F. L. Engledow 127 



For the cross here described ^i data are not available but Biffen 

 [(1) p. 38] found that all the grains borne by all the F^ (and therefore "/") 

 plants were identical in form and appearance and, within the limits of 

 natural fluctuation, of one common length-type. His conclusion is thus 

 precisely the one which has already been expressed in this paper — a 

 conclusion which appears to be decidedly out of harmony with the 

 double fertilization hypothesis. 



As already recorded, the F^ of the Polish x Kubanka cross contained 267 

 "/" plants and their grains afford material for a more extensive inquiry 

 than is possible with the necessarily limited number of F^ plants in any 

 cross. Table XXV contains the frequency distribution of the lengths 

 of these grains, and it exhibits clear iinimodality. That unimodality of 

 distribution is unreliable as an index of singleness of grain-type has, 

 however, already been demonstrated. By measuring the grains of the 

 more numerous F^ . I plants a distribution based on a greater number 

 of observations might be obtained but it could afford no more evidence 

 as to the existence of sub-types of length than does the F. . I distribution. 

 Only by measuring a great number of grains and grouping the lengths 

 of those which grew respectively into K, I, and P plants would it be 

 possible to determine whether the same length of grain (i.e. of endosperm) 

 was or was not associated with the K, /, and P types of embryo. 

 It is proposed to make this test next year but in the meantime, metrical 

 methods failing, the assistance of eye-judgment has been sought. 

 Suitably devised tests demonstrated the ability of the eye to pick out 

 infallibly the members of the three types from an artificially prepared 

 mixture of K, I, and P grains. Beyond this, however, the eye could 

 not go. Despite very extensive trials no evidence of sub-types either 

 for length or appearance could be found among the grains borne by "/" 

 plants. If, owing to " double fertilization," the endosperm is " hybrid " 

 in nature, the fact appears to find no expression in the length or appear- 

 ance of the endosperm both of which are such that they may be described 

 as " maternal " characters. 



The expectation derivable from an extension of the double fertiliza- 

 tion hypothesis is of interest. Correns (6) and others assume that the 

 secondary or definitive nucleus of the embryo sac exerts twice as much 

 influence on the endosperm as does the single generative nucleus of the 

 pollen grain — the reason for this being that the definitive nucleus is 

 formed by the union of tiuo of the polar cells of the original nucleus of 

 the ovule. The male generative nucleus unites, of course, with the 

 female definitive nucleus and from the divisions which follow this union, 



9—2 



