We hypothesized that plasma arginine vasopressin (AVP) concentrations in children with meningitis are appropriate for the children's degree of hypovolemia, even though the concentrations were higher than expected for the serum osmolality. A randomized study was conducted to compare the effect on plasma AVP concentrations of giving maintenance fluid requirements plus replacement of any deficit versus restricting fluids to two thirds of maintenance requirements for 24 hours. Plasma AVP concentrations and serum osmolality were measured before fluid therapy was begun and again after 24 hours. Nineteen children, 2 months to 17 years of age, were studied; 13 had bacterial meningitis (12 with Haemophilus influenzae type b). Ten children (seven with bacterial meningitis) received a mean of 1.42 times the calculated maintenance fluid requirements, and nine (six with bacterial meningitis) were restricted to a mean of 0.65 times maintenance. Children in the maintenance group also received significantly more sodium (mean = 6.3 mEq/kg/24 hr) than children in the fluid-restricted group (mean = 2.0 mEq/kg/24 hr). The two groups were comparable for plasma AVP concentration and serum osmolality before fluid therapy was begun. The plasma AVP concentration was significantly lower after 24 hours of maintenance plus replacement fluids than after fluid restriction (p = 0.005), and the change in AVP concentration correlated with the amount of sodium given (p less than 0.02). This study supports the hypothesis that serum AVP concentrations are elevated in patients with meningitis because of hypovolemia and become normal when sufficient sodium is given to facilitate reabsorption of water by the proximal tubule of the kidney. Patients with meningitis can be given maintenance plus replacement fluids but should be monitored for the development of the syndrome of inappropriate secretion of antidiuretic hormone.