Abstract

The effect of visual isolation on the density of territorial fish such as Atlantic salmon ( Salmo salar ) was examined. Kalleberg's (1958) hypothesis that the density of salmon can be increased by adding boulders was tested for the first time in a natural stream. Eight study reaches were established in Catamaran Brook and the Little Southwest Miramichi River (New Brunswick). Each reach was divided into three sites of 3 x 2 m, and received one of three treatments: boulder-added, 36 boulders (median diameter = 0.21 m) were added to increase visual isolation; boulder-removed, where all boulders were removed; and a control, where the reach was left unchanged. A three-fold increase in the density of salmon was observed in the boulder-added treatment compared to the boulder-removed and control sites. The impact of adding boulders was more important for the young-of-the-year than for the 1+ and 2+ salmon. However, for non-salmonid fishes, no effect was found. A Geographic Information System (GIS) viewshed analysis showed that the visible area and the distance to nearest neighbour were significantly smaller in the boulder-added treatment than in the other two treatments. Although mean water velocity was reduced in the boulder-added treatments, the snout velocity of individual salmon did not differ among treatments, suggesting that the boulders were not used as a velocity refuge. The results support Kalleberg's hypothesis that visual isolation is a key factor controlling salmon density, and that it might be used as a low-cost method for increasing the density of territorial fish.