Predicting the response of coral reefs to large-scale mortality induced by climate change will depend greatly on the factors that influence recovery after bleaching events. We experimentally transplanted hard corals from a shallow reef with highly variable seawater temperature (23-36°C) to three unfished marine parks and three fished reefs with variable coral predator abundance and benthic cover. The transplanted corals were fragmented colonies collected from a reef that was relatively undisturbed by the 1997-1998 warm-water temperature anomaly, one of the most extreme thermal events of the past century, and it was assumed that they would represent corals likely to succeed in the future temperature environment. We examined the effects of four taxa, two fragment sizes, an acclimation period, benthic cover components, predators and tourists on the survival of the coral fragments. We found the lowest survival of transplants occurred in the unfished marine parks and this could be attributed to predation and not tourist damage. The density of small coral recruits approximately 6 months after the spawning season was generally moderate (~40-60/m2), and not different on fished and unfished reefs. Coral recovery between 1998 and 2002 was variable (0-25%), low (mean of 6.5%), and not different between fished and unfished reefs. There was high variability in coral mortality among the three unfished areas despite low variation in estimates of predator biomass, with the highest predation occurring in the Malindi MNP, a site with high coralline algal cover. Stepwise multiple regression analysis with 14 variables of coral predators and substratum showed that coralline algae was positively, and turf algae negatively associated with mortality of the transplants, with all other variables being statistically insignificant. This suggests that alternate food resources and predator choices are more important than predator biomass in determining coral survival. Nonetheless, large predatory fish in areas dominated by coralline algae may considerably retard recovery of eurythermal corals. This will not necessarily retard total hard coral recovery, as other more predator-tolerant taxa can recover. Based on the results, global climate change will not necessarily favor eurythermal over stenothermal coral taxa in remote or unfished reefs, where predation is a major cause of coral mortality.