The superimposed twitch technique was used to study the effect of whole-body hypothermia on maximal voluntary activation of elbow flexors. Seven subjects [26.4 ± 4 years old (mean ± SD)] were exposed to 60 min of either immersion in 8°C water (hypothermia) or sitting in 22°C air (control). Voluntary activation was assessed during brief (3 s) maximal voluntary contractions (MVCs) and then during a 2 min fatiguing sustained MVC. Hypothermia (core temperature 34.8 ± 0.9°C) decreased maximal voluntary torque from 98.2 ± 1.0 to 82.8 ± 5.8% MVC (P < 0.001) and increased central conduction time from 7.9 ± 0.4 to 9.1 ± 0.7 ms (P < 0.05). Hypothermia also decreased maximal resting twitch amplitude from 17.6 ± 4.0 to 10.0 ± 1.7% MVC (P < 0.005) and increased the time-to-peak twitch tension from 55.4 ± 4.0 to 79.0 ± 11.7 ms (P < 0.001). During the 2 min contraction, hypothermia decreased initial torque (P < 0.01) but attenuated the subsequent rate of torque decline (control from 95.5 ± 4 to 29.4 ± 8% MVC; and hypothermia from 85.3 ± 8 to 37.3 ± 5% MVC; P < 0.01). Cortical superimposed twitches increased as fatigue developed but were always lower in the hypothermic conditions. Cortical superimposed twitches increased from a value of 0.4 ± 0.3% MVC prefatigue to 3.9 ± 1.4% MVC postfatigue (P < 0.001) in the hypothermic conditions and from 1.7 ± 0.9 to 5.5 ± 2.3% MVC in control conditions. Our results suggest that hypothermia decreases MVCs primarily via peripheral mechanisms and attenuates the rate of fatigue development by reducing central fatigue.