It is hypothesized that some of the variability in the conclusions of several human cold adaptation studies could be explained if not only were the changes in core and shell temperatures taken into account, before and after cold adaptation, but also the absolute temperatures and metabolic rate in both thermally neutral environments and in the cold. Such an approach was used in a group of volunteers before and after a ski journey (3 weeks at -20 to -30 degrees C) across Greenland. Eight subjects were submitted to cold tests (Tdb = 1 degree C, r.h. = 40%, wind speed = 0.8 m.s-1) for 2 hours. Thermoregulatory changes were also monitored in a neutral environment (Tdb = 30 degrees C). In the neutral environment, the arctic journey increased metabolic rate (11.2%; P less than 0.05) and mean skin temperature [Tsk: 33.5 (SEM 0.2) degrees C vs 32.9 (SEM 0.2) degrees C, P less than 0.05]. During the cold test, the arctic journey was associated with a lower final rectal temperature [36.8 (SEM 0.2) degrees C vs 37.3 (SEM 0.2) degrees C, P less than 0.01], a lower final Tsk [20.7 (SEM 0.4) degrees C vs 21.2 (SEM 0.3) degrees C, P less than 0.01] with no change in metabolic heat production. These observations are indicative of an hypothermic insulative isometabolic general cold adaptation, which was associated with a local cold adaptation of the extremities, as shown by warmer foot temperatures [12.3 (SEM 0.9) degrees C vs 9.8 (SEM 0.9) degrees C, P less than 0.001].