Progesterone was administered percutaneously to postmenopausal women in topical applications on the breast and chest areas in a hydrophilic (gel), lipophilic and an emulsion type base. Venous blood samples were taken 2, 4, 6, 24, 48 and 72 h following administration. The plasma levels were evaluated by radioimmunoassay. Time of maximum concentration (tmax) was, in all cases, in the neighborhood of 4 h. Mean peak plasma concentrations were: 1 ng/ml for the lipophilic, 1.24 ng/ml for the hydrophilic and 2.26 ng/ml for the emulsion type base. The areas under the curves (AUCs) were practically equivalent for the first two methods, but higher values were obtained for administration in the emulsion type base. The elimination was slow, with a half-time varying in the range of 3040 h for all three types of base, a value that was much higher than those obtained after administration of progesterone via vaginal suppositories. The AUCs were parallel with the peak plasma concentrations: almost 2-fold higher for emulsion than for the gel and lipophilic base. Fit for plasma levels using mono-, bi- and tricompartmental models furnished acceptable results only in the case of monocompartmental model, which raises a number of physiological and physico-chemical considerations. A 'pseudomonocompartmental' model was constructed to explain this 'anomaly'.