Seven active male subjects cycled for 60 min at 29.5 (0.8)% peak work rate on three separate occasions in a hot environmental condition [36.0 (0.1) degrees C, 60 (1)% relative humidity] in order to determine the reproducibility of a closed-pouch sweat collection technique for sweat composition at the scapula, forearm and thigh. To confirm that sweat composition was not influenced by between-trial variations in sudomotor drive, local sweat rate, whole-body sweat rate, heart rate (HR), rectal temperature (T(re)) and mean skin temperature (T(sk)) responses were also measured, consequently reproducibility was also established for these variables. Sweat composition did not differ among trials, with the mean coefficients of variation (CVs) for sweat [Na(+)], [K(+)] and pH being 10.4 (7.4)%, 8.1 (6.5)% and 1.3 (1.1)%, respectively. Local sweat rates did not differ among the three trials (P>0.05) although whole-body sweat rate was reduced in the third trial (P<0.05). The mean CVs were 11.0 (7.8)% and 4.7 (1.6)% for local and whole-body sweat rates, respectively. Between-trial differences were not evident for T(re), T(sk) or HR with mean CVs of 0.3 (0.2)%, 0.7 (0.6)% and 3.9 (1.7)%, respectively, although HR tended to be greater in the first trial ( P=0.08). It is proposed that moderate variations in sweat composition were influenced by variations in the local sweat rate, which were induced by application of the pouch.