» 



50 



thereafter; the first fruits from these have been found in 

 June. Ihe emergent, dicismogamous f lowers begin to blocro 

 when the stems reach the water surface, and are usually 

 conspicuous fron late June until August. Seed dispersed 

 largely takes place fron mid- to late summer. In 

 Washington, the sites are lower in elevation, and emergent 

 flowering begins during May (J. Gamon, pers. cxum.). In 

 Idaho in 1988, during v*u.ch near average climatic ocMiditions 

 occurred during the ^ring, cleistogamous flowers were in 

 bud c»i unbranched, submerged stems on 6 May. Plants were in 

 flower above the water surface on 14 June, and cleistogamous 

 fruits were near maturity. 



2. Relation to clinate and microclimate: Because H. aquatilis 

 is an aquatic species largely restricted to vernal ponds and 

 wetlands, its phenology is intimately tied to the climatic 

 factors influencing these habitats. These factors would 

 include precipitation (e^iecially winter sncwpack and 

 subsequent run-off, and spring rains) and summer weather 

 patterns. The current drought oonditioTs in the Pacific 

 Northwest have resulted in an eaurlier drying of some of the 

 habitats in Montana. A subsequent reduction in the total 

 amount of seed production would be expected, since the 

 actual duration of the plants and flowers would be shorter. 

 In Washington, the current drought ocxiditions have resulted 

 in some ponds remaining dry (or at least without pc^ided 

 water) throughout the ye2ur (J. Gamon, pers. oomm.). 

 However, drou^t cc^iditions experienced in northern Idaho 

 during the winters of 1986-87 and 1987-88 did not appear to 

 affect the water level of the pond; it was at high water 

 nark. 



D. Reproductive biology. 



1. IVP^s of reproduction: The breeding system of H. aquatilis 

 has been studied by Lesica et ai. (1988) . Anatomical 

 studies showed that in the cleistogamous flowers, the 

 corolla develops a small closed bud amd then drops off, 

 leaving an enlarging ovary. Although the chasmogamous 

 flowers develop fully, anther dehiscence and embryo 

 development before the flowers had opened was repeatedly 

 observed. In these flowers, as the corolla opens the stigma 

 pushes vp throu^ the filament tube in close proximity to 

 the dehiscing cinthers; this sequence would almost eissure 

 self-pollination if it had not previously taken place. No 

 evidence of agamospenty was observed; in both cleistogamous 

 and chasmogamous flowers, embryo and/or endosperm 

 develc^xnent was observed only after penetration of the ovule 

 by a pollen tube. AdditicMially, pollen stainability of 

 samples from the Condon Creek site in Montana was 93% 

 (s.d.=3%) , indicating normal fertility. All of these 

 (±)servations suggest that, althou^ not iirpossible, the 

 occurrence of outcrossing in this species is probably 

 extremely restricted, and that the breeding system 



