210 BOTANICAL GAZETTE [MARCH 
anthesis, it would be difficult to detect, because the young plants 
are not so dimorphic as the mature individuals. The time of 
reversal after anthesis may be at any stage until extreme old age. 
As stated, some individuals produce only imperfect stamens with 
defective pollen and indehiscent anthers, while others produce 
normal staminate flowers with dehiscent anthers and pollen which 
appears normal in every respect. In such cases the sexual expres- 
sion usually involves the entire flower, and the perianth is typically 
staminate, like the perianth on a staminate plant. The reversal 
from femaleness to maleness is of varying degrees, both as to the 
perfection of the stamens and the number of flowers produced. 
The staminate plants of an intermediate sexual expression are 
usually so at the beginning of anthesis, few staminate individuals 
developing carpellate structures at a later stage unless they do so 
from the beginning. Some, however, continue to be more or less 
intermediate up to the time of old age and death. If some method 
of rejuvenation could be employed, it is probable that plants 
purely staminate at first might be induced to become carpellate 
later, but under ordinary conditions the change in the staminate 
plants, as in the carpellate plants, progresses from femaleness to 
maleness. The reversal in the staminate plants is usually less 
complete than in the carpellate plants, probably for the reason 
that senility usually sets in soon after the beginning of anthesis, 
while in the carpellate plants the long active period after anthesis 
has begun permits the efficient environmental factors to have full 
effect in the growing vegetative tissues. 
A few special cases were carefully studied in relation to the 
progressive change in sexual expression. A number of individuals 
appeared normally carpellate and produced two or three normal 
seeds, and then gradually changed to the staminate condition, 
until finally, before they began to die of old age, purely male sex 
was being expressed. Nothing but typical staminate flowers with 
dehiscent anthers and normal pollen were being produced. Female- 
ness had been changed completely to maleness; and this change 
had taken place in plants which in the seedling stage had been 
determined as carpellate individuals with decided characters 
peculiar to the female state. There is but one inevitable rational 
