1. Parotid secretion has been studied by micropuncture in sodium-replete and sodium-deficient sheep. 2. The osmolality of unstimulated primary saliva was slightly higher than in plasma and fell following cholinergic nerve stimulation. In sodium-depleted animals the osmolality of final saliva was hypotonic and exhibited flow dependency, where as in sodium-replete animals it was always isotonic. 3. In sodium-replete sheep, the primary fluid sodium concentration was about 120-130 mmol l-1 but in final saliva it was about 167 mmol l-1 and showed little or no flow-dependency. In sodium-depleted sheep, the primary sodium concentration averaged only 82.2 mmol l-1 and it was concluded that sodium-depleted primary fluid contained some other unidentified solute that allowed it to remain approximately isotonic; in final saliva the unstimulated sodium concentration was about 40 mmol l-1 and it rose with increasing flow rate to a maximum of 114.9 mmol l-1. 4. The primary fluid potassium concentration in sodium-replete animals did not differ significantly from that seen in sodium-depleted animals and the values were uninfluenced by stimulation; the over-all mean value was 11.2 mmol l-1. In final saliva, in sodium-replete sheep, the potassium concentrations averaged 7.8 mmol l-1 but in sodium-depleted sheep the concentrations were between 5 and 10 times greater than in primary fluid. 5. It was calculated from the equilibrium pH that the primary bicarbonate concentration would have been about 35 mmol l-1. In final saliva, where bicarbonate was measured directly, the concentrations were much greater and increased with stimulation to about 115 mmol l-1. 6. The primary fluid phosphate and chloride concentrations were the same in both sodium-replete and sodium-depleted animals and were unchanged by stimulation; the mean concentration of phosphate was 1.30 mmol l-1 and of chloride, 53.0 mmol l-1. In final saliva the phosphate concentrations were little changed but the chloride concentrations fell to an average value of 20.0 mmol l-1. In final saliva it was found that the summed sodium and potassium concentrations exceeded the summed chloride, bicarbonate and phosphate (in mequiv l-1) concentrations, on average by 13.9 mequiv l-1, regardless of sodium status or flow rate. 7. The results indicate that secretion by the sheep parotid can be accounted for in terms of the standard two-state model. Phosphate seems to enter the saliva only in the primary fluid and potassium and bicarbonate appear to enter at both primary and secondary sites; sodium and chloride enter at the primary level and can be reabsorbed in the ducts. Salt depletion causes the primary fluid concentrations of sodium and chloride to fall and the content of an unidentified solute to rise markedly while, at the ductal level, it causes normally quiescent sodium and potassium transport processes to become activated.