170 



flowertube. At its base it is rather narrow, then it widens but 

 after that again becomes very narrow. This structure has been 

 represented in fig. ISA and B showing the two halves obtained 

 by a vertical section of labelluni and x. It shows at the same 

 time in which places x and labellum spring from the tube. 

 The latter is filled with honey to the brim (which is most 

 probably produced by the innerwall) and is only accessible 

 through a very narrow cleft. The question arises whether, apart 

 from the Marantaceae, there exist other flowers of this con- 

 struction, viz. flowers in which the flowertube is formed by 

 the apical portions of the carpels. Systematists may be pre 

 pared to answer this question. 



The fact that the rich store of honey is almost shut off calls 

 to mind a remark by Schumann about the rarity of insects 

 visiting Canna-flowers along the way that leads to pollinisation 

 of the stigma and on the other hand the frequency of burglary 

 at the base of the flower '). My own observations with regard 

 to pollinisation in Canna are limited. In the botanical garden 

 of Utrecht I noticed in the conservatory, which chiefly contained 

 the numerous specimens of this genus, only a couple of bees, 

 one wasp and two Syrphus-flies but remained in doubt whether 

 or not they would flod an opportunity in the very wide mouth 

 of the flower to convey the pollen from the broad side of the 

 style to the stigma or the stigmatic stripe (Eichler's Narben- 

 streife). In our fig. 3 we see how all the pollen produced by 

 the anther is being deposited on the wing or broad face ot 

 the style which takes place before the opening of the flower. 

 In ordinary flowers the style is to be considered double and 

 consisting of the real threadlike style with a two-lobed stigma, 

 as in figg. 16 {s') and 23 {st), and secondly of a wing fig, 23, id) of 

 much thinner substance growing out from its left side. This wing 

 bears a complementary stigma <? in the shape of a very thin and 

 simple ridge. The difference between these two divisions may be best 



i) Fritz Muller, however, observed humming birds visiting Canna in South- 

 America (IV p. 230 footnote). 



