Fecundation in Orchidee and Asclepiadee. 725 
granules were not often to be found even in the tube, especially 
after it had acquired considerable length*. 
Almost every grain in the mass had produced its tube, and 
the tubes were directed from all parts of it towards the point of 
dehiscence. In this state the mass had become more convex 
from the increased bulk of its contents. 
The tube so produced from each grain of pollen cannot be said 
to be emitted from it, but is manifestly an elongation of its mem- 
brane. "These tubes are transparent, cylindrical, about 1-2000th 
of an inch in diameter, neither branched nor jointed, with no 
apparent interruption in their cavity, and when of great length, 
which they often attain, are frequently without granular matter. 
I next proceeded to examine the course of the cord, which in 
most cases,—and indeed in all where the mass had remained a 
sufficient length of time in the fissure,—had opened a passage 
for itself through the membrane, or rather had separated the 
upper edge of this membrane from the base of the stigma, to 
which it was before united. Having effected this separation, 
it was found to proceed along the surface of the base of the 
stigma in a line exactly opposite to the glands seated on the 
apex of the same bevelled angle. The cord having passed 
along the surface of the attenuated base of the stigma until 
it arrives at its articulation with the two styles, then inclines 
towards the inner side of the apex of the style nearest to it, and 
actually introduces itself, wholly or in part, into the hollow of 
the apex, which in this stage is in some degree exposedt. But 
as the partial separation of the styles from the stigma, then 
taking place, is not always suflicient for the free admission of 
the whole cord, a few of the tubes not unfrequently become 
bent, in some cases even zigzag, doubtless in consequence of 
* Tab.35. fig. 7, & 10; and Tab. 34. fig. 12. 
t Tab. 34. fig. 7—9; and Tab. 35. fig. 4, & 10. 
the 
