STRUCTURE OF IMPATIENS FULVA. 149 
posterior, and is bent almost at right angles in its free portion, 
the point of the pistil being inserted into the bend. Springing 
from the posterior portion of the tube formed from the co- 
herent lower portions of the filaments, is a membranous kiud of 
wing, closely adpressed to the pistil—this wing, together with a 
slight projection from the anterior filament, eompletely closing 
in and covering the pistil. A section of the staminal tube, with 
the pistil, is shown at fig. 7, and, after the removal of the pistil, at 
fig. 8. 
The development of the inconspicuous or *cleistogenous ” 
flowers is entirely different. As already stated, the bud re- 
mains much much more minute, and the calycine and corolline 
whorls are much more regular. The buds never open, but are 
pushed off from the already fertilized ovary in the form of a cap, 
as shown in fig. 9. This takes place at so early a stage that it 
is very difficult to detect the process; indeed I examined hun- 
dreds of specimens without finding more than three or four 
instances of the half-expelled cap, which closely resembles the 
calyptra of a moss. The structure of the stamens is very dif- 
ferent from that found in the conspicuous flowers. They are 
equal in length, the filaments free for their whole length, strap- 
shaped, and, although in the earliest stage straight (fig. 10), become 
afterwards remarkably contracted (fig. 11). The anther is scarcely 
broader than the filament; and the quantity of pollen is exceed- 
ingly small compared with that in the conspicuous flowers. 
Owing to the very obscure nature of the stigmatie surface, the 
exact period of impregnation is very difficult to determine; nor 
could I detect any dehiscence of the anthers for the discharge of 
the pollen. Prof. Oliver informs me that it is very commonly 
the case in closed self-fertilized flowers for the pollen-tubes to 
penetrate the substance of the anther itself. The mode in which 
` the cap is thrown off presented a great difficulty to my mind. I 
find this to take place almost invariably at the very earliest stage 
with the first growth of the pistil (though this does not appear to 
be so much the case with the other species grown in this country, 
nor with Z. fulva in America, to judge from Prof. Gray's descrip- 
tion and drawing). Owing to the length of the petioles, the 
flowers are generally found beneath the leaves, so that the caps 
are lost. In a few cases, however, where there had been a dis- 
tortion of the petiole, or some other cause of disturbance of the 
