We quantified relationships between both fyke-net catch rates (catch/net-night) and electrofishing catch rates (catch/mi) and population densities (number/acre) of walleye Sander vitreus (formerly Stizostedion vitreum) for adult population estimates and total population estimates to determine whether catchability was density dependent. Fyke-net catch rates were modeled as a nonlinear function of adult walleye density and of four size-classes of the adult population, and electrofishing catch rates were modeled as a nonlinear function of adult and total walleye density and four size-classes of the adult and total populations. The results showed nonlinear relationships between catch rate and density for the adult and total populations. We accounted for measurement errors in catch rates and densities by estimating bias-corrected slopes by means of Monte Carlo simulations and estimated measurement-error ratios by means of an errors-in-variables model. We found that the bias-corrected slopes were higher than ordinary-least-squares regression estimates and that measurement errors were greater in catch rates than in density estimates. Lastly, we sought to explain the residual variability in the relationships between (1) fyke-net catch rates and adult walleye densities and (2) electrofishing catch rates and adult and total walleye densities. We found that the fyke-net catch rate was positively related to adult walleye density and percent littoral zone (percentage of lake surface area ???20 ft deep) and negatively related to conductivity. We found that the electrofishing catch rate of adult walleyes was positively related to adult walleye density and conductivity and that the electrofishing catch rate of the total walleye population was positively related to total walleye density. We concluded that the nonlinear relationship between catch rates and walleye abundance limits the use of catch rates to index walleye abundance in northern Wisconsin lakes.