OBJECTIVE Exposure of restorations in extracted teeth to cyclic thermal fluctuations to simulate one of the many factors in the oral environment has been common in many tracer penetration, marginal gap and bond strength laboratory tests. Temperature changes used have rarely been substantiated with temperature measurements made in vivo and vary considerably between reports. Justification and standardization of regimen are required. METHODS An assessment of reports describing temperature changes of teeth in vivo is followed by an analysis of 130 studies of laboratory thermal cycling of teeth by 99 first authors selected from 25 journals. A clinically relevant thermal cycling regimen was derived from the in vivo information, and is suggested as a benchmark standard. CONCLUSIONS Variation of regimens used was large, making comparison of reports difficult. Reports of testing the effects of thermal cycling were often contradictory, but generally leakage increased with thermal stress, although it has never been demonstrated that cyclic testing is relevant to clinical failures. However, should this be done, the standard cyclic regimen defined is: 35 degrees C (28 s), 15 degrees C (2 s), 35 degrees C (28 s), 45 degrees C (2 s). No evidence of the number of cycles likely to be experienced in vivo was found and this requires investigation, but a provisional estimate of approximately 10,000 cycles per year is suggested. Thermal stressing of restoration interfaces is only of value when the initial bond is already known to be reliable. This is not the case for most current restorative materials.