Lake Superior lake herring (Coregonus artedi) recruitment to 13-14 months of age in the Wisconsin waters of Lake Superior varied by a factor of 5,233 during 1984-1998. Management agencies have sought models that accurately predict recruitment, but no satisfactory model had previously been developed. Lake herring recruitment was modeled to determine which factors most explained recruitment variability. The Ricker stock-recruitment model derived from only the paired stock and recruit data explained 35% of the variability in the recruitment data. The functional relationship that explained the greatest amount of recruitment variation (93%) included lake herring stock size, lake trout (Salvelinus namaycush) population size, slimy sculpin (Cottus cognatus) biomass, the interaction of mean daily wind speed in April (month of hatch) and lake herring stock size, and mean air temperature in April (when lake herring are 12-months old). Model results were interpreted to mean that lake herring recruitment was affected negatively by: slimy sculpin predation on lake herring ova; predation on age-0 lake herring by lake trout; and adult cannibalism on lake herring larvae, which was reduced by increased wind speed. April temperature was the variable that explained the least amount of variability in recruitment, but lake herring recruitment was positively affected by a warm April, which shortened winter and apparently reduced first-winter mortality. Stock size caused compensatory, density-dependent mortality on lake herring recruits. Management efforts appear best targeted at stock size protection, and empirical data implies that stock size in the Wisconsin waters of the lake should be maintained at 2.1-15.0 adults/ha in spring, bottom-trawl surveys.