CEPHALAT^TTTEUA OR EPTPACTTS 71 



rangular" stigma of G. cucullata. We shall see later that that 

 species is untrustworthy as evidence of the existence of a quadrangular 

 stigma in the genus Cephalanthera, so that IJichard's diagnosis in 

 this respect still holds good. The explanation as to the pollen-grains 

 appears to be somewhat disingenuous. The tetrads of pollen in 

 Cei)lialantliera resolve into separate grains in sitic^ in Ejjipactis they 

 do not then separate at all. Only when they come into contact with 

 the viscid secretion of the stigma do they swell and disintegrate, as 

 is usually the case in the Orchidacese. Disintegration before polli- 

 nation is a very different thing from disintegration brought about by 

 the action of the stigmatic fluid. Even assuming that the author is 

 correct in his statement that in Cephalanthera the pollen-grains are 

 originally bound together in tetrads, the difference in the pollen- 

 grains pointed out by Richard is undiminished in imj)ortance. 



He next deals with the new character adduced by Keichenbach 

 fil. — the absence of a viscid gland in the case of Cephalanthera, its 

 presence in Epipactis. He saj^s, in effect — I could reasonably pass 

 over this character as of very little value, if indeed the very mention 

 of it does not prove the difficulty of fixing a boundary between the 

 two genera. Keal viscid glands, i. e. a thorough transformation of 

 the tissue of the rostellum contiguous with the anther into a viscid 

 gland, certainly occur in some species of JEpipacHs, yet it generally 

 decreases with the reduction of the rostellum, the species with a 

 short rostellum (e. g. E. palustris) show only small viscid glands, 

 whilst the development of the latter in species of Cephalanthera with 

 a scarcely discernible rostelhun entirely fails to take place. 



In this the learned author shows an absolute failure to understand 

 or appreciate the capital importance of the viscid gland, which is the 

 most outstanding and weiglity character in Orchidaceae, is peculiar to 

 that order, and is the root-idea of its floral construction. The one 

 unique character of the Orchidacese, which puts them on a different 

 plane from all other natural orders (except perhaps the Asclepiadaceae), 

 is the extremely ingenious mechanical device by means of which 

 insects are unconsciously induced to convey the pollinia from one 

 flower to the stigma of another, usually that of a separate plant. 



One of the most remarkable things in the morphology of the 

 order is the absence of a rostellum in Cephalanthera^ an absence 

 which reduces that genus from the exalted position of possessing one 

 of the most ingenious contrivances for the transport of pollen in the 

 vegetable kingdom, to the status of ordinary self -fertilized plants. 

 Self-fertilization is the simplest of all conceptions — it needs no honey, 

 no scent, no colour, no co-operation of insects, or even of wind or 

 water. Cephalanthera is a decadent genus, which has fallen from its 

 high estate, assuming that it is really the case that it is entirely 

 self -fertilized, and that we have not simply so far failed to understand 

 the mechanism of the flower. It was evidently originally designed 

 for insect-fertilization. From a biological point of view it is now 

 more widely separated from JEpipactis than is the latter from genera 

 possessing a rostellum. The presence or absence of a rostellum is of 

 much greater importance than the presence or absence of caudicles, — 

 though these latter characters are used to differentiate the two great 



