544 ME. H. O. FORBES ON 



begins to rotate into its natural position, and concurrent with the 

 commencement of this movement some influence (I know not what) 

 causes the top of the rostellum to become inverted close down to 

 the posterior wall of the stylary canal. It occurs even before the 

 anther has rotated a few degrees from its erect position. Along 

 with the invagination of the rostellum the lower lip of the 

 stigma is in consequence dragged upwards. Dissections of the 

 column show that the rostellum goes on gradually elongating 

 down the stylary canal, while the pollinia, slowly continuing to 

 rotate, finally precipitate themselves into the stigma, whose 

 flap-like sides embrace the pollinia within the anther-cap. On 

 the conclusion of these singular movements, no remains of the 

 stigma can be seen ; only the swollen and viscid pollinia show 

 slightly between the edge of the anther-case and what was the 

 lower margins of the stigma. As a rule all these operations are 

 concluded before the full expanding of the flower, which, after 

 remaining only a few hours open, fades and envelops the column. 

 The pollinia are thus swallowed up by the stigma, and being 

 bathed by the viscid matter, have only to emit their tubes 

 vertically downwards, while they retain their position in the 

 anther-case. I have found also that in some cases the rostellum 

 is not depressed down the stigmatic canal, but remains in what 

 is the more natural orchideal situation, as a broad flat floor to 

 the anther, and projects far over the stigma. "When the flower 

 of Arundina speciosa has this form, it invariably, as far as my 

 observations enable me to speak, falls off unfertilized. The 

 pollinia in this form lie far back in the anther, entirely hid by 

 the anther-case, which fits close down all round ; and between 

 its external edge and the rim of the rostellum there is a con- 

 siderable platform. An insect to secure the pollen would require 

 to land on this platform and lift up the anther-case to extract 

 the pollinia ; which supposing it to have successfully accom- 

 plished, it might wander far to find a stigma to which to apply 

 it. Flowers with this conformation, however, remain expanded 

 and fresh for several days, in marked contrast to those of the 

 first form, which close up in a very few hours. On the upper 

 surface of the rostellum, and in the median line, there is a well- 

 marked ridge which runs out to the tip to form the central pro- 

 montory of the rostellum. 



In describing Phaius Blumei, I remarked that there existed on 

 the back of the stigma a prominent ridge running down nearly to 





