Locomotor behaviour of particular person rainbow trout (Salmo gairdneri) was noticed in a chamber with facet-by-aspect flows of fresh water and 0, 1, 10, or 30% oil refinery effluent, with a steep gradient between the 2 flows. None of the concentrations caused statistically important avoidance or preference as measured by changes in time spent by fish in the four areas of the chamber (upstream or downstream and north or south side). Equally, the aspect of entry of the effluent was not associated to any differences in relative activity in the 4 areas, as measured by the numbers of movements fish made throughout the areas. When 30% effluent was current, the overall level of activity in all elements of the chamber appeared to be lower than activities for the opposite concentrations, however the distinction was at the borderline of statistical significance (P=0.06). Behavioural response to the effluent was much less vital than direct physiological results on trout, beforehand documented with the identical effluent. The tested batches of effluent had an total three-day LC50 of about one hundred% effluent. The effluent exceeded Canadian regulations in average stage of ammonia, was somewhat elevated in oil-plus-grease and residue, but was well beneath regulatory limits for phenols and sulphides.