Temporal and Spatial Variation in the Phenology of Flowering and Fruiting in a Tropical Rainforest

Abstract

The phenology of flowering and fruiting of forest plants was studied at the community level for one year and three months in submontane rainforest on Negros Island in the Central Philippines. Temporal variation in the percentage of plants in flower or fruit was generally less than twofold. Temporal variation in abundance of flowers or fruits was higher, up to fourfold among some months. A new statistical test for comparing proportional phenological data is described. A peak in flowering during the dry season and early part of the wet season was statistically significant; this was due to temporal variation in flowering patterns of trees and palms, but not vines. A minor peak in fruiting during the first half of the wet season was not statistically significant. The long-term relationship of rainfall to phenological patterns remains uncertain. Limited comparisons were possible hetween the two months sampled in both 1983 and 1984. Both flowering and fruiting levels were reduced by one-half in these months in 1984 compared to the same months in 1983, indicating a difference between years in either the timing or magnitude of peaks. Peaks in numbers and an index of volume of flowers or fruit corresponded roughly to peaks in the proportion of plants in flower or fruit, but at some times the proportion of plants flowering or fruiting was a poor predictor of abundance. One of three understorey plots was significantly different from the others, with tenfold differences in proportion of plants flowering or fruiting, and up to 100-fold differences in abundance or volume of flowers or fruits. Lower plant density, markedly different edaphic conditions at one plot and/or differential response to drought may account for these differences. Some large temporal differences in flowering and fruiting were not statistically significant, and larger sample sizes would have been necessary to determine if those differences were real. This highlights the importance of statistical treatment of data and of a concern for sample size in studies such as this.