Total densities of wild brown troutSalmo truttaages 1 to 7 varied widely (range, 130–1,304/hectare) among 12 sample sections in Spring Creek, a 35-km-long stream in a limestone area of central Pennsylvania. During a 2-year study we determined the relative importance of spawning habitat and other habitat features on spatial variation in density. Densities of age-0 and of all age-1 and older brown trout were positively correlated with redd densities. Embryo survival was lowest in sections with low brown trout densities, but embryo survival was not as important as redd density in determining population size. Redd density appeared to be a function of availability of suitable spawning substrate. Habitat variables such as depth, pool area, cover, and substrate were not correlated with brown trout density. Spawning habitat and other habitat features were quantified with the habitat suitability index model. Scores derived from the model were poorly correlated with densities of age-0 and of age- l and older brown trout. We concluded that juvenile brown trout do not disperse widely from natal areas, and that local population densities are largely a function of the availability of spawning habitat.