225 



The cavities of each separate theca — the loculaments — 

 are inverse to the placentas, i. e. the larger the latter, 

 the smaller, the former are. In the lowest sections they 

 are but little circles at large distances and containing 

 only a few pollengrains. 



Quite at the base of the anther they hâve altogether 

 disappeared and with them the pollen. Such is the case 

 with the stamens represented in figures III — VIL But in 

 figures IX— XII the transformation goes on much farther. 



A transition between the two sets is to be seen in 

 fig. IV. The ovules of the two splits touch one another, 

 caused, as is évident, by the anterior loculaments growing 

 smaller compared with the outer ones, i. e. the locula- 

 ments pointing away from the centre of the flower. In 

 fig. IX this shrinking becomes still more obvious, besides 

 this ovules appear on the filament, a phenomenon which 

 I often came across. 



When the anterior loculaments disappear altogether — 

 a not unfrequent occurrence — only one carpel-like part 

 (bu) remains, which part corresponds wit the outer anther- 

 valves with which the two placentas are firmly connected. 

 The figure moreover shows that the style (with the stigma) 

 is continuous from the top of the carpel in its very 

 middle. For clearness' sake we might once more point out 

 that bu is a transformation of the outer anther-valves, 

 belonging of course to différent thecae, whereas the origin 

 of bl is to be found in the inner valves. 



The latter (bi) do not always disappear : in fig. X and XI 

 it is, indeed, much narrower but présent without a doubt. 

 It may even happen that it ends into a style with stigma 

 and also that it is split up into both its components, 

 which lengthen into ribbon-like lobes. But ovules are as a 

 rule absent, because the placentas remain in connection 

 with bu. The more remarkable are therefore the exceptions, 

 of which one is represented in fig. XII. 



