1886. ] - BOTANICAL GAZETTE, - 269 
which has two pollinia’.” This is a plain admission that the 
view that the pollinia are inserted by the corpuscula will not ac- 
count for the formation of dichotomous combinations of pollinia, 
and involves the whole theory of the insertion of pollinia by the 
corpuscula. For if a pollinium in a dichotomous combination 
can be inserted by the knee, any other pollinium may be inserted 
in the same way, and the theory of insertion by the knee may be 
applied throughout without meeting any difficulty. The suppo- 
sition that the pollinia are inserted by the knees explains the for- 
mation of these as well as of unilateral ones, and shows the value 
of both forms; for the combination becomes the equivalent of a 
leg provided with a number of pollinia, each one of which may 
be inserted separately. at 
Finally, I have seen the pollinia of A. Sullivantii introduced 
into the stigmatic chambers in the manner above described. The 
pollinia and the entrance to the chamber in this species are very 
large and are easily seen, and hive bees move so slowly in effect- 
ing pollination that, after a knee is caught, one can see the pol- 
linium slowly disappear between the wings, so that there can be 
no doubt as to the manner of insertion. Commonly, however, 
slit. When a foot is held by a retinaculum the pollinium is found 
in the chamber, with every indication that it was introduced by 
the knee. 
ExpLaNATION oF Ficures on Pirate VIII.—1. Flower of A. Sullivantii 
a. anther wing. ¢. corpusculum. e. entrance to stigmatic chamber. k. 
knee. r. retinaculum. s. stigmatic chamber. p. pollinium. 
i gk Aik enantio 
14Tbid. pp. 191, 192. 
