An investigation of the site and mechanism responsible for the maternal-fetal electrical potential difference (PD) was done in 11 anesthetized guinea pigs at 54-56 days gestation. We removed the most distal fetus and placenta from one uterine horn and secured a catheter, thermistor, and Ag-AgCl electrode in the resulting pouch. The pouch was filled with Earle's solution. We placed another thermistor and electrode in the maternal abdomen. The PD between electrodes was monitored continuously; periodic samples of maternal blood and intrauterine fluid were taken. Thirty minutes after the uterus was filled, the PD (uterine cavity negative) averaged 29.6 +/- 4.5 (95% confidence interval of the mean) mV. Over 4 h, intrauterine K+ concentration [( K+]) decreased from 4.9 to 2.6 +/- 0.5 meq/l, against a chemical and electrical gradient. In eight animals, we measured bidirectional Na+ flux using 22Na and 24Na. The flux ratio was not distinguishable from unity despite a significant PD. Our data indicate that the maternal-fetal PD is probably generated by the endometrial epithelium and that Na+ and K+ both move across the epithelium by active transport or cotransport rather than simple diffusion.