Relative abundance of lake trout and burbot in the main basin of Lake Michigan in the early 1930s

We used general linear models to compare and to assess the sources of variation in CPUEs (catches per unit effort) of lake trout Salvelinus namaycush and burbot Lota lota netted in small-mesh gill nets from the main basin of Lake Michigan in the early 1930s, before populations of both species collapsed. Spring, summer, and fall CPUEs were modeled separately. In all seasons combined, CPUEs were over 2.5 times greater for lake trout than for burbot, and lake trout CPUEs were greater than burbot CPUEs in five of six regions of the lake. The hypothesized sources of variability in log(e)(CPUE) included lake region, depth (either as a linear [depth only] or a quadratic [depth and depth(2) included] relationship), number of nights out that the nets fished (also as either a linear or quadratic relationship), and the possibility that the effects of depth could vary among regions. The models predicted 57-68% of the variation in lake trout CPUE and 29-37% of the variation in burbot CPUE (log scales). Lake region was highly significant for both species in all seasons. In all seasons, the relationship between lake trout log(e)(CPUE) and depth was quadratic and varied among regions. For lake trout we did not find any effects of number of nights out. Burbot log(e)(CPUE) varied linearly with depth and with nights out in the spring. In the summer, log(e)(CPUE) for burbot varied linearly with nights out, and its relationship with depth was quadratic and varied among regions. In the fall, quadratic relationships between burbot log(e)(CPUE) and depth again differed among regions, but there was no relationship with nights out.