288 MR. J. C. WILLIS’S CONTRIBUTIONS TO 
the apex, where are two horns projecting beyond on either side of 
the anther (fig. 7). The further end of the column has a sharp 
keel-like ridge on its anterior (inner) face, and the margins ot 
the lateral wings are also folded inwards: two gradually nar- 
rowing grooves are thus formed leading to the stigmatic cavity, 
whose entrance is a long horizontal slit about 1 mm. wide (fig. 7). 
The cavity turns downwards inside the column (fig. 10), making 
a sort of pit. From its base a strand of conducting parenchyma 
runs downwards through the column. Above the stigma is a 
sharp ridge (e, fig. 7) bearing the anther. There are two pollinia 
attached by short caudicles to a pedicel of moderate length 
(fig. 8), which in its turn is attached to the upper side of the 
saddle-like portion of therostellum. This is hollow, and contains 
a quantity of a viscid substance which does not fill the cavity. 
It is continued into a long narrow projection (fig. 7) which runs 
downwards parallel to, and just clear of, the edge of the central 
ridge of the column. Each pollinium has a central hollow opening 
to the outside at its further end (fig. 9). 
The labellum and column are, except at one or two points to 
be mentioned below, of a most extraordinary slipperiness. They 
feel to the touch as smooth as ice. 
With this introduction, we may go on to consider the actual 
mode of fertilization. Ifa sharply pointed pencil be slid side- 
ways down the column, it follows the crest of the central ridge, 
and the point goes under the projecting tongue of the rostellum, 
and then reaches the viscid matter in the saddle. Owing to the 
hollow shape of this latter, already described, only a smal] portion 
of the viscid matter touches the pencil ; and as one goes on with 
the movement, the adhesion is not sufficiently firm, and the 
rostellum often springs back into place. Occasionally, however, 
the adhesion is firm; and it can always be made so by pressing 
the saddle against the pencil. In this case the pollinia are re- 
moved, leaving the empty anther still in position. The whole, 
t. e. tongue, saddle, pedicel, and pollinia, stand in much the same 
straight line. No movement appears to occur after removal. 
The appearance of the summit of the column after removal of 
the pollinia and of the anther is seen in fig. 10. There remains 
a small narrow portion of rostellum (x in fig. 10) projecting down 
the column. 
If, now, we again slide the pollinia upon the pencil along the 
column, they slide with the utmost smoothness along its slippery 
surface ; and presently, as they reach the central ridge, they go 
