The fluorescence retention and intensity of juvenile brown trout Salmo trutta marked during their first summer were monitored in a hatchery and in four natural streams. A handheld detector was used for direct examination. In the hatchery, three marking treatments (T) were compared: 3·5 min in a 0·5% calcein solution (T0·5-3·5), 7 min in a 0·5% calcein solution (T0·5-7) and 3·5 min in a 1% calcein solution (T1-3·5). The fish were raised indoors for 11 months and then outdoors until 18 months. The fluorescence retention rate was 100% in all treatments at 11 months, although T1-3·5 showed the highest mean fluorescence intensity, followed by T0·5-7 and T0·5-3·5. The fluorescence intensity was not correlated with the final total length (LT) of the fish in two treatments, although it significantly decreased with increasing LT in T1-3·5. At 18 months, <30% of the fish were still slightly fluorescent, suggesting a negative effect of sunlight exposure. In stream studies, the fluorescence intensity did not significantly differ according to final LT; an overall mean ± s.d. retention rate of 70·7 ± 26·6% was measured at 12 months with a decrease to 48·6 ± 24·6% at 24 months. Significant differences amongst streams and within reaches of the same stream were observed. Because of a significant positive effect of the shading index on the fluorescence intensity, the use of calcein should be restricted to streams unexposed to direct sunlight. Consequently, the marking method would be reliable for 1 year monitoring studies in shaded streams.