Water loss through inflorescences may place extreme demands on plant water status in arid environments. Here we examine how corolla size, a trait known to influence pollination success, affects the water cost of flowering in the alpine skypilot, Polemonium viscosum. In a potometry experiment, water uptake rates of inflorescences were monitored during bud expansion and anthesis. Corolla volume of fully expanded flowers predicted water uptake during bud expansion (R =0.61, P=0.0375) and corolla surface area predicted water uptake during anthesis (R=0.59, P=0.044). To probe mechanisms underlying the relationship between corolla size and water uptake, cell dimensions and densities were measured in several regions of fully expanded corollas. Corolla length was positively correlated with cell length in the middle of the corolla tube and cell diameter in the corolla lobe (Pearson's r from 0.26-0.33, n=86, P ≤ 0.05). Cell density was negatively correlated with cell dimensions in the upper corolla tube and lobe (Pearson's r from -0.39 to -0.42, P ≤ 0.0015). These findings suggest that more water may be required to maintain turgor in large corollas in part because their tissues have lower cell wall densities. The carbon cost of water use by flowers was assessed in krummholz and tundra habitats for P. viscosum flowering, respectively, during dry and wet portions of the growing season. For plants in full flower, average leaf water potentials were significantly more negative (P=0.0079) at mid-day in the krummholz (June) than in the tundra (July), but were similar before dawn (P=0.631). Photosynthetic rate at the time of flowering declined significantly with increasing corolla size in the krummholz (P=0.0376), but was unrelated to corolla size on the tundra (P>0.72). Plants losing water through large corollas may close leaf stomata to maintain turgor. If photosynthesis limits growth in this perennial species, then the water cost of producing large flowers should exacerbate the cost of reproduction under dry conditions. Such factors could select for flowers with smaller corollas in the krummholz, countering pollinator-mediated selection and helping maintain genetic variation in corolla size components of P. viscosum.

We designed field experiments using solar-tracking Ranunculus adoneus flowers to determine where photoreception occurred, which organs responded, and how movement was achieved. Flower peduncles bend eastward in the morning and gradually unbend over the course of the day. Peduncles were found to bend significantly more frequently in the middle region near the floral bracts, 1-3 cm below the flower, than elsewhere on the peduncle. Because the peduncle tip continued to track the sun even after the flower itself was removed, our experiments concentrated on shielding (or conversely, exposing) various portions of peduncles from (or to) sunlight. Photoreception occurred primarily in the portion of the stem just beneath the floral receptacle. By following the position of landmarks applied to the stem, we found that 40% more growth occurred on the shaded side of bent peduncles, compared to the sunlit side. In contrast, top-shielded peduncles did not solar track well and grew only 25% more on the shaded side than on the sunlit side. This growth differential corresponded to differences in cell length on the two sides of bent peduncles, with significantly longer epidermal cells occurring on the shaded side than on the sunlit side.Key-words: Ranunculus adoneus; snow buttercup; heliotropism; phototropism; solar tracking. INTRODUCTIONHeliotropism, the daily movement of organs to follow the sun, is usually differentiated from phototropism, which is merely turning towards a fixed light source, not tracking it (Hart 1990). Occasionally, it can be difficult to distinguish between the two, as evidenced by the long debate as to whether sunflowers, Helianthus annuus, actually track the sun or are fixed facing east (Kellerman 1890;Schaffner 1898Schaffner , 1900Polikarnov 1954;Shibaoka & Yamaki 1959;Morozov 1963). If a similar mechanism drives both flower heliotropism and the well-studied phenomena of seedling phototropism, then the dichotomy between the two processes may be trivial. However, very little is known of the physiology or mechanism of solar tracking in flowers.Solar-tracking movements of flowers and leaves have been known for over a century (Weisner 1879;Darwin 1880). Leaf heliotropism occurs in at least 16 plant families (Ehleringer & Forseth 1980), while floral heliotropism is mainly restricted to the Asteraceae, Papaveraceae, Ranunculaceae, and Rosaceae (Hocking & Sharplin 1965;Kevan 1975;Knutson 1981). Leaf heliotropism maximizes light interception under shady conditions or short growing seasons (Mooney & Ehleringer 1978;Ehleringer & Forseth 1980). Most reports of solartracking flowers come from arctic or alpine plant species (Hocking & Sharplin 1965;Kevan 1975;Knutson 1981). In many such species, heliotropism raises flower temperatures, enticing insects to remain, basking and foraging, enhancing their effectiveness as pollinators (Hocking & Sharplin 1965;Kevan 1972;Stanton & Galen 1989;Kudo 1995). In addition, heliotropism increases light interception by photosynthetically active floral parts (most often green ovary ti...