276 MR. A. W. BE3TST.TT OS CLEISTOGAMIC FLOWEKS. 



the vernal perfect flowers. The pedicels are shorter, and the buds 

 and the young fruit are pendent ; but the ripe capsule is erect ; 

 and in this flower, as in the sweet violet, I was unable to detect 

 the tendency, described by Michalet and Von Mohl, of the cap- 

 sules to bury themselves in the earth. The closed flowers are 

 about one sixth of an inch in length, and do not present by any 

 means the same diversity of structure from the normal perfect 

 flowers as is exhibited in the case of Viola. The five sepals are 

 lanceolate, green, mottled with brown" and never expand. The 

 corolla, which I never found to be entirely suppressed, presents 

 great diversity in its degree of development ; the petals are some- 

 times entirely included, and remain green, while sometimes they 

 assume a white colour, their tips being exserted beyond the calyx, 

 in that case all the five petals are pushed off in the form of a cap 

 by the growth of the ovary. The ten stamens all produce pollen ; 

 the five inner and longer ones have their anthers on a level with 

 and resting on the five stigmas (fig. 13); while Fig. IS. 



the anthers of the five outer stamens do not reacli 

 more than halfway up the ovary, and appeared 

 to me to be functionless. Darwin, however, 

 has seen pollen-tubes from these reach the 

 stigma ; while Michalet, on the other hand, de- 

 scribes them as sterile, or even frequently abor- q a j- s ^ ce t ose iu. 

 tive. The capsules contain numerous seeds, and Pistil and stamens 

 are provided, like those of the perfect flowers, J^ w ^ toglimi0 

 with elastic valves ; but I had no opportunity of 

 comparing the fertility of the closed and perfect flowers. 



Impatiens. 



I. noU-me-i anger e, Linn. — The small cleistogamic flowers are 

 produced in abundance, unquestionably at the same time as the 

 handsome open flowers, and on the same plant and same branch, 

 thus presenting a contrast to what I formerly observed in the 

 case of I. fulva*. The calyx presents a strong resemblance to 

 that of tfie closed flowers of the last-named species, constituting 

 a small brown cap which becomes detached at the base by the 

 growth of the capsule, and which completely covers up the organs 

 of reproduction. The five stamens are similar to those of the 

 closed flowers of L fulva ; the filament is strap-shaped, nearly 

 straight, transparent, and terminates in an arrow-shaped anther; 



* See 'Journal of Linnean Society, Botany/ 1872, vol. xiii. p. 147. 



