few flowers so situated that they can reach the air and these also 

 seem to have acquired the habit of self-fertilization. 



It does not follow that perfect flowers in the submersed 

 aquatics necessarily end in cleistogamy. It does seem, on the 

 contrary, as though there w^ere a general tendency away from 

 bisporangiate flowers, possibly to avoid self fertilization. This 

 is evidenced by the frequent occurrence of rudiments of the 

 suppressed parts in flowers of this habitat. In our flora ex- 

 amples are seen in Elodea canadensis, E. ioensis, Vallisneria 

 spiralis, etc., which are now dioecious but display more or less 

 prominent rudiments of the suppressed stamens or pistils. Cera- 

 tophyllum, Naias, Zannichellia, Zostera, PhyUospadix, Hydro- 

 charis, Enalus, and many other of these submersed seed plants 

 have functionally monosporangiate flowers, a large per cent of 

 them dioecious, with frequent expression of rudiments. 



Myriophylliim spicatum, though flowering slightly above the 

 surface of the water and having a complex inflorescence shows 

 itself to be in transition in an interesting way. The upper part 

 of the spike is composed of staminate flowers, which, as Knupp 

 (8) has sho^^Ti, contain mere rudiments of pistils. These abor- 

 tive pistils are increasingly developed in each lower whorl of 

 the spike until near the middle of the inflorescence they are func- 

 tional in association with the stamens of the perfect flowers foimd 

 there. Below this region the stamens are abortive becoming 

 smaller in each successive whorl and at the bottom of the spike 

 are represented by mere rudiments. There is thus a reciprocal 

 transition from either end of the spike toward the center and 

 the monosporangiate flowers are here without doubt derived 

 from perfect flowers like those still shomng near the center of 

 the series. With these comparisons in mind the close pollination 

 of Heteranthera duhia stands in sharp contrast to the highly 

 specialized devices favoring cross pollination found in many of 

 the submersed aquatics. This failure to overcome the barriers 

 to cross pollination may be due to the structure and arrange- 

 ment of the flowers, as discussed above. On the other hand it 

 would seem that the ease and certainty Avith which close pollina- 

 tion is accomplished may have operated to inhibit floral special- 

 ization. 



Ritzerow (9) has called attention to the frequency with 

 which cleistogamic flowers, either accompanied by chasmogamic 



55 



