Manual performance while wearing a whole body covering NBC garment was studied at -10 degrees C. Hands were protected by thin cotton gloves, which were covered with rubber gloves. The test subjects were exposed for 40 min in one of the four conditions: standing at -10 degrees C, standing for 10 min followed by walking (5 km/h) for 30 min on a treadmill, standing while holding a solid steel bar (see section 2.2), or standing at 20 degrees C. Six different manual tasks were performed after each 40-min exposure. All tests were also performed with bare hands at 20 degrees C. Moreover, the effect of contact cooling on skin temperatures and rewarming thereafter was examined by means of gripping a steel bar (-10 degrees C) during cold exposure. During exposure to -10 degrees C conditions finger skin temperature rapidly decreased to 10.7 +/- 2.2 degrees C (mean +/- SD). Improvement of body heat balance by exercise increased finger temperatures to 19.6 +/- 9.0 degrees C. Hand temperature remained at a higher level both during rest and exercise at -10 degrees C (20.1 +/- 1.7 and 20.6 +/- 6.1 degrees C, respectively). Cold exposure deteriorated manual performance and especially those functions that are related to finger dexterity. Finger skin temperature had high correlation with screwing, peg-board and magazine loading tests (r = -0.90, r = -0.77 and r = -0.72, respectively, p < 0.01) but no relation was found with hand grip strength (r = -0.03). Manual performance was impaired in every test both by gloves and cooling. Contact cooling decreased skin temperatures on the palm side of the hand and fingers around twice as effectively in normothermic subjects and 3.9-6.5 times more effectively in cooled subjects in comparison to cooling by cold air alone. Contact cooling had no significant effect on skin temperatures on the dorsal side of the fingers. The rewarming rate after the release of the steel bar was clearly higher in the dominant hand in comparison to the non-dominant hand. In conclusion, the present results show that the thermal insulation of rubber gloves was clearly insufficient, allowing unacceptably low finger temperatures during work in the cold. However, only those tasks requiring finger dexterity were clearly adversely affected. Heat production by physical exercise was able to increase finger skin temperature and to partly restore manual performance. Handling of cold tools is especially harmful for the palm side temperature of the non-dominant hand.