INHERITANCE OF SEX IN MERCURIALIS ANNUA 429 
assumption that there are either two kinds of eggs as Bateson (1909) claims 
for the female Bryonia dioica, or two kinds of sperms as Correns (1907) 
claims for the male Bryonia dioica. The eggs of the homosporous fern 
cannot be ^assumed to be of two kinds, because the resultant sporophyte 
is not sexually differentiated, or at best it is potentially hermaphroditic. 
Nor, for the same reason, can the sperms be conceived as being of two kinds. 
The results obtained in selfed Mercurialis annua females show that, 
although there m^y be gradations in strength of femaleness, no male plants 
are ever produced, which, on Bateson's assumption, should be possible. 
It is to be noted that Bateson has drawn the conclusion from his crosses of 
Bryonia dioica X B. alba that the eggs of the female of B. dioica are of two 
kinds, male-producing and female-producing, and that this accounts for 
the approximately one-to-one ratio of the sexes in a population of males 
and females of B. dioica. Mercurialis females, if their eggs are of two kinds, 
male-producing and female-producing, should yield both male and female 
offspring on Bateson's assumption. The offspring of selfed female plants of 
Mercurialis are, however, all female or prevailingly female. 
Male gametophytes arise from microspores. These microspores have 
their sex determined before the reduction division. The male gametes 
are produced by the male gametophyte which is the haploid generation. 
If there is to be a differentiation of the male gametes into those bearing 
male determiners and those bearing female determiners, it must occur within 
the microgametophyte — that is, after the reduction division has occurred, 
through either the loss or the addition of cellular material, unless it be 
assumed that the whole gametophyte generation is of two kinds, male- 
determining and female-determining. This process is not at all analogous 
to the production of male gametes in the males of animals. If we assume 
with Correns (1907) that there are two kinds of pollen grains, male-produc- 
ing and female-producing, we must assume that the differentiation occurred 
at the reduction division. Yet we find that pollen grains are male haploid 
generations resulting from male spores. Neither Bateson nor Correns 
has adequately discussed the relation of the gametophytic generations which, 
in plants, are interpolated between the spore and the gametes, to problems 
of sex determination. The results obtained in male cultures of Mercurialis 
annua negative Correns' assumption of two kinds of pollen grains. Correns 
like Bateson arrived at his conclusions from the results of crosses of Bryonia 
dioica X B. alba. According to Correns, the presence of two kinds of pollen 
grains, male-producing and female-producing, in a dioecious form such as 
'Bryonia dioica accounts for the approximately one-to-one ratio iii a popula- 
tion of males and females. On that assumption, selfed males of Mercurialis 
annua should yield both male and female offspring. However, the offspring 
of selfed males of Mercurialis annua are male or prevailingly male, and thus 
negative Correns' assumption of the two kinds of pollen grains. Further, 
because of the occurrence in plants of one species of hermaphrodites, mixed 
