We tested the effectiveness of stable isotopes as recorders of physiological changes that occur during coral bleaching and recovery. Montipora capitata and Porites compressa fragments were bleached in outdoor tanks with seawater temperature raised to 30°C (treatment corals) for one month. Additional fragments were maintained at 27°C in separate tanks (control corals). After one month, (0 months recovery), buoyant weight was measured and a subset of fragments was frozen. Remaining fragments were returned to the reef for recovery. After 1.5, 4, and 8 months, fragments were collected, measured for buoyant weight, and frozen. Fragments were analyzed for stable carbon and oxygen isotopic compositions of the skeleton (δ¹³C_s; δ¹⁸O_s) and nitrogen and carbon isotopic compositions of the host tissue (δ¹⁵N_h; δ¹³C_h) and zooxanthellae (δ¹⁵N_z; δ¹³C_z). δ¹³C_s decreased immediately after bleaching in M. capitata, but not in P. compressa. δ¹⁸O_s of both species failed to record the warming event. During the remaining months of recovery, δ¹³C_s and δ¹⁸O_s were more enriched in treatment than control corals due to decreases in calcification and metabolic fractionation during that time. Increased δ¹⁵N_h of treatment P. compressa may be due to expelled zooxanthellae during bleaching and recovery. Increased δ¹⁵N_z at 1.5 months in treatment fragments of both species reflects the increased incorporation of dissolved inorganic nitrogen to facilitate mitotic cell division and/or chl a/cell recovery. Changes in δ¹³C_h and δ¹³C_z at 1.5 months in treatment M. capitata indicated a large increase in heterotrophically acquired carbon relative to photosynthetically fixed carbon. We experimentally show that isotopes in coral skeleton, host tissue and zooxanthellae can be used to verify physiological changes during bleaching and recovery, but their use as a proxy for past bleaching events in the skeletal record is limited.