Conditions for discontinuous polyacrylamide gel electrophoresis have been defined in which progesterone receptors of chick oviduct cytosol and a variety of steroid-binding proteins from other sources are stable and amenable to quantitative analysis. The essential modifications from standard procedures include the use of (1) separation gels in which the cross-linking agent/acrylamide monomer = 15:85, (2) glycerol (10% v/v) in all phases of the Trisglycine-HCl buffer system (pH 10.2 in the separation phase during electrophoresis at 0 degrees), and (3) a layer of a charged reducing agent, thioglycolate, beneath the sample layer. Electrophoresis of untreated oviduct cytosol labeled with [3H]progesterone +/- competing steroids revealed a heterodisperse slow peak and a sharp fast peak. Both peaks displayed the steroid-binding specificity and saturability that are characteristic of intracellular receptors. Recovery of steroid from both the slow and fast components increased linearly with sample load up to 60 mul of cytosol (1.2 mg of protein)/gel (6 mm diameter). The specific progesterone binding detected by this technique was comparable to that detected by charcoal-dextran treatment or ion exchange filtration. Relative electrophoretic mobilities (Rf) of globular protein standards and steroid-protein complexes in cytosol and chick serum were measured in separation gels with total gel concentrations (T) systematically varied from 5 to 15% (w/v). Data were processed by computer programs to obtain weighted linear regressions of log Rf on T (Ferguson plots) and the joint 95% confidence limits of the slopes (-KR) and intercepts of these plots. Molecular radii (R) of the binding components and apparent molecular weights (M) were calculated from the linear correlation of R with KR 1/2 for the standards. The value of M is approximately 158,000 obtained for the cytosol fast component was independent of the length of the separation gel, the presence of a stacking gel or prior exposure of the cytosol to KCl. It was higher than expected from the sedimentation coefficient of 4.2 S in the same pH 10.2 buffer. Electrophoresis in 170-mm separation gels without stacking gels revealed that KCl extracts of protamine-precipitated cytosol contain a different receptor form, of lower net negative charge than the cytosol fast form. The results demonstrate the utility of electrophoresis in highly cross-linked gels of several concentrations to discriminate between various receptor forms and steroid-binding components of serum. This method may lead to overestimates of M for highly asymmetric receptor forms.