The trade-off between noise and temperature and their combined effects on discomfort were studied on 108 lightly clothed subjects (0.6 clo), individually exposed for 2 h in a climatic chamber. Every 10 min of the first hour, subjects could modify the experimental conditions by deciding a change in temperature or noise. However, any change in one parameter was experimentally associated with a fixed change in the other parameter according to eight predetermined designs and all trials for thermal improvement were detrimental to acoustic comfort and conversely. Four initial exposures started at thermoneutrality (24 degrees C) in a noisy environment (85 dBA, recorded fan noise), the reduction of noise being linked to a temperature change towards cool or warm climates. The other four conditions started at a low noise level (35 dBA) but in a cool (14 or 19 degrees C) or warm (29 or 34 degrees C) environment, the reduction of thermal discomfort towards 24 degrees C leading to a louder noise. After six possible voluntary changes, the environment was kept constant for 1 h. Ambient parameters, skin temperatures, and subjective estimates were recorded. Results showed that females accepted noisier environments than males, suggesting that thermal comfort is dominant for women. Noise was rated as the most unpleasant factor when initial conditions were noisy whereas temperature was the most disturbing factor when subjects began the experiment with thermal conditions far from thermoneutrality. Finally, although the combined effects of noise and temperature did not influence the physiological data, our results suggest that noise may alter thermal pleasantness in warm conditions.