96 ■ MISSOURI BOTANICAL GARDEN. 



and the other bright green pollen*. He also suggests here and 

 in his "Different Forms of Flowers," that L. indica may be a 

 heterostyled formf. Fritz MiillerJ fertilized a Lagerstroemia, 

 infertile to its own pollen, with yellow and green pollen from 

 another variety or species in other gardens. Both kinds pro- 

 duced apparently good seed, though only some of those from 

 the green pollen germinated. In his paper on dimorphism 

 in the anthers of pollen flowers, Hermann Muller§ assigned 

 the same interpretations to the structure of Lagerstroemia as to 

 other forms in which there is a color dimorphism in the anthers, 

 namely, that the central yellow anthers serve as organs of 

 attraction and furnish booty to the visitor while the blue 

 or green anthers, which are not observed by the visitors be- 

 cause of their resemblence in color to the background formed 

 by the petals, furnish pollen for fertilization. Koehnet sug- 

 gests that in respect to this character the genus Lagerstroemia 

 is still in the process of evolution, since the forms may be 

 arranged in a complete series from those in which there is no 

 differentiation in the androecium to those in which the 

 dimorphism of the stamens is most highly developed. The 

 reference to green stamens by some writers is apparently a 

 slip of the pen. Knuth in his Handbook reviews Miiller's 

 work. Koehne in his recent monograph suggests the possi- 

 bility that the forms in which the stamens arc not dimorphic 

 have been reduced from those in which the dimorphic condi- 

 tion prevails, but he inclines much more to the opinion that 

 the dimorphism is in a state of evolution, supporting this 



* I have seen several stamens intermediate in size between the typical 

 large and small. Darwin noticed one petal with a furrow near the base 

 of the limb containing pollen. In several flowers I found one or all of 

 the large stamens with the anthers more or less completely metamor- 

 phosed into the orbicular crispate limb, while the filament represented 

 the claw of the typical unguiculate petal. I mention these cases because 

 of the similarity of the filament of the large stamens to the claw of the 

 petal. 



t Koehne states that all species are homostylous. The styles of all 

 the flowers I have seen are of sensibly the same length. 



t MuUer, F. Nature. 27 : 364. 1883. 



§ Muller, H. Kosmos. 13:254.1883. 



11 Koehne, E. Engler's Bot. Jahrb. 6 : 45. 1885. 



