mr r . J 



512 COL. :\I. J. (iODFlT.Y ox THE 



fertilizct] by its own pollen. 



fm 



constantly self-fertilized (Fert. Orel). ISitO, p. 85), and stated (I.r. p. 242) 

 that die rostellum and its vessels -werG completely aborted, and that the 

 poUen-o'rains ^Yere not cemented into tetrads as in almost ;i]l other orchids 

 {I.e. p. 80). Ho expressed his belief that Cephalanthera was a degraded 

 Kpipacth, that is, an KpljMuih which has lost its rostellum, and with it the 

 possibility of cross-pollination by insects. 



In March 1^20 I was at Hyeres, on llio Kiviera, where ensifoUa was 

 fret^nent. On dissecting its tlowers I found that in many cases the pollinia 

 had disajtpenred without leaving a trace behind. Presently I opened a 

 flower in whieh loth pollinia were still in the anther perfect and intact, 

 but adherino- to the stigma were one whole pollinium and portions of 



another. These must evidently have come from some other flower. It was 

 now clear (1) that pollinia are by some moans bodily removed from a large 

 number of flowers, (2) that in some flowcM-s pollen is deposited on the stigma 

 from some outside source. This seemed to indicate that insects visit the 

 flowers, remove the pollinia, and transfer them to the stigma of another 

 flower, but how could this bo possible, in view of the absence of a rostellum 

 to jjrovide the necessary viscid material? 



I next observed that if the anther was pressed back on its hinge, it at once 

 resumed its former position when released. Normally it hvins forward, so 

 that the curved pollinia rest on the edge of the stigma. To ascertain whether 

 they become anchored there by jiollen-tubes, as described by Darwin in th<^ 



¥ 



It iHoved easily. 



carrying- the iiolliniu ^vitli it. These Avore quite free, "without the .slightesi 

 trace of adlie.-^ion to the odoo of the stio;nia. AVlieii I rtdeuseJ the anther, it 



. -- — "^D 



sprung smartly back^ with so niucli ela^jticity that in one catc both pollinia 

 were flung bodily out of tlm flower. AVhat is the object of this elastic 



spring-hinge ? 



Dolpino suggested (Ult. Osserv, snlla dicog. Tart ii, 1875, p. 149) that 

 tlio viscid secretion with which the stigma is always coated might be sufficient, 

 if an insect became smeared with it, to remove the pollinia. To test this 1 

 inserted a dry camel's-hair brush into a flower, and caused it to sweep gently 

 upwards along tlie middle of the stigma and th<^ face of the anther. It came 

 out wMth one wdiole pollinium attached to it by the middle, the ends pointing- 

 outwards. With a lens I could see a little viscid matter olistenins on the 



fry tt> 



brush. But why was onl}' one pollinium removed ? On looking into the 

 flower I found the anther was empt}' — evidently one pollinium had been 

 previously removed. I did the same thing wdtli another flowerjand this time 

 both pollinia "were readily withdrawn. I repeated the cxperiuKMit with 

 flowcT after flower, always witli the same result. To make sure that the 

 removal of tlic pollinia was really due to the viscid matter of the stigma, 



