Plasma cortisol levels of fingerling rainbow trout were measured as an index of the stress resulting from various procedures used for transport of the fish for stocking. When transported under ' normal ' conditions, which included water at the hatchery acclimation temperature (10-1 1" C), 0, saturation or supersaturation, and neutral pH, there was a marked increase in plasma cortisol levels within 0.5 h, which was maintained over the next 4 h of transport; there was a significant decrease in plasma cortisol by 8 h of transport. It was found that the plasma cortisol levels at 4 and 8 h were not appreciably altered by transport under partial 0, desaturation, 0, saturation, 0, supersaturation, or 0.5% NaCI, or by anaesthesia with tricaine methanesulfonate (MS 222) prior to capture and transport in MS 222-free water or 0.5% NaCI. A 15 min exposure to an immobilizing dose ofbuffered or unbuffered MS 222, or 2-phenoxyethanol, caused an increase in plasma cortisol ofabout 2 h duration, indicating that anaesthetics are themselves stressful. Exposure to chilled water (1" C) caused a large increase in plasma cortisol levels by 4 h after initiation of exposure; plasma cortiso! had decreased at 1 day, and by 2 days a constant level was reached which was above the level in fingerling trout under ' normal ' hatchery conditions. Trout acclimated to chilled water for 24 h and transported in chilled water had an increase in plasma cortisol during transport. Anaesthesia prior to transport or addition of salt did not reduce the stress oftransport as judged by plasma cortisol levels. The results indicate that stress from capture and transport during stocking cannot be avoided using present methods.