To improve the management of the European eel (Anguilla anguilla) in freshwater, it is essential to define important lotic habitats. Electrofishing data from 289 wadeable, hard-bottom sites in 69 Swedish coastal rivers and streams, originally surveyed for salmonid monitoring, were used to evaluate the effects of sampling-and habitat-related factors on eel occurrence. Probability of eel occurrence, as influenced by sampling procedure (sampled area, number of consecutive runs and ambient water temperature) and habitat characteristics (size of catchment, dominating bottom substrate, shade, water velocity, mean depth), was evaluated for small (total length B 150 mm) and large ([ 150 mm) yellow eels. Data were analysed in a mixed presence/absence generalized linear model with dispersal (distance to mouth from sampled site), habitat and samplingrelated variables as covariates. The two models explained variation in occurrence to 81.5% for small eel and 76.2% for large eel. Probability of eel occurrence decreased with distance from the river mouth, and increased with sampled area, number of runs, water temperature, coarser substrate and size of river. We suggest that future eel habitat restoration should focus on lower reaches of larger rivers with suitable coarse bottom habitats. Stocking of young eel should be carried out in comparable accessible habitats in the upper reaches where eel densities are low. The results also strongly indicate that eel may be sampled together with young salmonids with DC electrofishing in wadeable habitats.