Experiments were performed on weaned lambs and adult sheep of various breeds (15-52 kg body weight)in the U.K. Experimental procedures were performed as previously described (Buckle et al. 1995), comparing stimulation of the chorda-lingual nerve continuously at 4Hz with the equivalent intermittent pattern of bursts of 40 Hz for 1 second in every 10 (40Hz 1:10). Anaesthesia was induced and maintained with sodium pentobarbitone (1530mg kg^-1 I.V. then 0.1-0.3 mg min^-1 kg^-1 I.V.). Animals were eventually killed by lethal injection of barbiturate (ca. 15ml 20% I.V.). Data are given as mean±S.E.M. and analysed by paired and unpaired Students t-test; n = number of animals. The fall in submandibular vascular resistance was unaffected by the pattern of stimulation, (−69.4±2.7% at 4Hz; −71.4%±1.7% at 40Hz 1:10, n=6). Blockade of de novo synthesis of nitric oxide attenuated the vascular response at 4Hz, (−69.4±2.7%; −44.4±6.2% before & after L-NAME, P<0.05), and at 40Hz 1:10, (−71.4%±1.7%; − 46.6±8.4% before & after L-NAME, P<0.05, n=6) to a similar degree. Subsequent administration of atropine did not attenuate the vascular response further, (42.5±6.0% at 4Hz, 38.2±5.36 at 40Hz 1:10, n=3). Continuous stimulation caused a significantly greater fluid secretion than intermittent, (0.047±0.004; 0.038±0.005 ml min^-1 [g gland]^-1 at 4Hz and 40Hz 1:10 respectively, P<0.05, n=6). Blockade of de novo synthesis of nitric oxide did not affect the volume of secretion at 4Hz, (0.047±0.004; 0.040±0.005 ml min^-1 [g gland]^-1 before & after L-NAME) or 40Hz 1:10, (0.038±0.005; 0.037±0.005 ml min^-1 [g gland]^-1 before & after L-NAME). Subsequent administration of atropine reduced the salivary flow to only a few drops in a ten minute period. Administration of L-NAME significantly reduced the protein output at 4Hz, (50.9±5.6; 42.2±5.3 µg min^-1 [g gland]^-1 before & after L-NAME, P<0.05, n=6), but had no effect at 40Hz 1:10, (55.0±10.8; 59.1±12.6 µg min^-1 [g gland]^-1 before & after L-NAME). With continuous stimulation at 4Hz the secretion of fluid is known to be nitric oxide-independent and almost entirely cholinergic. Secretion of protein can be achieved by both nitric oxide-dependent and -independent mechanisms, demonstrated by the difference in protein output between continuous and intermittent patterns of stimulation. We suggest either that the secretion of protein is predominantly mediated by the release of a neurotransmitter with a nitric oxide-dependent 2nd messenger pathway, or that nitrergic facilitation of neurotransmitter release occurs, when stimulation is continuous, but not when it is intermittent. The vascular response is partially nitric oxide dependent.