Flexible pedicels and hummingbird pollination of jewelweed (Impatiens capensis).

Most biology textbooks treat hummingbird-flower pollination statically in terms of a match between bill and flower dimensions: the bird simply inserts its bill into the flower, and pollen is picked up and delivered, much like putting a sword into a sheath. Studies in my lab indicate that many hummingbird flowers move during hummingbird visits, and that these movements enhance flower pollination while either raising or lowering the costs of feeding for birds, depending upon the flower species.

Consider jewelweed (Impatiens capensis), a wildflower native to eastern North America. The plants have orange-yellow flowers which resemble miniature "horns of plenty", and hang from flexible pedicels, or stems. Earlier researchers have hypothesized that the flexible pedicels of hummingbird flowers were an adaptation which prevented the flowers' ovaries from being pierced by the hummingbird's bill. The problem with this explanation, at least for jewelweed, is that the ovaries of the flower aren't located anywhere near to where the bird inserts its bill, and moreover, no one has found any evidence that hummingbirds damage the ovaries of flowers with their bills.

Our films indicate that flexible pedicels may be a general mechanism for hummingbird-flower pollination. When a hummingbird feeds at a jewelweed flower, each lick of the tongue into the curved nectar spur of the flower pushes the flower back, and then the spring-like pedicel pulls the flower forward, smearing pollen on the upper bill of the bird (see video clip).

To investigate whether the flexible pedicel and movements of jewelweed flowers enhanced their pollination, Allen Hurlbert, Aki Hosoi, Paul Ewald, and I experimentally immobilized a sample of jewelweed flowers by securing small pieces of wire around the pedicels of flowers so they couldn't move. We found that natural, mobile flowers deposited approximately 50% more pollen onto their hummingbird visitors than experimentally-immobilized flowers, but that mobile flowers significantly increased the foraging time costs of birds, by about 25% in comparison to immobilized flowers (see 1996 Oecologia 105:243-246).

Jewelweed flowers are also pollinated by bees, and our studies indicate that these flowers do not move during bees' visits, implying that mobile flowers in this case may be an adaptation for hummingbird pollination. I am presently investigating the extent to which other floral traits of jewelweed reflect adaptations for bee versus bird pollination. For example, jewelweed flowers exhibit considerable variation in the curvature of their nectar spurs, and our studies suggest that the maintenance of such variation may be a consequence of differences between bees and hummingbirds in pollination and feeding efficiencies at spur types.