290 REPORT OF THE CONFERENCE ON GENETICS 
left free in the protoplasm. It often happened that these free chromosomes, 
instead of remaining in the protoplasm, were taken into the nuclei, but 
the process then was an irregular one. It must therefore happen 
sometimes that all the D. longifolia-chromosomes are concentrated within 
one nucleus and the other must consequently form pure D. rotundifolia- 
nuclei. 
If it is true that the characters of a hybrid are determined by the 
combination of the chromosomes, it is very important to gain a knowledge 
of the pollen grains themselves and of the number and arrangement of 
their chromosomes. The pollen grains are generally of the same size and 
structure even among species of the same genus. In the case mentioned, 
however, the pollen grains of D. rotundifolia are sufficiently different 
from those of D. longifolia, and an examination of several pollen grains 
of the hybrid might afford some evidence as to whether the ideas of the 
cytologists and the hybridisers on the segregation of characters in the 
germ-cells are correct. Especially in the case of the hybrid Drosera where 
such different proportions of longifolia- and rotundifolia-chromosomes 
may occur in the pollen grains, the pollen grains should be a very favour- 
able subject for a demonstration of the point at issue. 
The four pollen grains of one tetrad, when ripe, lie free in the anthers, 
and it is therefore difficult to note which pollen-cells are sister-cells, and 
it is very necessary to know this, if one wishes to use the structure of the 
pollen grains as evidence of the correctness of our views as to Reduction. 
In Drosera, however, the conditions are very favourable for demon- 
strating the point at issue. The four pollen grains are always joined in 
tetrads, even when they have reached the stigma. We therefore have 
here a rare case of germ-cells in a hybrid, where the parent germ-ceils 
are-unequal and the four cells derived from each mother-cell always remaim 
connected with one another. 
As I said, it must happen that sometimes the D. rotwndifolia- 
chromosomes in the hybrid must be purely segregated in one nucleus, and 
the D. longifolia-chromosomes in another. Of the four nuclei in a tetrad 
two and two have quite the same chromosomes, since the second division is 
an equation-division. Now in the Drosera-hybrid most of the pollen 
grains are like those of D. longifolia, but sometimes I found a very 
interesting structure of the tetrads: two cells of them were quite 
D. rotundtfolia-like, while the two others were D. longifolia. 
I therefore think that we have here a very good demonstration of the 
results of a Reduction-division, the important act in plants and animals, 
through which the characters of the sexual nuclei got fixed. The two 
pollen grains are quite D. rotundifolia, and the other two D. longifolia. 
I also think that this isa good illustration of the hypothesis of gametic 
segregation put forward by Mendel. 
Finally, I will in a few words touch upon another question con- 
cerning the cytological structure of hybrids. It is known that many 
hybrids show the peculiarity that their F,-generation is not homogeneous, 
but exhibits a graduation in characters between those of father and 
mother. 
I have had the opportunity to study such a case. I had the pleasure 
to examine from a cytological point of view the very interesting material of 
