— 145 - 



wards leaves m^, catches hold of m^, and clings to it, feeding 

 itselfby tlie pollen of m\ and trying (probably successfully) to 

 squeeze the juice out of it. 



In the next flowers it touches the stigma before the anthers 

 andproduces cross-fertilisation. Very often the Syritta proceeds 

 irregularly.creeping in various directions over the different parts 

 of the flower. 



5. Eristalis tenax feeds itself by the pollen of m^. 



6. Bombus muscorum proceeds in the same way as the hive- 

 bee, We could not observe N»^ 5 and 6 accurately. 



7. Pieris Napi tries to suck nectar, apparently without 

 success. 



The function of the three kinds of staraens is therefore the follo- 

 wing: l°the stamens m- are the fertilizing ones.2'' the pollen of m^ 

 is coUected or devoured by thevisitors; a sraall quantity of it is 

 kept for the fertilisation. 3» the very small supply of pollen 

 produced by the anthers m' is devoured or coUected by insects, 

 but the connectives of those anthers are large, and transforraed 

 into juicy organs, which take the place of the wauting nectaries. 



We may supposethat the flowers of the ancestors of Commelyna 

 were regular, resembling the flowers of Tradescantia virginica, 

 Gradually the stamens became differentiated into two sets : three 

 (superior) shorter ones, the nutritive anthers, and three (inferior) 

 longer ones, the fertilisiug anthers. In this stadmm the flowers 

 were nearly constructed like those of Tinnantia undata. 



In the latter species (according to H. Muller) the median-inferior 

 stamen (corresponding to m' in Commelyna) is as long as the 

 two lateral-inferior ones; it lies uuderthe style, almost quite out 

 of the reach of visitors, and is therefore rather useless. Each 

 superior stamen bears upon its filament a brush of yellow hairs, 

 which makes it conspicuous. 



In this stage of evolution, the flowers of the ancestors o 

 Commelyna were already adapted to long-tongued bees(l). But 

 we know that Apis and Bombus (H. Muller, fert, of flow., 



(1) In the same way as Solanum rostratum etc. — See J. E. Todd, 



on the flowers of Solanum rostratum and Cassia chamaecrista.Americ. 



Naturalist, 1882, p. 231, with woodcuts. 



10 



