732 
DUNCAN S. JOHNSON 
absence of the latter organs in certain cases may simply mean 
that these starved prothallia fail to attain that degree of matu- 
rity (whatever that may mean) at which archegonia are normally 
produced. This view seems distinctly confirmed by the experi- 
mental work of Miss Wuist ('10), on the prothallia of Onoclea 
struthiopteris. Campbell ('05, p. 314) has stated that these 
prothallia are constantly dioecious. Miss Wuist finds that ordi- 
nary soil cultures show about 1 per cent of protogynously monoe- 
cious prothallia. When prothallia, from a soil culture, which have 
borne only archegonia, are grown in Beyerinck's fluid for five to 
seven days they begin to bear antheridia. Similar results were 
obtained when prothallia bearing archegonia only were trans- 
ferred from distilled water to Knop's solution. Miss Twiss (^10, 
p. 168) has shown that the prothallia of Aneimia and Lygodium 
are not really dioecious but merely protandrously monoecious. 
Goebel ('05, p. 220) suggests that the evidence for persistent 
dioeciousness in the ferns is everywhere inadequate. 
It seems possible, from what has been said, that all the cases 
of dioeciousness described among homosporous pteridophyta are 
attributable to external factors. This belief is strengthened by 
the work of Correns ('08, p. 661) on Satureia hortensis. He has 
shown that, in this gynomonoecious species, individuals that 
under normal conditions produce 15 per cent of normal females, 
7 per cent of imperfectly hermaphrodite and 78 per cent of func- 
tionally hermaphrodite individuals can, by cultivation on poor 
soil, or with insufficient illumination, be induced to form a larger 
percentage of pure females. By the combined action of these 
two agencies the male tendency may be so greatly inhibited that 
but 17 per cent of hermaphrodite flowers are formed while the 
proportion of pure females rises to 79 per cent. Correns also 
notes that the percentage of hermaphrodite flowers is greatest at 
the height of the blooming season, while, at the beginning and 
end of the flowering period, female flowers preponderate. 
Another interesting case with a bearing on this question, and 
one that perhaps brings us a step nearer the immediate causal 
factor, is that of Mercurialis annua, studied by Strasburger ('09^, 
p. 507). He found that '^.ertain isolated plants of this species 
