176 



left side of x '). To this extent Eichler might be in the right 

 but as we have already seen not in respect to the calyx and 

 the corolla. 



How then explain an irregularity which after Eichler would 

 be a unicum in the Vegetable Kingdom? 



I venture to propose the following solution. Starting from 

 the fact that the plan of the flower is completely regular and 

 that in the flowertube (fig. 6) the vascular bundles indicate 

 the presence of three styles and three exterior alae each with 

 two bundles, we assume for simplicity's sake that in both 

 flowers of a cicinnus the ala « is directed downwards (fig. 18). 

 (The real direction in to be seen in fig. IB, VI, p. 172). 



The flowers being antidromous, s and .s^' make reflected images, 

 but S occupies the top in both figures. As .v and labellum show 

 little essential difference it seems admissible that in the second 

 or left flower, ^^ takes the labellum-shape and the labellum the 

 .^'-shape. And further that the anther is in A a product of y 

 but in B of d, and finally that the wing of 'S' in both flowers 

 grows in the direction of the staminode which it links to the 

 fertile stamen. 



In order to explain this hypothesis and make it admissible 

 I would recall the remarkable fact described by Fritz Muller 

 about the synema (labellum) in a species of Alpinia (VIII). 

 From his observation (which Dr. SairrH and myself were able 

 to confirm'^)) it appears that whenever the synema (labellum) 

 takes the superior or the inferior place in the flower it assumes 

 in former case the shape of a labellum but in the latter it 

 breaks up in its components : two antipetalous staminodes + 

 one antisepalous. 



Why should we not in the case of .^' and labellum of Canna 

 suppose the same interchange? 



1) In Fig. 5, p. 31 (XII) the anther seems to be on the right. This could be 

 explained by supposing either s' to have developed to a fertile style instead of S 

 or X and anther having exchanged fertility and sterility. 



2) Ann. du Jard. bot. de Buitenzorg, 2e Serie, IX, p. iOl. 



