Stress in wild brown trout Salmo trutta was assessed by sampling blood and measuring the concentrations of plasma cortisol and blood glucose in fish collected by electrofishing and immediately anaesthetized with metomidate. In the River Nidelva, Trondheim, Norway, the resting blood plasma cortisol concentration in the juvenile (0 + year) brown trout was 52 ± 44 nM (mean ± s.d.) in December and 2·3 nM (detection limit) in January. The corresponding blood glucose values were 1·8 ± 0·9 and 1·2 ± 0·2 mM, respectively. After electrofishing, handling and transport to the artificial stream, plasma cortisol and blood glucose levels increased significantly in both experiments. A maximum plasma cortisol level of 239 ± 120 and 71 ± 32 nM and a maximum blood glucose level of 3·9 ± 0·9 and 3·0 ± 0·9 mM were measured in the December and January stream channel experiments after transport, respectively. After introduction to the artificial stream, the blood plasma cortisol level returned to resting values within 24 h in the January stream channel experiment. The blood glucose levels remained at a higher level compared to the reference group throughout the December experiment, while it returned to resting values after 24 h in the January stream channel experiment. The major difference between the December and January experiments was the temperature within the artificial stream, 15–17° C in December and 7–9° C in January. This may have influenced the blood glucose levels. After dewatering of the artificial streambed there was a significant increase in plasma cortisol both in the December and January experiments, and after 24 h the plasma cortisol returned to the resting level in the January experiment. The blood glucose also increased during dewatering, although not significantly.