sac may be pulled out as a single tangled mass of spores and 

 pollen tubes, many of which probably never find their way out 

 of the anther (Fig. 2). 



Further elongation of the style throw^s it into a contorted and 

 folded form as it elongates in the narrow confines of the space 

 wdthin the closed perianth. Despite the many pollen tubes and 

 the favorable conditions for fertilization, many sterile ovules are 

 found. Probably about fifty per cent of the ovules do not de- 

 velop into seeds but seem to abort immediately following the fer- 

 tilization stage. With many flowers to the plant, however, and 

 an average of perhaps fifteen seeds to the flower, the total output 

 of seeds per plant for the summer is considerable. 



Discussion 



Heteranthera duhia has developed a vegetative body well 

 adapted to the submersed life as sho\Mi by its vigorous and suc- 

 cessful gro^vth. It has failed, however, to attain floral speciliza- 

 tion adequate to insure cross pollination. Wliile the reasons for 

 this failure can not be demonstrated, it seems to the writer that 

 they are correlated with the perfect flowers and the ease with 

 which close pollination could be accomplished. This would sug- 

 gest that cleistogamy in this form is initial and not derived, and 

 that it has perhaps operated to inhibit floral evolution. 



The writer has elsewhere outlined (5) the alternatives pre- 

 sented to a submersed aquatic, one or more of which have to 

 be adopted if it is to avoid self-pollination and still reproduce 

 by seeds. Certain of these should perhaps be mentioned at least 

 in relation to the possible adjustments of Heteranthera duhia. 

 Many of the submersed aquatics with less highly specialized 

 flowers raise an inflorescence above the surface of the Avater at 

 the time the flowers mature. The Potamogetons and Utricularias 

 follow this plan which is essentially a continuation of the habits 

 of the ancestral land plants and permits either perfect or mono- 

 sporangiate flowers to be used in safety. Heteranthera duhia 

 can not adopt this plan as its flowers are scattered and inter- 

 mixed with the vegetative leaves and there is an indefinite flower- 

 ing period. RanuncuJus aquatilis with flowers similarly scat- 

 tered along the axis, gets its blossoms into the air by developing 

 a floating stem, or one that is entirely free floating, Avhich lies 

 horizontal so near to the surface that a slight elevation raises 



53 



