1. The effects of acute I.V. administration of morphine on putative oxytocin neurones of the supraoptic nucleus were studied in urethane-anaesthetized female rats which had been exposed to I.C.V. infusion of morphine (up to 50-mu-g h-1) or vehicle for 5 days. 2. In vehicle-infused rats, I.V. morphine inhibited the spontaneous activity of six out of seven putative oxytocin neurones. Increasing doses of morphine were given, from 1-mu-g kg-1 to 5 mg kg-1. The median cumulative threshold dose to produce significant inhibition was 20-mu-g kg-1 (seven cells in six rats); six out of seven cells were inhibited at 161-mu-g kg-1. The highest doses tested inhibited by approximately 90% (excluding one unaffected cell). Inhibition was fully reversed by I.V. naloxone without overshoot, indicating a lack of acute dependence. 3. Injection of morphine I.C.V. inhibited firing at doses that wer ineffective by I.V. injection and the effects of I.C.V. morphine were reversed by I.V. naloxone 4. Acute morphine (500-mu-g kg-1 I.V.) reduced the plasma concentration of oxytocin, measured after 15 min by specific radioimmunoassay, by 34% (n = 14). 5. In lactating rats I.C.V. injection of morphine (1-2-mu-g) inhibited the activity of supraoptic neurones identified as oxytocinergic by their responses to suckling. 6. In seventeen rats infused with I.C.V. morphine the initial firing rate of twenty-eight spontaneously active, non-phasic neurones was significantly less, by 24%, than thirty-four similar cells in control rats, indicating incomplete tolerance to I.C.V. morphine. Morphine (up to 161-mu-g kg-1 given I.V.) inhibited none of nine active non-phasic neurones (P < 0.01 compared to control rats), but at higher doses inhibited four of nine cells; the overall median threshold cumulative dose (1660-mu-g kg-1) was significantly greater than in vehicle-infused controls, indicating tolerance to I.V. morphine In contrast with control rats, some cells (5/9) were modestly excited by low doses of morphine. Naloxone (5 mg kg-1 I.V.) produced withdrawal excitation: the firing rate of putative oxytocin neurones increased to approximately 260% of the pre-I.V. morphine value, indicating dependence in mechanisms regulating the firing rate of these neurones. 7. In morphine-infused rats, the basal firing rate of nineteen phasically active putative vasopressin supraoptic neurones was not different in nineteen phasic cells in controls (6.4 +/- 0.7 vs. 4.2 +/- 0.6 Hz). 8. Thus morphine potently inhibits the firing of magnocellular oxytocin neurones in the female rat, inhibiting oxytocin secretion. Morphine tolerance and dependence develop during I.C.V. infusion of morphine for 5 days. Similar tolerance to and dependence upon endogenous opioids during pregnancy may be important in the preparation of oxytocin neurones for parturition.