86 THE JOURNAL OP 330TANY 



would not otliei*wise have done. By the time the next flower a>)ove 

 it has opened, the viscid gland has disappeared, leaving a small 

 brownish mark which indicates its former position (fig. A 3, r.). 

 Fertilization bv insects does not occur, or if it happens on rare 

 occasions, only minute fragments of pollen can be transferred. 



The anther projects over the upper edge of the stigma for 

 upwards of half its length. The club-shaped poUinia are joined at 

 the apex, whilst their thickened lower ends are slightly divergent, 

 forming an inverted V. Owing to the forward position of the anther, 

 the viscid gland, instead of being immediately opposite the point of 

 junction of the pollinia, as in M. lafifolia, is opposite the V-shaped 

 space between them. As the anther is face downwards in its natural 

 position, when the anther-cells open, the ^xtllinia sink downwards, and 

 their apices not being arrested by the viscid gland, they pass over it, 

 sliding down over the sloping upper edges of the stigma, and finally 

 coming to rest on its frontal viscid surface (fig. A 3, p.). At the 

 same time a marked change comes over them. The}^ increase notice- 

 ably in size, and become fluffy in appearance. An outgrowth of 

 pollen-tubes occurs, causing them sometimes to appear hairy under a 

 powerful lens. These penetrate the stigma, anchor the pollinia, and 

 fertilize the ovary in the usual way. Self-fertilization thus appears 

 to be inevitable, and the subsequent vigour of the capsules shows its 

 effectiveness. 



In E. viridiflora var. diinensis (J. B. 1913, p. 343, and 1918, 

 p. 1) the pollen is so friable that, even before the flower opens, 

 numerous tetrads of pollen, looking like single pollen grains, fall on 

 the lip. I did not observe a single instance of this in lepfocliila. 

 I have seen minute portions of ix)llen adhering to the inner walls of 

 the anther-cells, but as a whole the pollen remains in situ on the 

 pollinia. Hermann Miiller says that in the Westphalian plant the 

 whole pollinium becomes felted together by pollen-tubes, so that in 

 the later stages pollen can only be detached by forcible removal 

 (Verhandl. des N. H. Vereines der preuss. Rheinlands, &c., 1868). 

 The same is the case with our plant. He also says that nothing 

 whatever is to be seen of a rostellura (which with us is visible in bud 

 and early flower), and that the whole pollinia emerge over the edge of 

 the stigma, forming two pyramids, whose bases rest on its upper 

 surface. With us only their upper portions so emerge, their bases 

 remaining behind the stigma. 



There can be no reasonable doubt that E. viridiflora was 

 originally fertilized by insects. The fact that the cup at the base of 

 the lip still glistens with nectar, and the presence of a viscid gland in 

 newly-opened flowers, show clearly that the flower was designed to 

 attract insects. 



Miiller believed that cross-fertilization might occasionally occur 

 with E. viridiflora. He found aphides sucking nectar, in one case 

 with adherent grains of pollen. He also many times observed small 

 insect larvae (Thrips?), sometimes with a few pollen-grains on the 

 head and back (Z. c). I saw similar larvie and aphides in the flowers 

 of the Surrey plant, and, like Miiller, noticed one or two specimens 

 of the latter stuck fast on the stigma, and dead. It is quite possible, 



