Floral Neighborhood and Pollination Success in Four Hummingbird-Pollinated Cloud Forest Plant Species

Abstract

In a cloud forest at Monteverde, Costa Rica, was examined pollen loads received by self—compatible flowers of two pairs of plant species pollinated by hummingbirds: Hansteinia blepharorachis and Razisea spicata (Acanthaceae), and Besleria triflora and Drymonia rubra (Gesneriaceae). Each pair consisted of one species (Hansteinia or Besleria) pollinated by short—billed hummingbirds and a related species (Razisea or Drymonia) pollinated by long—billed hummingbirds. At three different times per species, separated by 1—3 mo, we examined flowers on 28—40 focal plants from a wide variety of floral neighborhoods, ranging from plants isolated from conspecifics, either by distance or by other flowering species pollinated by the same hummingbirds, to plants surrounded by conspecifics. Because short—billed hummingbirds often restrict foraging to areas of high flower density, and because short—tubed flowers adapted for hummingbirds often have similar pollen placement, we predicted that short—tubed flowers isolated from conspecifics would receive fewer conspecific grains and more heterospecific grains than short—tubed flowers surrounded by conspecifics. Because long—billed hummingbirds often forage over large areas and because long—tubed flowers adapted for hummingbirds tend to diverge in pollen placement, we predicted that pollination of long—tubed flowers would be relatively unaffected by floral neighborhood. Effects on pollen loads of floral neighborhood (nearness to or isolation from other flowers) followed few patterns consistent with our prediction or with conventional theory. (1) There were no consistent effects of floral neighborhoods on numbers of heterospecific grains deposited on stigmas; in all four species, regardless of corolla length, effects of particular neighborhood variables (as determined with stepwise multiple regression) were as likely to run exactly counter to conventional models as to corroborate models. (2) In none of the 12 sampling runs did increases in absolute densities of neighboring heterospecific flowers adversely affect pollination. (3) However, in two runs, loads of conspecific grains increased with increases in the absolute density of neighboring conspecific flowers, and/or (in three runs) with increases in their relative density (proportion of conspecifics among neighboring flowers). These runs all involved short—flowered species rather than long—flowered species, tending confirm our initial prediction, but half the sampling runs, even of short—flowered species, failed to show any density—dependent effects from neighboring flowers pollinated by the same hummingbirds. Flowers frequently received fewer conspecific grains than they had ovules to be fertilized. Therefore the potential existed for floral neighborhoods to affect seed set and fitness of plants. Nevertheless, even though neotropical hummingbird—pollinated flowers have been cited as examples of species whose flowering peaks are displaced through competition for pollination, competitive effects from neighboring heterospecific plants were only sporadic in the species we examined, and were particularly infrequent in those species with long flowers adapted for long—billed hummingbirds.