EPIPACTTS VTKIDIFLORA 35 



the former threaten the continuity of the species. Consider, for 

 instance, the anther in Anacaniptis ])iirainklaUs^ which has to 

 be placed witli such unfailing accuracy that the caudicles of the 

 poUinia shall find their way to and attach themselves exactly to the 

 minute viscid disc of the rostellum (which is part of the stigma). 

 Tlie viscid disc is saddle-shaped, and when withdrawn by an insect 

 curls round its proboscis, therebj^ causing the pollinia to diverge, so 

 that, after the movement of depression has taken place, they are 

 pointing forward and outward at such an angle that their tips will 

 touch the stigmas when the proboscis is inserted into another tlower. 

 In this plant there are two stigmas separated by a median spice. If 

 the ends of the caudicles fasten themselves to the viscid disc too close 

 together, they will not diverge enough when the disc curls round the 

 proboscis, and would strike the median space between the two stignras. 

 If, however, they are cemented too far apart, they would diverge too 

 much, and would touch points outside the effective stigmatic area. 

 This example is enough to show the necessity of accurate standardiza- 

 tion in the reproductive parts of the tlower. Any new departure in 

 these essential organs is therefore of paramount importance. If 

 differences in the leaves, bracts, and floral envelopes are sufficient 

 ground for specific distinction — and they are commonly so regarded — 

 diiferences in the essential mechanism of reproduction are of still 

 gi'eater weight. 



If adequate consideration is given to the above-mentioned morpho- 

 logical differences between IE. latifolia and E. viridijlora, i\\ej 

 should be amply sufficient to prove that these plants are specifically 

 distinct. The clearness of view of the elder Reichenbach is all the 

 more remarkable, in that he recognized this apart from the evidence 

 of the structural differences in the reproductive organs. It is curious 

 that the younger Reichenbach, who made such excellent anatomical 

 drawings of so many orchids, should have omitted to do so in the 

 case of E. viridiflora. He does not seem to have suspected any 

 difference of structure as compared with E. latifolia. 



Moreover, these morphological differences involve remarkable 

 functional changes. E. latifolia is fertilized by wasps, to whose 

 forehead the viscid gland becomes firmly attached, so that the pollinia 

 are bodily and cleanly withdrawn, and carried to another flower. The 

 pollen is not friable, and only when it comes in contact with the 

 viscid stigma of another flower are fragments detached. Self-fertili- 

 zation appears to be impossible ; the pollinia lie at the back of the 

 stigma, and can only be removed by some external agency. If not 

 withdrawn, they wither in situ. 



In E. viridiflora var. leptocliila the viscid gland is present in 

 the newly-opened flower, but it is inoperative. If touclied with any 

 suitable object, it does not adhere to it as a whole ; a few strings of 

 viscid matter can be withdrawn, but these are far too weak to remove 

 the pollinia, which ordinarily do not come into contact w4th the ros- 

 tellum. Once, indeed, I succeeded in removing a small fragment of 

 pollen adhering to the viscid gland, but I am inclined to think that 

 in touching the gland with a pencil I probably pressed it out of 

 place, and caused it to come in contact with the pollinium, which it 



1)2 



