1886. ] BOTANICAL GAZETTE. 267 
no difficulty, on the supposition that the pollinia are introduced 
by the knees, in the observation of H. Mii ‘In 
several flowers which I dissected the corpuscula and pollen masses 
were still in their places, though pollinia, which must have come 
from other flowers, had been inserted into the* stigmatic cham- 
bers.” 
In a number of cases observed by me on A. Sullivantii the 
movement which turns the knees toward the flower is completed 
in about seven minutes, though it has proceeded sufficiently in 
five minutes to turn the flexure far enough to render its inser- 
tion quite likely. With the view of pollination stated in this 
paper, the slow movement is plainly of advantage, because a 
knee can hardly be caught by the wings until it has occurred. 
Now, since only one pollinium is inserted into the stig- 
matic chamber, there is an economic disposal of the pollinia. 
The interval between the flexures is so great that only one of 
them can be caught, while the other passes by the slit uninjured. 
There are, therefore, two chances of a pollinium being transfer- 
red to another plant. If both pollinia were left at each act of 
pollination, there would be bat one chance of cross fertilization. 
Moreover, the chances of cross fertilization would be reduced 
from the fact that the stigmatic chambers would be more nearly 
filled by pollinia from the same source, while if they are intro- 
duced singly, there are more chances of a chamber receiving a 
pollinium from a distinct 
As two pollinia are fastened together, there would be but one 
chance of pollination occurring if the pollinia were inserted by 
the corpuscula, while there are two chances of a knee -being 
caught. When a pollinium has been removed, the broken retin- 
chamber, of course, there would be ie aa 
taking place without regard to the number of pollinia. Such an 
accident would be likely to destroy the whole combination. As 
already observed, on the large flowers of A. Sullivantn and A. 
2Tbid, p. 400. 
