X 
grains by means of an opening in the sides of the latter; an observation 
which 1 have not succeeded in repeating. I find that the pollen-grains 
have been correctly described as extending into a pollen-tube, after 
being applied to the surface of the stigma, and it appears to me that 
the active granules pass out of the pollen-grains into those tubes. 
The Ovary adheres firmly to the tube of the calyx, and is often 
so twisted, when the flower is about to expand, that its back and the 
floral envelopes belonging to it is turned to the front. It consists of 
three perfect carpels, stationed alternately with the stamens, opposite 
the petals, and bearing the placentae in their axis, and of three other 
pieces alternate with the first, destitute of placentae, and eventually 
separating from them when the fruit is ripe. These pieces may be 
either regarded as the united margins of the carpels, or as imperfect 
carpels ; the principal objection to the latter hypothesis consisting in the 
intermediate pieces being external with regard to the undoubted carpels. 
Dr. Brown takes a different view of the position of the carpels (On 
the sexual organs , Sfc. in Orehidece and xisclepiadece , p. 12). He con¬ 
siders them to be stationed opposite the sepals; consequently the in¬ 
termediate pieces above described must be in his mind representations 
of the dorsal sutures of the carpels, and he must regard each placenta 
as of a double nature, half belonging to one carpel and half to another. 
To this opinion he appears to have been led both by his own theoretical 
views of the general structure of an Orchidaceous flower, and also by 
supposing in common with all other Botanists at that time, that placentae 
must necessarily alternate with stigmata. But the evidence of Oroban- 
ehaeea? shews, as 1 have elsewhere explained, that in this respect the 
theories of Botanists were unsound; and that placentae are beyond all 
doubt produced by the axis of carpels as well as by their margins. 
The stigma is a viscid excavation in front of the anther, and just 
below it. In most cases it is quite simple, merely terminating in a glan¬ 
dular dilatation of the upper margin, called the rostellum. It is lined 
with a lax tissue composed of minute ascending jointed hairs, and has 
a direct communication with the cavity of the ovary, either open or but 
imperfectly closed up. The glandular dilatation, in all Vandeae and 
Ophrydcae, and in many genera, separates from the stigma and adheres 
