certain Varietal and Hybrid Ferns . 203 
admission that the whole case for individuality and differentiation amongst 
the chromosomes rests on a far more insecure and speculative foundation 
than, on other grounds, would be generally admitted. We do not urge 
these considerations in any sense as suggesting that we doubt the state¬ 
ments that have been made respecting the process in Drosera. Our only 
motive for alluding to this case is because it is, from the relatively few 
number of the chromosomes, a much more favourable case than is our plant 
for ascertaining the facts, and at the same time to indicate that, in our 
opinion, the special importance which has been attached by some writers to 
the course of events in Drosera can nevertheless hardly be maintained as 
being of general application. Furthermore, that there are also, as it seems 
to us, obstacles which from a theoretical standpoint make the interpretation 
which has been put on the pairing of the chromosomes in this and in similar 
cases difficult of acceptance. 
The important results which have recently been published by Gates, 1 
as the result of his investigations on hybrid Oenotheras, prove that there are 
other factors connected with the possible modes of chromosome distribution 
of the nature of which we are as yet practically ignorant. Gates finds that 
the hybrid Oenothera lata x O. gigas has 21 chromosomes in its pre- 
meiotic cells. At meiosis this premeiotic number reappears, owing to the 
failure to pair on the part of the chromosomes, and finally each of the result¬ 
ing nuclei receives as nearly as possible the half of this number, i. e. either 
10 or 11 respectively. There are irregularities also, but these are of no 
consequence for our present purpose. The feature around which our 
interest centres is that despite the fact that the two parents contribute 
different numbers of chromosomes, an equal, or almost an equal, number 
are distributed to each of the daughter-nuclei in the heterotype mitosis. 
That is to say, the distribution is not the result of ‘ a pairing and separation 
of homologous chromosomes of maternal and paternal origin, but the 
segregation tends to be into two numerically equal groups V 2 These 
results are in harmony with those we have obtained in Polypodium Schneideri, 
for the chromosomes, whether they are bivalent or not, are about equally 
distributed between the two daughter-nuclei at the heterotype mitosis, 
in spite of the fact that the share, numerically speaking, which is con¬ 
tributed by the two parents respectively is so markedly dissimilar. We 
are far from desiring to suggest that these results affect, at any rate 
necessarily, the views which are entertained by many respecting the allelo¬ 
morphic distribution of chromosomes in a normal plant. It may well be 
that the peculiarities that characterize the behaviour of the chromosomes in 
these hybrids are directly correlated with their sterility. But however this 
1 R. Ruggles Gates, The behaviour of the chromosomes in Oenothera lata x 0 . gigas. Bot. 
Gaz., xlviii. 2 Ibid., p. 195. 
