Extrapolating results of laboratory bioassays to streams is difficult, because conditions such as temperature and dissolved metal concentrations can change substantially on diel time scales. Field bioassays conducted for 96 h in two mining‐affected streams compared the survival of hatchery‐raised, metal‐näive westslope cutthroat trout (Oncorhynchus clarki lewisi) exposed to dissolved (0.1‐μm filtration) metal concentrations that either exhibited the diel variation observed in streams or were controlled at a constant value. Cadmium and Zn concentrations in these streams increased each night by as much as 61 and 125%, respectively, and decreased a corresponding amount the next day, whereas Cu did not display a diel concentration cycle. In High Ore Creek (40 km south of Helena, MT, USA), survival (33%) after exposure to natural diel‐fluctuating Zn concentrations (range, 214–634 μg/L; mean, 428 μg/L) was significantly (p = 0.008) higher than survival (14%) after exposure to a controlled, constant Zn concentration (422 μg/L). Similarly, in Dry Fork Belt Creek (70 km southeast of Great Falls, MT, USA), survival (75%) after exposure to diel‐fluctuating Zn concentrations (range, 266–522 μg/L; mean, 399 μg/L) was significantly (p = 0.022) higher than survival (50%) in the constant‐concentration treatment (392 μg/L). Survival likely was greater in these diel treatments, both because the periods of lower metal concentrations provided some relief for the fish and because toxicity during periods of higher metal concentrations was lessened by the simultaneous occurrence each night of lower water temperatures, which reduce the rate of metal uptake. Based on the present study, current water‐quality criteria appear to be protective for streams with diel concentration cycles of Zn (and, perhaps, Cd) for the hydrologic conditions tested.