232 Atkinson. 



all the egg cells bearing- the paternal characters likewise were sterile, 

 then biennis would be a hybrid by the union of the paternal and maternal 

 hereditary characters. 



As de Vries remarks (1. c. 1913) other explanations of patrocliuy 

 are possible, and the one above mentioned would need other reinforcing- 

 hypotheses. To apply the hypothesis of a dissimilarity of the gametes 

 in the parents to the production of the twin and triplet hybrids from 

 reciprocal crosses between Oe. pycnocarpa and nutans would require 

 several very ingenious reinforcing- hypotheses to account for the blend- 

 ing- of the veg-etative characters in one hybrid and a sorting and 

 redistribution of parental characters in two other Fi generation hybrids. 

 It would require several different kinds of functional gametes in one 

 or both parents, and in such a case the progeny of the parents in self 

 fertilization would split up into different types in the Fi generation, 

 which is not the case. 



In certain animals it is known that two kinds of sperms, are pro- 

 duced, but these relate to sex determination and a cytological basis for 

 their dissimilar nature has been found in chromosome differences. No 

 cytological evidence of dissimilar gametes in plants has been discovered. 

 The hypothesis is an interesting one, but it is still a hypothesis and 

 would necessarily become very complex if one should attempt to apply 

 it to the Fl segregations described in this paper. 



3 rd. Theory of a differential division or sorting in the 

 zygote. This theory, like several other theories of heredity, rests on 

 the assumption of a material basis for the hereditary "characters", and 

 that more or less definite parts of the chromatin, or other portions 

 of the cell are responsible for, or bear in some way, the heredity 

 "tendencies" or "factors" of the organism. The theory of material 

 bearers for the factors supplying the impulse for the repetition in the 

 progeny of characters or processes possessed by the parent has, at the 

 present time, a very large following. Whether or not the hypothesis 

 expresses even a small measure of the truth underlying the mysterious 

 processes of inheritance, it has served, and will for some time serve, 

 an important purpose in stimulating investigation on the role of cell 

 structures, and on the practical operations of plant and animal breeding. 

 In the use of such expressions as "bearers of characters", "factors", or 

 "tendencies", etc., we do not imply that in any manner the characters 

 of the adult are preformed in minute particles of the cell substance, 

 nor that every minute character of the adult is represented by a cor- 



