We used a simple empirical model to predict denitrification rates from measurements of bulk soil properties. Boundary analysis was used to define relationships between denitrification rate and each of air-filled porosity, respiration rate and mineralizable-C content. The ratio of measured denitrifying enzyme activity to the maximum measured value was used to account for variation in amounts of enzymes and numbers of denitrifiers in soil. Nitrate content had little effect on denitrification rate and was excluded from the model. Because the model did not account for microscale variability, it did not accurately predict rates in individual soil cores. Nevertheless, population means and distributions of predicted and measured values were similar. The seasonal patterns of mean values of predicted and measured denitrification rates were also similar over the second half of the sampling period, which extended from May to November. The model did not account for appreciable denitrification on three dates in May. This discrepancy indicated that environmental regulation of denitrification may not be uniform over the season. The model was not sufficiently sensitive to factors influencing episodic events. Key words: Denitrification rate, model, boundary line