179 



stameDS and three threadlike unwinged styles. Monstrosities 

 examined by Dr. Smith and myself show that in fact some of 

 the petals are really petaloid instead of reseaibling and imi- 

 tating sepals (XI, p. 64). 



Has the flower of Canna now attained its highest possible 

 stage of reduction^ Is it probable that the flower, which in 

 Distemon only shows h labellum, stamen ./; and a winged style, 

 is going to face further reduction in the same or in some other 

 direction? We find some indication in the so called "third'' 

 flower which we made mention of in 1914 (XIII, p. 127). This third 

 flower of the cicinnus, which is commonly rudimentary or 

 altogether absent, turned up in Dr. Smith's private garden at 

 Buitenzorg in quite a sufficient number to permit of a close 

 examination. The majority of the specimens were, apart from 

 the normally constructed perianth, as represented in fig. 19 a and b 

 (taken from XIII). There are in b two pairs of staminodes, placed 

 decussate, of which the outer correspond with the alae a 

 and /? and the inner with x and labellum. In a a similar 

 flower is represented with the omission of « '). Both flowers 

 compared show^ that the notch generally significative of the 

 labellum is as such not fully reliable as it may also be in 

 possession of an outer staminode. I had noticed this before and 

 saw it again a few weeks ago in the garden of Utrecht where 

 a flower expanded ala ^ with a notched top (and a lateral 

 anther). The only thing on which to rely in identifying the 

 parts of the flower is the place which they occupy. 



Of the sixteen specimens of the jthird flower as many as 

 fourteen showed no trace of the fertile stamen, a; being com- 

 pletely regular and not to be distinguished from the labellum. 

 Ala / was wholly absent. Only in the two other flowers it 

 was present, together with a rudiment of the stamen. There 

 is yet another fact worth mentioning in this "third"' flower, 

 viz., the shape of the style. In one case only did this exhibit 

 a complete wing, in one case one of half the ordinary width 



\) In the list of figures of XIII erroneously indicated as lahellum. 



Ann. Jard. Bot. [luitonz. 2^ Ser. Vol. XIV. 1:5 



