1. Oxytocin was pressure injected through a glass micropipette into a supraoptic (SON) or paraventricular nucleus (PVN) while recording the electrical activities of oxytocin cells in a contralateral nucleus, to see whether oxytocin acts locally in the magnocellular nuclei to control their bursting activity and whether the oxytocin cells of the four magnocellular nuclei were functionally interconnected during suckling. To test the rapidity of these relations, similar intranuclear injections were realized with acetylcholine, known to rapidly increase the background activity of oxytocin cells. The effects of intranuclear injections of oxytocin and acetylcholine were tested before and after interhemisphere sections of various dimensions. 2. Injecting oxytocin (1 ng in 100 nl) into a magnocellular nucleus (5 times into the PVN and 15 times into the SON) facilitated the occurrence and increased the amplitude of bursts of the oxytocin cells in both the contralateral PVN and SON. This facilitatory effect was similar to that induced by intraventricular injection of the same dose of oxytocin, though slightly delayed and lower. 3. Injecting acetylcholine (0.6 microgram in 100 nl) into the SON (7 times) induced a rapid and sustained increase in the background activity of oxytocin cells in both the contralateral PVN (2 times) and SON (5 times) within the same delay (less than 15 s). This excitatory effect was similar to that induced by an intraventricular injection of 5 micrograms acetylcholine. The effects on bursting activity were not considered in this study. 4. Neither the injections of oxytocin or acetylcholine outside but near the magnocellular nuclei (200-500 microns), nor the intranuclear injection of 100-200 nl of cerebrospinal fluid-like medium, modified the background activity, the frequency and amplitude of bursts of the oxytocin cells in the nucleus contralateral to the injection site. 5. After interhemisphere sections most oxytocin cells were silent, bursts occurred in an erratic manner, and their amplitude was attenuated and irregular (more than the 20% variation normally recorded in non-operated rats). Moreover, the amplitudes of successive bursts of pair-recorded supraoptic-supraoptic (SO-SO) oxytocin cells, highly related in control conditions (correlation coefficient, r = 0.68 to 0.98) were no longer correlated after interhemisphere section (r = 0.24 to -0.61), but all bursts remained synchronized.(ABSTRACT TRUNCATED AT 400 WORDS)