KuRK — The Pollination of certain Species of Dendohrium. 51 



I frequently tried in this and other species to see if the polliuia would 

 rebound from the lip on to the stigma, when the anther was suddenly 

 dislocated, as C. Darwin has described in the case of D. chrysanthum ; in 

 D. capillipes, after many trials, I obtained this result once. In all my 

 trials the flower was held in its natural position. If such self-pollination 

 occurs in nature in D. caTpillipes, it must be of very rare occm-rence, and most 

 probably would not be followed by the formation of a capsule. I have had 

 no opportunity of examining D. chrysanthum. 



Dendrobium senile. — In this orchid the filament is elastic ; and the anther, 

 on dislocation, falls downwards and forwards. Still access to the stigma is not 

 completely prevented, as the filament is relatively short, and does not allow 

 the anther to fall down sufiicieutly to entirely block the passage. Each 

 shoot of this species only produces one or two flowers. Frequently a whole 

 aggregation of plants growing naturally may be observed with only one or 

 two flowers ; so it may often happen that a bee, after visiting a flower, returns 

 to that same flower, and pollinates it with its own pollen, having meanwhile 

 made a fruitless search for another flower. Such a flower would then prob- 

 ably produce a good capsule, as in both the flowers which I experimentally 

 pollinated with their own pollen, good capsules were produced. The flowers 

 of this orchid remain fresh for a much longer period than those of any other 

 Dendrobium which I have observed. One of the flowers was self-pollinated 

 after having been in bloom for a month. 



Dendrohium tortile. — The success of self-pollination in this species is 

 somewhat remarkable. The anther is rather small, but still when pushed 

 backwards it adheres to and protects the surface of the stigma. 



Dendrobium crepidattim. — In this species the filament, though elastic, is 

 very slender, and tlie anther is readily knocked off, in which case free access 

 to the stigma is allowed after the flower has been already visited. Not only 

 did self-pollination succeed in the two flowers experimented with, but in 

 another plant of this species cleistogamy occurred, three buds which did not 

 expand forming capsules. In this case I believe the pollen tubes penetrated 

 the rostellum which normally, before the flower opens, comes to lie in close 

 proximity to the poUinia. 



I have examined the flowers of a few species of Dendrobium not 

 belonging to the section Eudendrobium. The following are my results : — 



Dendrobium lasioglossum Eeichb. f. (?), D. draconis Eeichb. f., D. for- 

 tnosutn Roxb. — In these three species the filament is not elastic. When 



