The distribution of soil nitrogen (N) in profiles of plots receiving annual plant applications of 0, 45 or 68 kg anhydrous ammonia N ha^-1 prior to planting cotton in experiments established in 1955 was studied by determining the amounts of N in various organic and inorganic fractions of soil cores collected May 1988 in 15 cm increments to a depth of 150 cm. The data obtained indicated that N cycling in Norwood soil cropped to cotton is highly efficient, and that fertilization led to a greater portion of total soil N in the profile residing as organic N near the soil surface. The amounts of total N in the surface 45 cm were 34-68% greater in fertilized plots than in unfertilized plots, whereas the amounts of organic and clay-fixed N at depths between 60 and 105 cm were greater in unfertilized plots. These findings suggest that mineralization of organic N was the principal source of plant-available N in unfertilized plots, and that clay-fixed N in surface and subsoils is involved in N nutrition of cotton. Regression analyses relating seedcotton yield to the distribution of soil N showed that a model including the rate of N applied and the amounts of organic N and mineral N from the surface 15 cm only provided the greatest predictive value (r²=0.97). Although these experiments indicate that nitrate concentrations and indices of N availability are less effective than measures of organic or total N in predicting the`response of cotton to fertilizer N, additional experiments on a Commerce silt loam showed that substantial amounts of residual nitrate are available when excessive amounts of fertilizer N are applied the preceding season.