158 MR. R, TBIMEN ON THE STRUCTURE OE BOKATEA SPECIOSA. 



and subconvergent. Its apex is similarly produced, but the pro- 

 cess is shorter, and so recurved as to leave the front of the hood 

 widely open*. The membrane at the anterior edge of the lateral 

 horns is folded back throughout, forming grooves in which the 

 caudicles of the poUinia (enveloped to their extremity in a conti- 

 nuation of the thin divided investing membrane of the pollen- 

 masses) lie. The two disks are situated upon the apices of these 

 horns, not far apart, with their viscid surfaces facing rather back- 

 ward and towards the sides of the helmet. On each side of the 

 base of the helmet, a little below the external edge of the groove 

 containing the caudicle, is a small, thin, rounded winglet, which 

 Is an expansion of the membrane of the rostellum. These wing- 

 lets evidently serve as guards against the escape of the very long 

 and elastic caudicles from the grooves, which just there are 

 widest and extremely shallow. It is very noticeable, that, imme- 

 diately behind the winglets, the Investing membrane of the cau- 

 dicles (which lies free in the grooves) is firmly coherent with the 

 column. The two widely-separated stigmatic surfaces are seated 

 at the apices of a pair of wonderfully elongated white processes, 

 having a common origin back in the neck, and at the base of the 

 helmet, immediately behind the orifice of the nectary. These pro- 

 cesses, cylindrical and apparently solid, are in reality tubular 

 the edges of each flattened expansion curling upward, and meeting 

 very closely for about two-thirds of their length. Beyond this, 

 the line of contact Is marked, and the edges soon after diverge 

 widely at the somewhat spoon-shaped and slightly concave surface 

 of the stigma. The upper surface of the processes (forming the 

 lining of the tubes) is very smooth and glossy, but not at all 



viscid. 



struc 



ture described to the removal of the poHinia and the subsequent 

 fertilization of one or more flowers. An insect approaching the 

 flower finds the opening of the nectary surroxmded and guarded 

 by the helmet-like rostellum. It seems probable that, as Mr. 

 Darwin has shown in the case of Angrcecum sesqtiipedale, the 

 length of the nectary and the amotmt of nectar it contains are so 



hungry 



um 



and in such a case the attachment of one or both of the viscid 



m ■ 



* The outer surface of the hehnet (except the anterior portion of the lateral 

 horn?), as tvoII as the iipper surface of the recurved horn, is very minutely 

 "shagrccned " or roughened with points. 



