SUMMARY1. One hundred and thirty-two antidromically identified paraventricular (PV) and supraoptic (SO) neurones were studied during milk ejection evoked by the suckling of the young, in thirty-four lactating rats anaesthetized with urethane (1.1 g/kg i.P.). The spike activity of these neurones was analysed to determine some of the factors governing their orthodromic activation. Changes in spike activity were compared with the amount of oxytocin released from the neurohypophysis, as measured by the rise in intramammary pressure at milk ejection.2. Sixty-five units, from both the PV and SO nuclei, displayed a characteristic burst of accelerated activity 12-18 sec before milk ejection. With a suckling stimulus of nine or ten pups, these neurosecretory responses were both regular in their occurrence (recurring every 4-8 min) and uniform in character, i.e. all responsive cells accelerated to 30-80 spikes/sec for 1-4 sec. Two units deviated from this pattern and both were recorded in animals which failed to display a detectable milk ejection. Both cells displayed diminutive responses (10-20 spikes/sec, for 0 5 sec) at regular intervals.3. The number of suckling pups was the most important factor controlling both the occurrence and expression of the neurosecretory response. When successive pups were removed from the nipples there was a progressive and linearly related fall in the magnitude of the unit responses, i.e. the peak firing rate and the number of spikes in each response declined. Neurosecretory activation was normally not observed in anaesthetized and stereotaxically restrained animals when six pups or less were placed to suckle. D. W. LINCOLN AND J. B. WAKERLEY neurosecretory activation had an important influence on the expression of the unit response. On a population basis and ignoring variations created by differences in the number of suckling young, units with background firing rates of > 1*0 spikes/sec displayed responses that were significantly (P < 0.05) greater (peak rate x number of spikes) than those units firing at < 1-0 spikes/sec. This correlation was also observed with single units, particularly where large fluctuations in background firing rate were encountered.5. When five to seven pups were placed to suckle -a threshold number, two factors were observed to occasionally trigger a neurosecretory response. These were the injection of 0 05 ml. saline into one of the cannulated mammary glands, and the addition or removal of a pup. In either case the response was immediate and occurred with both ipsilateral and contralateral stimulation.6. The amount of oxytocin released in response to each period of neurosecretory activation was always proportional to the magnitude of the response displayed by the individual unit, thus a larger pulse of oxytocin was released in response to the suckling of nine pups than seven pups. All suckling-induced milk ejections were the result of oxytocin being released in pulses of 0-2-1'5 m-u.7. We found no conclusive evidence to suggest that the number of pups or the suck...