Leaf-fall, leaf-flush (newly expanded leaves), flowering and fruiting were recorded monthly between April 1982 and May 1985 in 703 individuals of 99 species of trees, shrubs and herbs in a logged lowland, complex rainforest. Weekly water availability and demand were modelled using pan evaporation, rainfall, and a range of estimated maximum soil water storage. The water balance model indicated that rainfall distribution was an inadequate measure of the seasonal availability of water for growth and reproduction. Intermittent shortages in available water were predicted between October and February, well into what is the wettest season of the year in terms of long-term rainfall averages. Adequate water was available in soil water stores through most of the August–October ‘dry season’.

Leaf-fall, flowering, and fruiting all followed distinct seasonal patterns although appreciable levels of activity occurred in all months. Community leaf-fall increased rapidly following an annual low in the very wet, February–May period, to a peak in August. The rapid increase in leaf-fall activity was accompanied by pronounced depressions in flowering and leaf-flushing and coincided with the period of minimum temperatures and insolation. There was no evidence to suggest that this leaf-fall peak was accompanied or induced by drought. This ‘winter’ was followed by rising temperatures and insolation at a time when adequate soil water was usually available, and flowering activity (number of species, number of individuals, and total abundance) increased rapidly in August and peaked from September to October. The flowering peak was accompanied by a major leaf-flush. Leaf-flushing continued intermittently from October to February, a period characterized by maximum temperature, maximum insolation, and periods of predicted water stress when soil water stores were depleted during short, rainless periods. A second peak in leaf-flushing coincided with the wet in March-April. This was accompanied by a relatively small increase in leaf-fall activity. Fruiting showed a bimodal peak from October to April each year. Prediction of functional relationships between climate and species and individual behaviour from community phenological patterns was considered to be inappropriate.