306 Saunders. 
The conclusions given under (4) and (5) are also unsupported by 
evidence. The pollen grains of the Stock races employed all appear 
to be similar and all “good”. There is no indication from their 
appearance that a large proportion are sterile as has been found by 
other observers to be the case in some of the Oenothera forms. The 
suggestion that all the mature grains may be good but that perhaps 
two of the daughter cells of each tetrad do not mature is purely 
hypothetical, and if not confirmed on investigation would constitute 
another large assumption lacking all proof. Pending the production 
of evidence on these points, there is perhaps little to be gained by 
further discussion, and I will pass to the conclusions dealing more 
particularly with Stocks, where the theory comes into direct conflict 
with the facts. 
Firstly, with regard to the excess of doubles otained from the 
ever-sporting forms. The fact of the excess cannot be denied, and, 
as I have suggested, can be explained on the supposition of partial 
coupling between two factors. For as it has been shown that the 
male gametes in these forms do not carry singleness, a coupling of 
this nature between factors carried only by the female gametes will 
give unequal numbers of the combinations in which both factors are, 
(= singles) and are not (=doubles) present, the latter being always 
in excess. 
How does GOLDSCHMIDT deal with this undeniable excess of 
doubles?) He admits that his scheme of sex-limited inheritance only 
provides for a ratio of equality, and that any conception of partial 
coupling of the nature suggested could not easily be incorporated 
into it. He therefore finds himself obliged to take up the position that 
although the actual numbers obtained in some hundreds of families 
represents .an excess of doubles corresponding to a ratio of about 
9d:7s or 8.5d:7-55, either this excess is so slight that we must 
assume that the real ratio is nevertheless 1:1. Or, alternatively, that 
some unknown condition causes the distribution of the sex factors to 
deviate from the ratio 1:1, and that the same cause may bring about 
a similar deviation in regard to the factor determining singleness, in 
which case the ratio suggested might represent the actual ratio. The 
first supposition cannot, I think, be maintained in face of the facts. 
If this view is correct then we should expect that the majority of 
the families recorded would show a rough equality, and that among 
the rest about as many would vary in the direction of deficiency of 
doubles as in the direction of excess. But where are the families in 
