The reported incomplete and dose-dependent absorption of chlorothiazide in humans was demonstrated in six rats after five oral solutions at doses of 0.93, 2.55, 9.23, 25.6, and 70.2 mg/kg. Mean 48-hr urinary recoveries of intact drug were 57.3, 50.4, 36.7, 22.8, and 15.3%, respectively. A similar degree of dose dependency in absorption was found in rat, dog, and human when the doses were related to unit body surface area (BSA) but not on unit body weight, indicating similar interspecies absorptive capacity in terms of unit BSA. This finding may be partly rationalized by marked similarities in the reported solution transit time (2-3 hr) in the small intestine as well as in the calculated gross surface area of the small intestine per unit BSA (0.163 for rat and 0.132 for human). Contrary to the previous postulation of a specific absorption site, the drug was absorbed from different regions of the GI tract with apparent 1-hr absorption rates, studied by the in situ closed-loop method, in the following rank order: jejunum (34.6%) greater than duodenum (32.7%) greater than large intestine (20.1%) greater than ileum (18.0%) greater than stomach (12.4%). Different from the commonly assumed first-order absorption process, the intestinal loop absorption was concentration-dependent, suggesting a saturable mechanism. For example, the absorption rate at 0.008 mg/mL was higher than that at 0.2 mg/mL in ileal loops (61%, p less than 0.01) and jejunal loops (22%, p less than 0.1). In addition, the absorption rates at pH 6 and 7.4 were statistically identical, indicating a lack of ionization effect that is important in the passive absorption process. The solubility-limited absorption could probably be ruled out at doses below 2.55 mg/kg for rat and 125 mg for human in view of higher aqueous solubilities at 37 degrees C (e.g., 1.3 mg/mL at pH 7) found in the present study. Contrary to the previous hypothesis of low membrane permeability as a limiting factor for absorption, the "intrinsic" partition coefficient in 1-octanol/aqueous buffer was moderate, 0.6. Furthermore, the absorption in ileal and jejunal loops was enhanced by an apparent increase in mesenteric blood flow by caffeine. The existence of prolonged oral absorption in rats and humans is discussed.