Two hypotheses have been proposed to explain the dynamics of sympatric populations of deepwater sculpin (Myoxocephalus thompsonii) and slimy sculpin (Cottus cognatus). The first hypothesis is that slimy sculpins negatively affect survival of deepwater sculpins, and therefore deepwater sculpins coexist with slimy sculpins only when a keystone predator, lake trout (Salvelinus namaycush), is abundant. According to the second hypothesis, changes in the abundances of the sculpins are driven by interactions with fishes other than sculpins. To evaluate both hypotheses, we applied regression analyses to long-term observations on abundances of both sculpin populations in Lake Michigan during 1973-2002. For slimy sculpin abundance, we considered the predation effect by lake trout and the effect of deepwater sculpins on slimy sculpins. For deepwater sculpin abundance, we considered the effect of alewife (Alosa pseudoharengus) on deepwater sculpins, the predation effect by burbot (Lota lota), and the effect of slimy sculpins on deepwater sculpins. An information criterion was used to select the best regression model explaining the temporal trends. The best model to explain trends in slimy sculpin abundance was the model that included the lake trout predation term only. The best model to explain trends in deepwater sculpin abundance was a model including the alewife and burbot predation terms. Thus, a negative effect of slimy sculpins on deepwater sculpins was not essential in capturing the sculpin community dynamics. Therefore, our results supported the second hypothesis. Further, our results supported the contention that control of the alewife population was a prerequisite for restoration of deepwater sculpin populations.