MR. A. W. BEKTSETT ON CLEISTOGAMIC FLOTVEBS. 



273 



Fig. 1 0. 



Fertile sta- 

 men. 



Fig. 11. 



in V. cucullata. The fertile stamens (fig. 10) resemble in im- 

 portant points of structure those of the last-named species, but 

 the terminal prolongation of the connective is much 

 broader and larger, in shape like a hood, and toothed 

 at the side ; the two anther-lobes are often a consi- 

 derable distance apart, and are ciliated at the mar- 

 gin. The number of fertile stamens varies. Most 

 usually there are only two, which are coherent by 

 their edges, the two hood-like connectives overlap- 

 ping one another and forming a cap which is closely 

 applied to the stigmatic surface, completely covering 

 it up (fig. n) ? an( j on ij detached from it with diffi- 

 culty, the filaments being ruptured at the base by F- sylvatica. 

 the growth of the ovary, as in the last species. Oc- 

 casionally there are four fertile stamens, which are 

 then coherent in pairs, each pair forming a cap. In one instance 

 I found five ; but the fifth, which stood by itself at the back of the 

 ovary, had only a single anther-lobe. The remaining 

 stamens are almost always present in a rudimentary 

 condition, and consist of an obovate or subulate brown 

 filament, nearly as long as the fertile stamens, denti- 

 culate at the margin, and without a trace of anther- 

 lobes (fig. 12). The anther-lobes open by apical 

 Dores, which are in immediate contact with the an- 

 terior lip of the stigmatic cavity ; and the pollen- 

 tubes may be distinctly seen ascending vertically 

 through the pores. The pollen-grains are extremely 

 few in each anther-lobe, minute, ellipsoidal, and with _ 



a * KlfiP Y1CW Ol 



transparent cell- wall. In order to test the fertility p i 9t ii,withsta- 



of these cleistogamic flowers, I counted the ovules men attached 



in a number of unfertilized ovaries, and the seeds in 



a number of mature capsules which were evidently the product of 



closed flowers. In six unfertilized ovaries 3 



au average of 13 ovules, the number varying from 



10 to 18. In twenty mature capsules the number of 



fi ceds varied between 4 and 20, the average being 



125. It is evident, therefore, that, as far as these 



uumbers offer any indication, but few of the ovules 



remain unimpregnated. It should, however, be noted 



V. sy I vatic a 



founa 



Fig. 1 



every 



very large number) 



