The onset of large-scale coral bleaching events is routinely estimated on the basis of the duration and intensity of thermal anomalies determined as degree heating weeks. Degree heating weeks, however, do not account for differential rates of heating. This study aimed to explore the relationship between different rates of heating above the documented regional winter threshold, and resultant bleaching of the reef-building coral Acropora formosa. Under a relatively low light field, rapid heating of 1 degrees C day(-1) from 29 degrees C to 32 degrees C lead to a 17.6% decline in F(v)/F(m), concurrent with a rapid increase in xanthophyll de-epoxidation sustained into the dark, whereas slower heating rates of 0.5 degrees C day(-1) lead to no decline in F(v)/F(m) and no change in dark-adapted xanthophyll cycling. At the winter bleaching threshold of 30 degrees C, areal net O(2) evolution exceeded the control values for rapidly heated corals, but was lower than the controls for slowly heated corals. At the maximum temperature of 33 degrees C, however, both treatments had net O(2) fluxes that were 50% of control values. At 30 degrees C, only symbiont densities in the slowly heated controls were reduced relative to controls values. By 33 degrees C, however, symbiont densities were 55% less than the controls in both treatments. The rate of heat accumulation was found to be an important variable, with rapidly heated corals attaining the same bleaching status and loss of areal O(2) production for half the degree heating week exposure as slowly heated corals. The study revealed that it is incorrect to assume that significant dark acclimation disables non-photochemical quenching, because 75% of an increased xanthophyll pool was found to be in the de-epoxidated state following rapid heat accumulation. This has important ramifications for the interpretation of chlorophyll fluorescence data such as dark adapted F(v)/F(m).