Disturbance and Predictability of Flowering Patterns in Bird-Pollinated Cloud Forest Plants

Abstract

The distribution and flowering patterns of hummingbird—pollinated plants were compared from July 1981 to June 1983 in three patch types in cloud forest at Monteverde, Costa Rica. Study plots were: (1) four recent, large (1100—2500 m2) disturbances ("cutovers") produced by cutting vegetation, (2) six recent, smaller (200—600 m2) disturbances caused by treefalls, and (3) four plots (1600—1800 m2) of canopied forest. Based on published literature dealing with communities that characterize different regimes of disturbance, we tested one assumption and two hypotheses. Assumption: Plant species composition differs among the three patch types. Hypothesis 1: Phenotypic specialization by plants for co—evolved interactions with hummingbirds will be lowest in large gaps, highest in forest, and intermediate in treefalls. Hypothesis 2: Predictability of flowering phenologies and nectar production will be lowest in large gaps, highest in forest, intermediate in treefalls. Neither the assumption nor the hypotheses were supported by the results. The patch mosaic in this cloud forest was not associated with major differences in species composition of bird—pollinated plants. Most species studied were self—compatible. Most abundant in cutovers were species with long corollas, relatively specialized for attracting long—billed hummingbirds. Species with short corollas, which can be visited by many hummingbird species and some insects, were most abundant in treefalls and forest. Variation in phenological patterns showed no consistent trends among patch types. Predictability of flower and nectar production tended to be greatest in treefalls, which are foci of concentrated flowering activity by all species. Discrepancies between our results and previous studies can be ascribed to two facts. (1) Much of the literature dealing with ecological consequences of disturbance has dealt with large—scale anthropogenic disturbances such as old fields of the eastern USA, whereas we studied small, natural, or quasi—natural disturbances. (2) Studies of forest disturbance have focused on the tree layer, whereas we studied the understory herbs, shrubs, and epiphytes. Natural disturbance usually involves death and replacement of one or more trees, whereas individuals of other life forms may persist through the disturbance.