265 



pollen-cells, but so far there is no connection between the two 

 elements. 



The filament of the anther is winged and the wings are 

 produced vertically upwards into two linear structures of 

 varying shape, which bridge the space between anther and 

 stigma, and protect the pollinia after dehiscence against dis- 

 placement and the drying influence of the air. 



Dehiscence is now about to take place, and the rostellar- 

 disc begins to secrete freely. The anther is rigidly supported 

 behind by the fleshy curved base of the dorsal sepal. In front 

 and at the sides, the whole of the sexual apparatus is embraced 

 by the still erect labellum and its lateral lobes ; the raised 

 line or lines at the base of the lamina are exerting an in- 

 creasing pressure against the stigmatic-plate in their pro- 

 gressive development, thus forcing the sexual elements of the 

 flower more closely together. Dehiscence occurs, and for the 

 first time union is established between the male and female 

 elements of the plant. The apices of the pollinia become 

 adherent to the back of the rostellar-disc. There is no 

 caudicle. The flower expands; the insect-visitor removes the 

 "viscid disc" together with the attached pollinia, and cross- 

 pollination is then effected in the usual manner. 



It is interesting in this connection to examine the 

 gynostemium of Prasophyllum, which has reached a higher 

 degree of specialization than that of Diuris. Here the fila- 

 ment of the anther is still retained, but it has become adnate 

 to the base of the style, so as to form a very short column. 

 The wings of the filament are very similar to those in Diuris, 

 but in most members of the genus they have become adnate to> 

 the sides of the stigmatic-plate, so as to still further con- 

 solidate the column. The pollinia are composed of sectile 

 pollen, and they form an attachment to the "viscid disc," 

 not directly as in the case of Diuris, but through the inter- 

 vention of a well-marked caudicle. The structure of the 

 column is most easily studied in P. elatum, in which the 

 flowers are relatively large. In this species the wings of the 

 anther-filament have not yet become adnate to the stigmatic- 

 plate. 



Prasophyllum is almost exclusively an Australian genus, 

 its only extension of range being to New Zealand. 



These two genera do not appear to have developed along 

 the usually accepted lines of evolution, which are presumably 

 as follows : — 



1. Suppression of some of the sexual parts and confluence 

 of the rest to form a column. 



2. Aggregation of the pollen into pollen-masses. 



3. Formation of a rostellum. 



