1 The ability of phylogenetic, water stress and 'shared-pollinator' hy
potheses to explain flowering times was tested for the flora of Barro
Colorado Island, Panama. Both quantitative (217 species) and qualitati
ve (1173 species) flowering data were used. 2 Descriptive metrics calc
ulated from quantitative flowering data indicated that mean flowering
times fell in the final two months of the dry season and the first two
months of the wet season for 56% of species and that flowering was co
ncentrated in a short, predictable part of the year for the great majo
rity of species. 3 The 'shared-pollinator' hypothesis predicted that c
ongeners should have temporally segregated flowering times. This hypot
hesis was rejected for 57 of 59 genera. 4 The water stress hypothesis
predicted that species from drier habitats and life forms with limited
access to soil water should flower in the wet season. This hypothesis
was also rejected. 5 There were, however, strong phylogenetic pattern
s. Both the mean and the variance of flowering times were similar amon
g congeners. The variance of flowering times was also similar among co
nfamilials, and mean flowering times were concentrated in the wet seas
on for monocotyledons.