The dissolution rates of theophylline from six commercially available products (three uncoated and three sustained-release formulations) were determined in distilled water using the USP and rotating-filter dissolution apparatus. The effect of pH on the dissolution of these products was also examined by both methods. In addition, the effect of stirring rate on the dissolution of theophylline from these products was studied using the rotating-filter apparatus. The data obtained under all conditions were reproducible and well-described by a first-order equation. There was no significant difference between the percent of labeled content dissolved in 30 min (D30) and in 60 min (D60) obtained by the USP method and those obtained by the rotating-filter apparatus. The product-to-product variation in D30 and D60 was significant (p less than 0.001) for both the sustained-release and uncoated dosage forms. The pH of the dissolution fluid had a significant effect on the dissolution of theophylline from the products. The data obtained from the dissolution and absolute bioavailability studies in the rabbit were subjected to linear least-squares regression analysis, and good correlations were obtained between the dose-normalized peak serum level, time-to-peak, percent of the dose absorbed at 1 h and at 6 h, or the dose-normalized area under the curve from t = 0 to t = 00 and from t = 0 to t = 6 h and D30, D60, or the rate constant for dissolution. The linear relationship assumed for two of the products was used to predict bioavailability parameters from dissolution variables. The values predicted by this method were not statistically different from the actual values of these parameters.