Fifteen normal male volunteers received 400, 800, and 1200 mg doses of ibuprofen as 1, 2, or 3 tablets, respectively, in crossover fashion, then 420 mg in solution form during the fourth week. Plasma concentration of ibuprofen was measured by an HPLC method. Individual subject concentration-time (C,t) data following the solution were analyzed by two different methods, and results unequivocally indicated the open two compartment model with first order absorption. However, the computer fitting of both arithmetic and geometric mean concentrations led to a different model. A method was developed to obtain absorption data (fraction of drug absorbed, Fa, versus time) for a multicompartmental system from oral data alone, without intravenous data. The method assumes that Vp is constant intrasubject and that absorption is complete following administration of both the solution and tablets. The method was successfully applied to the ibuprofen tablet data. It was shown also that such a method is necessary to obtain ibuprofen absorption data since intrasubject variation of the microscopic rate constants k12, k21, and kel (as reflected by the intrasubject variation of the hybrid rate parameters lambda 1 and lambda 2 or beta and alpha) is of the same order of magnitude as intersubject variation. Absorption of ibuprofen from tablets was shown not to be simple first order as for the solution. The absorption profiles following one tablet were S-shaped, while those following 2 or 3 tablets had partial linear segments indicating zero order absorption.