Keur — The Pollination of certain Species of Dendrolnum. 49 



the passage, as lu the case of the first flower. If, liowever, a bee attempts 

 to enter a flower which has received a previous visit from that or any 

 other bee, it finds the anther blocking the entrance to the passage; and 

 if it attempts to force its way in, it only pushes the antlier backwards, 

 till tlie latter comes in contact with the stigma (Plate VI., fig. 3], to which it 

 adlieres and whicli it completely covers, thus preventing the deposit of any 

 more pollen on the stigma, or, if no pollen has already been deposited, of 

 any pollen at all. 



When the flowers are being frequently visited by bees, many may be 

 found with their empty anthers pushed backwards on to the stigmas. 



This mechanism prevents self-pollination, and allows of only a single visit 

 to the nectary of each flower. 



In the case of the first flower visited by a bee, pollination cannot take 

 place, as the bee enters without pollinia, and on its retreat leaves the anther 

 blocking the way to the stigmas. 



I found that when a flower of Bendroblnm Dalliousieanum had its pollinia 

 removed, and placed on its own stigma by artificial means, the flower soon 

 withered away ; but when the pollinia were placed on the stigma of the flower 

 of another plant, that flower formed a capsule. Thus if self-pollination could 

 take place by natural means, it would probably be of no benefit to the flower. 

 I artificially pollinated seven flowers of D. Daltiousieaninn ; of these six were 

 pollinated with their own pollen, and all withered without showing any 

 swelling of the ovary ; the seventh was pollinated with pollen from the flower 

 of another plant, and formed a good capsule. I did not experiment on a 

 sufficient number of flowers to enable me to say definitely that a self-pollinated 

 flower would never produce a capsule ; but from these experiments, and from 

 the mechanism of the flower, it may be concluded that a capsule is never 

 naturally produced by self-pollination in this species. 



Most of the Dendrobiums belonging to the section Eudendrobium show 

 the same mechanism as D. Dallwusieanum. In all the species belonging to 

 tliis section examined by me the fllameut was elastic, and the anther, after 

 dislocation, was jerked downwards in front of the entrance to the passage. 

 In the following list I give the results of artificial pollination in various 

 species of this section. In many cases the number of flowers experimented 

 with is too small to enable definite conclusions to be drawn with regard to 

 that species ; but the results, taken all together, tend, with a few exceptions, 

 to support one another. By the term " cross-pollinated," I mean that a 

 flower has been pollinated with the pollen from a flower on another plant. 

 Under the heading " Self-pollinated " I have included eleven flowers of 

 Z). capillipen, which were pollinated with the pollen from other flowers on the 



e2 



