The ratio of intracellular potassium to sodium in the mammary cells of the lactating mouse was compared with the ratio of potassium to sodium in the mouse milk to determine whether the sodium and potassium concentrations in milk are governed by a Donnan equilibrium as postulated earlier (Peaker 1977a, 1978). An efflux technique was used to determine the average intracellular sodium of 23.0 +/- 1.2 mumol/g tissue (+/- S.E. of mean). The intracellular potassium, determined by calculating the amount of potassium contained in both the interstitial and milk spaces and subtracting these values from the total tissue potassium, was 62 +/- 1 mumol/g tissue. The mean intracellular potassium to sodium ratio, calculated from individual efflux experiments, was 2.7 +/- 0.2. The total, interstitial, and milk water spaces were measured by tissue drying, sodium efflux, and lactose efflux, respectively. The average values (+/- S.E. of mean) obtained were 0.700 +/- 0.004 ml/g tissue, 0.150 +/- 0.016 ml/g tissue and 0.064 +/- 0.004 ml/g tissue. Based on these values the intracellular water space was 0.49 +/- 0.02 ml/g tissue. Intracellular concentrations of sodium and potassium calculated from the intracellular amounts of sodium and potassium and intracellular water space were 47 +/- 3 mM and 129 +/- 5 mM, respectively. The concentrations of potassium and sodium in mouse milk were 47 +/- 1 mM and 26 +/- 1 mM. The mean potassium to sodium ratio (+/- S.E. of mean) calculated from individual milk samples was 1.8 +/- 0.1. The milk ratio of potassium to sodium is significantly different (P less than 0.001) from the intracellular water ratio of potassium to sodium. This finding, in a tight epithelium such as the lactating mouse mammary gland, suggests that both sodium and potassium cannot be distributed passively across the apical membrane and an active transport process must exist for one or both of these ions in this membrane.