Surface and subsoil samples were collected from agriculturally important soils of Coastal Plain, Piedmont and Ridge and Valley regions of Virginia for the purpose of determining the contribution of different soil N fractions to plant available N. Soil samples were analyzed for exchangeable and non-exchangeable NH₄⁺-N, NO₃⁻-N, total N, and organic matter contents. The samples were also subjected to the anaerobic incubation procedure as an index of organic N availability. Plant available N was measured by N uptake of successive wheat crops grown in the greenhouse. Multiple linear regression models for different groups of samples were used to determine the contribution of the different soil N fractions to the plant available N supply, and to predict N uptake by wheat. Best models were selected considering fit, significance of the regression coefficients, and predictive ability. Due to the high correlation among the different soil N fractions, important collinearity was present and affected the linear models. These effects were reduced by utilizing biased techniques. All the soils provided significant amounts of N to the wheat in both the first and second crops. Exchangeable NH₄⁺-N and NO₃⁻-N were the major initial sources of plant available N. Non-exchangeable NH₄⁺-N was also a significant contributor to the plant available N supply for most soils. The results of the biological and chemical indices of organic N availability were highly correlated among each other, and with plant N uptake by the first and second wheat crops. Plant N uptake was associated with the variation observed in the different soil N fractions, indicating that wheat is a good indicator crop for plant N availability experiments. The procedures used to detect and combat collinearity were effective in producing more stable models with better predictive ability. Further research should be conducted under field conditions to study the contribution of non-exchangeable NH₄⁺-N to plant N availability. / M.S.