334 THE BIOLOGY OF FLOWERING PLANTS 



It seems to be quite easy to modify the sex distribution 

 in species the individuals of which produce two or more 

 kinds of flowers. It is scarcely possible to change the sex 

 of an individual with only one type of flower. A striking 

 example of the difficulty of this is Satureia hortensis, which 

 has gynomoncecious and female individuals. The indi- 

 viduals with hermaphrodite and female flowers are easily 

 influenced. Under bad conditions they produce only 

 female flowers ; with specially favourable nutrition the 

 proportion of hermaphrodite flowers is much increased. 

 But the individuals with only female flowers can by 

 no means be made to form hermaphrodites. Indeed 

 Schaff'ner's experiments with Arisaema are so far the only 

 indication we have that it may be possible to influence 

 such individuals, and this plant normally has many 

 intermediates. In strictly dioecious plants no means 

 are as yet known by which the sex distribution may 

 be controlled. It may be altered by Ustilago violacea in 

 the case of Lychnis, and a slight degree of intergrading 

 may exist. The potentialities of both sexes must be present 

 in each individual, but the natural determination is precise 

 and firmly fixed. It requires a profound and intimate 

 influence, which we cannot yet imitate, to change it. 



The question then arises whether the determination of 

 the potentiality which is to appear is caused by some 

 inherited factor. There is good evidence that, in some 

 cases at least, this is so. Correns (1907) first obtained 

 evidence from the behaviour of the dioecious Bryonia dioica 

 when crossed with B. alba, which has hermaphrodite flowers. 

 Whichever plant is used as pollen parent, the off^spring are 

 all unisexual, which shows that, in the first place, Bryonia 

 dioica has some factor which prevents the formation of a 

 hermaphrodite flower. If pollen of B. alba is used on 

 B. dioica the off'spring are all females. If B. dioica is used 

 as the pollen parent, half the offspring are males and half 

 females. This behaviour is explained on the hypothesis 

 that the male B. dioica has an inherited factor which sup- 

 presses femaleness ; but, as it is effective in half the offspring 



