When separating free and antibody bound digoxin, dextrancoated charcoal suspensions in protein-free urine dilutions develop an improve sieve effect which improves with aging of the suspensions. On addition of albumin to the dilution medium (buffer), the adsorptive capacities become stabilized. In urine dilutions less than of 1:50, false positive digoxin concentrations were measured. Determinations of 0-5 ng digoxin/ml in the diluted urine (1:100) and in the buffer solution show identical results. The recovery of added digoxin (1 ng and 2 ng/ml) is 100%; the lower detection limit is 3.6 mug/l urine. Between-assay variation for means of duplicates shows in the main measuring range (1.5-2.8 ng digoxin/ml) a variation coefficient less than of 4.5%; for a commercial test this value is distinctly higher. Dihydrodigoxin is measured at 3% only. Application of this method for measuring cumulative glycoside elimination of 5 healthy persons (0.5 mg digoxin daily intravenously during 2 weeks) results in a minimal interindividual variation of 5.1%. In a test preparation with a bioavailability of 94.6% the variation is 5.2%.