "false", and the hybrid is highly fertile, and segregation does not 

 take place in the second and succeeding generations. In a certain sense 

 hyhrida pycnella is an intermediate hybrid since it possesses characters 

 from both parents. But since these characters are inherited in their 

 entirety, it may be regarded as a segregate. In hyhrida tortuosa, the 

 indications are for the reverse supposed arrangement of the chromosomes 

 in the zygote. 



The provisional hypothesis which I have suggested in the previous 

 paragraph, of a differential division or sorting in the zygote, was briefly 

 outlined (Atkinson, 835, 1914) before I became aware of Swingle's 

 zygotaxis theory. It is very similar to it in some respects, particularly 

 in the chance association of the bearers of "factors" from the two parents 

 in the Fi zygotes and in the equal role played by each member of a 

 homologous pair in the production of true intermediate hybrids or blends. 

 But it differs from Swingle's theory in regard to "chromosome" con- 

 stitution of the two segregate hybrids (pycnella and tortuosa). It is in 

 reality a Qiodification of the zygotaxis hypothesis. There is not, 

 however, any cytological evidence that the positional arrangement of 

 chromosomes in the cells of hybrids bears any relation to their morpho- 

 logical characters. 



5 th. The reaction theory. In view of the fact, which seems 

 to be now quite well established, that certain hereditary qualities are 

 governed by enzymatic action, it may not be improbable that all the 

 factors, or bearers, may work through enzymatic or catalytic or other 

 chemical processes. But when the substances mixed are alike in all 

 cases and the conditions are the same, it is not intelligible, from a 

 chemical point of view, why in some eggs one reaction should take 

 place, while in others the reactions should be totally different. Even 

 if we accept in principle the reaction theory, the different reactions 



