ABSTRACT: Gyrodactylus salaris is a major threat to wild populations of Atlantic salmon Salmo salar in Norway. Inter-river dispersal of this parasite is considered to be mainly linked to infested fish. Dispersal through movement of contaminated water is considered possible but has not been quantitatively evaluated. To approach this problem, we built a simple model to estimate (1) the water concentration of free-living G.salaris and (2) the probability that given water volumes would contain free-living parasites. The concentration was back-estimated from reported experimental observations on the number of salmon juveniles found to be infected after being caged in a heavily infested river in Norway, protected from direct contact with the substrate or fish. Monte Carlo simulation was used to account for uncertainty and variability. The average concentration of G. salaris in the water column during the experiment was estimated to be 0.12 m^–3 (95% CI = 0.05 to 0.24). The probability that a given volume would contain at least 1 parasite was 1.2 × 10^–4 (95% CI = 5 × 10^–5 to 2.4 × 10^–4) for 1 l, increasing to 0.67 (95% CI = 0.39 to 0.91) for 10000 l. Thus, only high volumes were likely to represent a significant risk. Since we used data from a heavily infested spot, and despite the fact that time and space variations in water concentration of G. salaris may occur, we conclude that the risk of inter-river dispersal due to transfer of infested water through common river activities by humans is probably low.