49 



This subsequently extends backwards and becomes attached ta 

 the top of the column, thus forming a little raised falcate 

 eminence, which fulfils a highly important function in accur- 

 ately guiding the pollen masses when they fall from the 

 anther-case into the most advantageous positions. 



The anther is hemispherical, with a slight anterior pro- 

 jection over the tip of the rostellum. In the bud it may be 

 lifted backwards, showing its attachment to the column and 

 displaying its two loculi, each containing a perfectly spherical 

 yellow^ pollen-mass. There is as yet no signs of dehiscence, 

 nor is there any connection whatever between pollinia and 

 rostellum. There can be seen, however, on the lower surface 

 of each loculus an oblique sulcus corresponding to the depres- 

 sion between the two lobes of each pollen-mass. It is along 

 this sulcus that dehiscence subsequently takes place. The two 

 loculi do not nearly fill the anther-case. They occupy only 

 the upper and anterior portion, and as the case rests like a 

 small inverted basin on its rim upon the rostellar roof, the 

 pollen-masses are in this way kept raised a little above the 

 stipites. 



The stigma (figs. 5 and 6) is elliptical in shape, its long 

 axis being transverse. It is placed immediately below the 

 tip of the rostellum. It is deeply concave, and communicates 

 in its lowest part directly with the cavity of the ovary by 

 means of the stigmatic canal. Situated near each end of the 

 ellipse is a separate tiny depression. This depression is just 

 the correct size to receive a pollen-mass. 



Dehiscence of the anther is completed just as the seg- 

 ments of the perianth begin to separate in the act of expan- 

 sion, and the pollinia still retained within the dehiscing 

 loculi are found to be very sticky on their lower surfaces. The 

 contour of the rostellar roof is exactly adapted to receive 

 them when they are subsequently released and fall. 



A very little later, and before expansion of the flower is 

 complete, the anther-case becomes very hard and rigid, the 

 pollen-masses are liberated, and each drops upon the stipes a 

 little above the distal end, being guided accurately into posi- 

 tion by the falcate mesial band already referred to. Bilobing 

 of the pollinia is very slight, and only noticeable as a groove 

 on the under-surface. When they drop from the anther they 

 do so in such a manner that each is received by the stipes in 

 this particular groove and becomes adherent by its own sticky 

 secretion. This secretion is possibly selective in its action, as 

 the pollinia never appear to become adherent to any of the 

 adjacent parts. Possibly, however, this may be due to the fact 

 that the stipites stand up in relief from the underlyins^ struc- 

 tures. The pollen-masses do not harden immediately they 



