For determination of the dynamics of parotid urea secretion in conscious sheep, a previously standardized transit time ultrasonic flow metering system was used to measure bilateral parotid flow. Six ewes fed for ad libitum consumption were prepared under halothane anesthesia with ultrasonic probes around both parotid ducts; these ducts were also cannulated orally. After probe encapsulation (8 d), parotid flows were recorded during 24 h, and samples of saliva and blood for urea determination were obtained hourly. Jaw movements were recorded by means of a submandibular balloon to monitor feeding behavior. Urea concentration in parotid saliva was 60 to 74% of that in plasma (a positive linear correlation existed) and was poorly influenced by the parotid flow. The amount of urea secreted with parotid saliva was directly related to the salivation rate. To calculate the urea secretion in parotid saliva, a multiple linear regression model was developed from computer-calculated parotid flows over 1-min periods and plasma urea concentration. The model was accurate because the plot of calculated vs measured values was not significantly different from the line of identity. The daily parotid urea N varied from .35 to 1.02 g among ewes. The higher urea secretion rate found during rumination and eating (1.32+/-.42 and .98 +/-.33 mg/min, respectively) vs. during rest (.60+/-.39 mg/ min, P<.05) was due to higher salivation rates (5.17 +/-1.46, 3.56+/-.90, and 2.04+/-.52 mL/min, respectively, P<.05) rather than to changes in saliva urea concentrations (saliva:plasma urea ratio = .65+/-.04, .67+/-.04, and .68+/-.03, respectively). Of the daily parotid urea output, 40.8% was secreted during rest. The contribution of parotid urea N to the ruminal N pool was relatively small (1.2 to 3.7% of the N intake, which was 23.0 to 33.6 g/d). These techniques allowed direct and precise measurements of parotid urea secretion without disturbing the animal or altering the physiological regulation of salivary secretion.