Most tropical plants flower synchronously at species-specific times. Low and high temperatures, drought and rain, day length, and variation in solar insolation have all been hypothesized to act as environmental cues for tropical flowering. This abundance of hypotheses has been confronted by a paucity of data, precluding rejection of even one hypothesis. I will use new long-term data sets from Barro Colorado Island (BCI), Panama and a model selection framework to begin the winnowing. The data extend from 1987 to the present and include more than 260,000 flower records obtained in 1,565 weekly censuses of 200 passive traps and meteorological variables obtained just above the forest canopy. Every proximate cue hypothesized to control tropical flowering times was evaluated for 55 well represented species. Hypotheses concerning seasonal variation in temperature, day length, rainfall and photosynthetically active radiation (PAR) best matched the data for zero, five, seven and 32 species, respectively. Many species previously believed to respond to seasonal changes in moisture availability actually respond to seasonal variation in cloud cover and PAR. None of the hypothesized proximate cues matched the data for the 11 remaining species. Solutions to the environmental control of tropical flowering times remain to be discovered.