According to the most customary exercise protocols, core temperature (Tc) rises in parallel with workload (WL) and experimental time. Physiological variables, however, may be related to each of these factors. To investigate effects of WL independent of experimental time and body temperature, we employed four moderate WLs in 4-min steps between 35 and 65% peak O2 uptake (VO2 peak) in randomized order. To investigate independent effects of body temperature, the same work protocol was performed both after resting in comfortable ambient temperature [control test (Cont)] and after a double cold exposure [precooling test (Pret)], where Tc and the temperature set point are decreased by approximately 0.6 and 0.3 degrees C, respectively. Eight male subjects (24 +/- 1.9 yr, VO2 peak 4.9 +/- 0.5 l/min) worked on a cycle ergometer in a climatic chamber. Heart rate (HR) and breathing frequency (BF), but not preferred pedal rate (PR), were positively correlated to Tc, the slopes amounting to 17 and 3.75 min-1/degree C for HR and BF, respectively. The regression appeared linear over the whole temperature range, and the regression lines were not shifted by precooling. PR was increased by time, but Pret-Cont differences of PR and Tc were inversely correlated (r = -0.50, P < 0.01). The effects of WL were highly significant on HR, O2 uptake, and rate of perceived exertion but not on BF, PR, and sweat rate. The relation of rate of perceived exertion to HR was shifted by precooling.