Chap. VIII. CLEISTOaAMIC FLOWERS. 331 



fertile stamen ; the style is almost obsolete, with the 

 three stigmatic surfaces directed to one side. Both 

 the perfect and cleistogamic flowers produce seeds.* 



The cleistogamic flowers on some of the Mai- 

 pighiacese seem to be more profoundly modified than 

 those in any of the foregoing genera. According to 

 A. de Jussieut they are differently situated from the 

 perfect flowers ; they contain only a single stamen, 

 instead of 5 or 6 ; and it is a strange fact that this 

 particular stamen is not developed in the perfect 

 flowers of the same species. The style is absent or 

 rudimentary ; and there are only two ovaries instead 

 of three. Thus these degraded flowers, as Jussieu 

 remarks, " laugh at our classifications, for the greater 

 number of the characters proper to the species, to the 

 genus, to the family, to the class disappear." The 

 calyces of the perfect flowers are studded with glands, 

 and their absence on the cleistogamic flowers may 

 probably be explained by an observation of Fritz 

 JU tiller, who informs me that in the one species, 

 Bwnchosia Gaudichaudiana, the fertilisation of which 

 he has often witnessed, the perfect flowers are regu- 

 larly visited by bees belonging to the genera Tetra- 

 pedia and Epicharis. These bees sit down on the 

 flowers, gnawing the glands on the outside of the 

 calyx, and in doing so the under sides of their bodies 

 are dusted with pollen, by which- afterwards other 

 flowers are fertilised. Such visits to the cleistogamic 

 flowers would be useless. 



As the Asclepiadous genus Stapelia is said to pro- 

 duce cleistogamic flowers, the following case may be 

 worth giving. I have never heard of the perfect flowers 

 of H(yya carnosa setting seeds in this country, but some 



* Dr. Kirk, ' Journ. Linn. Soc' f ' Archives da Museum,' torn, 



vol. viii. 1864, p. 147. iii. 1843, pp. 35-38, 82-86, 589, 598. 



