Abstract Four physically fit but not heat-acclimatized male subjects were exposed to sessions of 90min, comprising three successive 30min units of (i) cyclic work-rest periods in a heated room, (ii) sitting under neutral ambient conditions and (iii) repetition of unit 1 of cyclic work in the heated room. Each cycle included consecutive periods of a few minutes of: (i) heavy work at half aerobic capacity (05 [Vdot]O2 max); (ii) light work at 025 [Vdot]O2 max and (iii) sitting. The heavy work was performed under a radiant heat source where tg was 20°C above tdb. The heated room was under one of the following tdb°C/twb°C combinations: 38/27 (warm/humid), 48/25 (hot/dry) and 48/30 (hot/wet). The unit of rest was under tdb/twb 22/16°C. The 30min unit in the heated room included three cycles of work and rest. The design of the cyclic work in the heat and the 30min recovery outside the heat was based on the inverse relationship between endurance and work as a fraction of the maximal aerobic capacity (f- [Vdot]O2max). The f [Vdot]O2 max was determined from the predicted cumulative increments in heart rate (HR) due to work, the tdb, tg and the air vapour pressure. The adequacy of the work schedule was judged by the levels of the peak and average HRs, and by the rise in rectal temperature (Trc). The stress of the ambient conditions was evaluated from the heat balance calculated as the required cooling (Ereq) due to the metabolic, radiative and convective heat loads and the ambient capacity for evaporation (Emax) Although the total exposure inside and outside the hot area involved favourable heat balance (Ereq/ (Emax <l), the cyclic work in the hot area was at the limit of tolerance (Ereq/ (Emax =l). Consequently, the HR responses proved the design of the work-rest schedule adequate, but a creeping up of Trc seemed to indicate that longer recovery under neutral conditions were needed. This was attributed to the heat-acclimatization state of the subjects.