Three methods, equilibrium dialysis, ultrafiltration through dialysis bags, and ultrafiltration through ultrafree filters were compared for the determination of propranolol binding in serum. The method of choice was equilibrium dialysis, which gave values for the free fraction which were not significantly different from those obtained by ultrafiltration through the filters. Of the three methods, equilibrium dialysis also had the best precision and required the least sample volume (1 ml). Similar values for the free fraction were also obtained using either 14C- or 3H-radiolabeled propranolol by this method. Investigation of the effect of temperature and pH of serum indicated that small changes in either parameter could result in up to twofold alterations in the free fraction, but changes in propranolol concentrations over the range 10-500 ng/ml had little effect on the free fraction. Using a standardized equilibrium dialysis method (i.e., 4 hr dialysis against isotonic phosphate buffer, at 37 degrees C, pH 7.4), a mean free fraction of 14.7% (range 8.4-23.3) was observed in 34 samples from 21 healthy subjects. This was significantly greater (p less than 0.001) than that (6.8%, range 3.1-12.3) found in nine elderly hospitalized patients with acute disease. The patients also had significantly raised levels of alpha 1-acid glycoprotein (AAG), which is a major binding protein for propranolol in serum. A significant linear correlation (r = 0.884, p less than 0.02) was found between the binding ratio for propranolol and serum AAG concentrations in these subjects.