660 CANNON: A CYTOLOGICAL Basis 
pigeons, and Juel, from studies of hybrid Syringa, arrive at the 
conclusion last given, although this must be implied from Juel’s 
results, as for example, his account of how a Syringa hybrid pollen 
grain may become pure as respects the chromosomes of its nucleus. 
The pigeon hybrid was a fertile one and the Syringa infertile ; the 
possibility thus comes up of the variation in the hybrid pigeons 
being caused and brought about by the normal, rather than the 
irregular maturation mitoses. I have for two years past been 
studying the spermatogenesis of a fertile cotton hybrid, and I have 
attained results similar to those of Guyer. In the case of the cot- 
ton, however, the abnormal divisions were so clearly such, that 
sex cells arising from them would, in all likelihood, not be capable 
of continuing the race. It, therefore, seemed to me that, at least 
in the cotton, variation in the hybrid offspring must come about 
either because the maturation mitoses were such as would induce 
them, or quite independently of these nuclear divisions, since, i 
fertile hybrids the mitoses are normal. 
The nuclear divisions from which the pollen grains arise, as 
commonly understood by botanists to take place, would surely not 
induce the variation in the hybrids after the regular manner de- 
manded by the law of Mendel, and, believing that this variation 
does not occur independently of these divisions, I venture to Sug” 
gest a kind of maturation division which would, I believe, account 
for the variation as above given, and at the same time agree fully 
with the present day observations on the divisions if not with the 
conclusions derived therefrom. 
This matter finds an apparently adequate explanation if we ac- 
cept the results of Riickert and others (Wilson, The Cell, 257 
and 273) based on the study of pure forms of both vertebrates a0 
invertebrates. These results may be stated in brief as follows: 
The chromosomes derived from the father and the mother une 
in synapsis and separate in the metaphase of one of the maturation 
divisions, and also a single longitudinal division occurs, $0 that the 
end is attained that the chromatin is distributed in such a WaY 
that two of the cells receive pure paternal, and two cells p 
ternal chromosomes, and no cells receive chromosomes from both 
is the father and the mother. In this manner é¢ has been demonstf ated eae 
: that pur € races of animals may, and normally do, organise Sev cel a 
of pure descent, — 
ure Mae. 
