To test the hypothesis that a substance present in the amniotic fluid could serve as a regulator of amniotic fluid volume, we drained and discarded amniotic fluid while replacing it with lactated Ringer's solution that was isotonic to amniotic fluid. Seven ewes with singleton fetuses at 119±1 days of gestation (mean ± sem) were instrumented with multiple indwelling catheters in the pedal artery, pedal vein and amniotic cavity. During the exchange periods, an average of 3,019± 171 ml of lactated Ringer's solution was infused per day into the amniotic cavity while an equal amount of amniotic fluid was pumped out and discarded. During the control period amniotic fluid composition and volume was not altered. Exchange and control periods started with the same amniotic fluid volume, lasted 3 or 4 days duration and were randomized as to order. Amniotic fluid volume measured by vacuum drainage was 556±98 ml at the end of the control period and 986±209 ml (P = 0.03) at the end of the exchange period. Fetal arterial blood gases, hemodynamic parameters and the osmolality gradient between fetal plasma and amniotic fluid were not altered by the exchange process. A linear relationship between the control amniotic fluid volume and volume at the end of the exchange period (P = 0.003) suggests that the animals with larger control volumes responded to isovolemic dilution with a larger increase in volume. We conclude that amniotic fluid may contain a substance that regulates amniotic volume.