-514 COL. M. J. GODFERY ON THE 



I l):^lif've thai Cephdanthera Is a very ancuMif jronus, wliicli sliows us the 

 ■<^arliosl. inethoJ of cross-pollination in the Orohidiicea^, lj(^Eore a roslelluni 

 luul yet been evolved in that Oi-tler, If it IkkI once possossed a rostelluni 

 .and .suhsofjuently lost it, we should almost certainly be able to find some 

 rudiment still extant, In the same Avay that staminodes exist to the present 

 day in many orchids, incluiling Cephalant/teray to represent suppressed 

 ^stamens. No sig'ii of even a rndimcMitary rostelliim can be found^ nor could 

 Darwin detect any trace o£ the spiral vessel belonging to it in grandijlora 

 i(/, r. pp. 239, 2-12). Further evidence of the antiquity of Cephalantliera is 

 affor<Ied by tlie fact that the poUen-grains are not cemented into tetrads as 

 in almost all other orchids. This seems to indicale lliat it shows us tlie 



:Simph^st primjcval form in which the polhui-grains e\ist(Ml before their 

 .assend)ly Into tetrads, the massing of tetrads into packets, the building up of 

 packets to form polhnia, and the development of caudieles had yot tak(ui 

 .place. 



Ceplialantliera not only shows us how cross-polh'nation was possible before 



:a rostellnm liad yet come into existencCj but it also enables us to form some 



idea as to liow that unicpie organ first came to be evolved* It is easy to sec 



that pollination through insects becoming smeared with tlie A'iscid secretion 



•of tlie stigma had tliis disadvantage, that it weakiuicd tlu? power of the stigma 



to deiach poUen from any poUInia brought into contact with it, especially if 



several insect-visits occurred in succession. If, therefore, tlie upper of tlu^ 



three stigmas, which from its position was most likely to suffer such loss, 



were to secrete more abundant viscid matter (which might very well happen 



through a natural (effort to replace loss), it would be a distinct advantage to 



the plant. Tliis iaicrease of secretion might easily be augmented by natural 



.seh^ction in th<^ course of time, till the upper stigma sc^creted enough viscid 



fluid to form a dist'uitit drop^ when we should have a rostellum in emlu'yo. 



It only th(Mi needs to acquire a very thin covering mendH*an(^ — just enough 



io ].)r(^vent the enclosed viscid matter from drying np — to j)res(Mit us witli the 



simplest form of rostellum, such as we lind in Epipaciis. 



It was unfoi'tunate that Darwin selected gnindijiora for his experiments 



an exceptional species fertilized in an exceptional manner, — and that hn was 



■t.em[)teil to generalize from one species to the whole genus and prononnce 



Cephdantliera a dcgradeil Epipactis, Far from this btiing the case^ Ce/dudan- 



Akera was probably a wcll-establi>hed genus before Epipariis came inio bein<r. 



Cephalaathera was not derived by retrogression from Epipadis^ it is much 



more like!}' that the latter was formerly b^rtilized in some such manner as 



.now nbtains in Ceplialantliera , and has subsequently acquired its rostellum 



Had Darwin been able to exanune ensifoUa and ru!>ra^ botli of which ^ywwy in 



I^ngland, though they are extremely rare and local, lie would assuredly have 



-discovered their simple method of cross-j.iolliiuition. 



