r". 



■"' -■ "iT.ITfft 



\ 



h- 



%v 



510 THE FERTILIZATION OF CErHALANTHKRA. 



■ ciiso tlmt it Is entirely self-fertilized, iiml that we have not siinply so far failed 

 to understand the tneehanistn of the fiDU'er/' I had then had no opportunity 

 of stud vino- tlie fertilization of enslfolia und rubra. Now tliut I have doijo 

 so T am convinced that both these species are wholly cross -pollinated by 

 insects, and that this is also the case occasionally with grandiflora, though 

 its subsequently acquired faculty of self-fertilization has now become the 

 dominant factor in its reproduction. 



I do not now believe that there Ijns been iuiy decadence or degeneration in 

 Ceplialanthera, but thatjt presents a case of persistence to the i)resent day of 

 an extremely ancient method of cross-polllnntion, whicli possibly prevailed 

 universally in the Orchidacese (except in Diandric) in the remote period 

 before a rostellum had been evolved in thut Order. 



\ 



