1. Salivation has been studied in the submandibular gland of the rabbit. A very slow spontaneous salivation took place when all possibility of nerve influence had been excluded. Salivation was not due to ultrafiltration.2. The ;spontaneous' saliva had a mean K concentration of 148 mM and Na concentration of 46 mM. With increasing salivation rate produced by parasympathetic nerve stimulation, K concentration fell to a plateau level of about 30 mM whilst Na concentration fell rapidly to reach the low value of 3 mM, then began to rise again at the higher flow rates.3. Ligation of the submandibular duct produced a reversal of the ion concentrations in spontaneous saliva. By 4 days K concentration was lower and that of Na higher than control values until by 2 weeks the effect was maximal with mean concentrations of 25 mM for K and 153 mM for Na.4. Ouabain increased the spontaneous salivation rate and ethacrynic acid slowed or prevented it altogether. On the basis of the known sites of action of these drugs it is postulated that two pumps are involved in the regulation of spontaneous salivation. There appears to be basal activity of an acinar mechanism pumping NaCl into the lumen, taking water with it. This pump is activated directly or indirectly by the intracellular Na concentration which itself is controlled by an Na-K exchange pump.5. Excitation of the sympathetic trunk produced a small, though definite, increase in salivation rate. There was evidence that myoepithelial cells might also be involved in the sympathetic response and that they were activated by alpha receptor stimulation. Salivation evoked by sympathetic nerve stimulation would seem to be a response to beta receptor stimulation, but the possibility that activation of both alpha and beta receptors was required could not be excluded entirely.