G. M. Taomson.—Notes on Cleistogamie Flowers of the Genus Viola. 417 
while still in the anther cells, may be seen emitting their tubes to the 
stigma. The style is nearly as long as that of the conspicuous flowers, 
but in order to bring the stigma within reach of the 
pollen, it lies coiled and twisted on the summit of the 
ovary. In some of the flowers all the stamens were 
reduced to mere filaments except one; in others there 
occurs one stamen with both anthers; another with a 
single anther, the rest being represented by filaments ; 
while others again had the rudiments of anthers, but 
no pollen. The size of the pollen grains varied some- 
S9 what in the two kinds of flowers. In the large, con- 
> Q spicuous flowers, the grains appeared uniform in size and 
in shape, and averaged about byth of an inch in length; 
2, a. Si ves from in the cleistogamic flowers, the larger grains were similar 
ar. end X 36. in shape and length to those of the larger flowers, while 
2, b. Pollen grains the majority were rounder and thicker, and not more than 
uu WA | T tooth of an inch in length. 
Viola cunninghamii, Hook. f. 
This species, as a rule, flowers earlier than the preceding. When 
examined by me during the month of January, it bore abundance of cleisto- 
gamic flowers, while the empty valves of their capsules showed that the 
other flowers had already ripened and shed their seed. The ordinary 
flowers are very similar to those of V. filicaulis, and are produced on pedun- 
cles of similar length. The cleistogamic flowers are produced on very short 
peduneles, which lengthen in fruit to or 3 of an inch. All the parts 
in these flowers are very small. The sepals are all present, and of the 
normal form, but the petals are wanting; their disappearance seems to be 
complete, as there was no trace of their presence. 
The stamens are very feebly developed. Two of them have both 
their anthers developed, but these are very small and do not contain 
much pollen. The filament is narrow below, but has its 
upper part extended into a hood. These two stamens are 
placed on opposite sides of the pistil, and their hoods com- 
pletely overlap and enclose it, their anthers being in contact 
with the short, obtuse stigma. 
In all the flowers examined by me, there seemed no 
departure from this type of structure, nor was there the 
variability which I noticed in the cleistogamic flowers of V. 
f filicaulis. 
ab 
8, a. Hooded stamen from cleistogamic flower of V. cunninghamii x 25. 
9, b. Rudimentary stamen from the same x 26, 429 
