266 BOTANICAL GAZETTE [Oct., 
ward, the retinacula finally stand nearly at right angles to the 
leg of the insect. The retinaculum, near the point where it joins 
the pollinium, is bent rather suddenly, so that the pollinium ap- 
pears to be flexed upon it. This flexure, which Mr. W.H. Leg- 
gett” has called the knee, is very prominent. There is quite an 
interval between the knees, and the membrane of the retinaculum 
at the knee is expanded transversly to the plane of the polliniunr 
(fig. 5 k.) Robert Brown" has observed in A. purpurascens, 
that the part of the retinaculum extending between the knee and 
the pollinium remains attached to the latter when it is found in the 
stigmatie chamber. This expansion of the membrane serves to 
prevent the withdrawal of the pollinium after it has been inserted 
knee has reached the upper part of the cavity and will go no 
further, the retinaculum is torn across and escapes, leaving the 
pollinium in position to effect fertilization. 
In repeated trials at artificial pollination of the flowers of A. 
Cornuti, A. Sullivantii, and A. incarnata, I succeeded three times, 
in the case of A, Sullivantii, in separating the pollinium from 
the retinaculum without withdrawing the latter from the slit, and 
thus was enabled to insert a pollinium and to draw out a corpus- 
culum at the top of the ale with its two pollinia by the same 
pusculum at the top remained intact. Accordingly, I have seen 
| ee es Ce Vel ins UC‘ “‘“‘“‘ ; 3O*; 
_ Trans. Linn. Soe. Lond., Vol. XVI, p. 724, 
