174 
$ 185. Cleistogene—Flowers — Zhe cleistogene flowers of Viola 
cucullata are greatly deformed and yet very efficient. Their every 
ovule is impregnated, and their capsule full of seeds. The flower 
is closely inflected on its stalk, conical in shape, having only its 
pale green calyx to cover it, which it does closely, never opening 
until forced open by the growth of its ovary. The anthers are in 
close contact with the stigma and open towards it, discharging on it 
all their pollen : at least, while the pollen is plainly seen under the 
stigma suspended by its tubes, I could find none in the anther cells. 
The average number of seeds in these capsules is 60, and the 
grains of pollen in each anther are certainly few, if any, more than 
12. Here then is the economy of cleistogene flowers. They dis- 
pense with nectar, with superfluous pollen, and with the flaunting 
corolla, and yet, by husbanding all their pollen, secure a better crop 
even than the open vernal flowers. 
Are the flowers of Gentiana Andrewsii properly cleistogene ? 
They afford now an interesting study in connection with the ques- 
tion of insect agencies in fertilization. On the 19th of September 
I examined numerous specimens in all stages of advancement. In. 
every flower save one the corolla was completely and doubly closed, 
as usual, by the interpetalary appendages in contorted folds, which 
are again covered by the true petals. The appendages are 2-lobed 
and in our specimens not at all fringed, as sometimes described. 
The 5 anthers, slightly cohering, closely encircle the short style just 
beneath and in actual contact with the 2 recoiled stigmas. After 
impregnation, the stigmas, by the growing ovary, are carried far 
above the anthers. 
In every flower whose anthers were open pollen was found on the 
stigmas, but not scattered on other parts ; all that was not on the stig- 
mas apparently remained yet in the anthers. This was the case even 
with that flower whose corolla was open by a small round aperture. 
The grains are very small—four times smaller than those of the 
Morning-Glory. I estimated their number to be about 400 in each 
anther—z,ooo in the 5 anthers. I also estimated the number of 
ovules in the ovary and seeds in the capsule, and found them to be 
ro0o (in some a little more, in others a little less.) 
Nectar abounds in this as in other flowers, secreted in the 
‘depths of the corolla tube. 
From these observations I infer that the flowers of the closed 
Gentian are properly cle‘stogene ; that insects have little or nothing 
to do in their fertilization. Of the 50 flowers examined both at 
night and morning, only one was opened (as if by an accident). 
None showed signs of disturbance of pollen previous to impregna- 
tion. In some of the half-grown capsules worms were found, having 
eaten their way in. The plant is doubtless completely self-fertilized. 
The nectar goes untasted, serving some’ other purpose than a bait 
for bees. The pollen is not wasted, and there is none_or little to 
_ spare, being only 2 grains to each ovule. ac AS ‘ 
§ 186. Cereus serpentinus, Lag.—A few days ago I’ was much 
_ surprised to see one of my, night-flowering Cerei, perfect its bloom 
in the early hours of the day, Sept. 3d, instead of late in the even- 
