Human milk sodium ([Nal]) and potassium ([K)] concentrations were measured every 4 h for 24 h in 28 subjects 3.5 to 32 wk postpartum. A diurnal variation in milk Na was seen, which was reciprocal to K. Significant negative correlations between Na and K were seen in these periods: 3.5 to 6; 8 to 18, 20 to 32 wk postpartum (p less than 0.01 for each). The mean 24 h milk sodium concentrations (x(Na)) decreased between 3.5 and 18 wk postpartum (p less than 0.005 by paired t test). Changes in mean potassium (-XK) were not statistically significant. Diet apparently does not affect milk Na. Administration of a low Na diet: 10.8 +/- .9 (SD) mEq Na/day and 60 to 100 mEq K/day for 2 days did not change x(Na). But urinary Na decreased 7-fold as aldosterone increased 5-fold. No significant correlation was seen between 24 h Na excretion in urine and x(Na) in milk (n = 51). A significant positive correlation was seen between urinary K and -XK in milk (r = 0.36), p less than 0.001).