SUMMARY1. The effects of continuous stimulation of the peripheral end of the ascending cervical sympathetic nerve were compared with those of intermittent stimulation, so arranged as to deliver the same total number of impulses, in cats under chloralose anaesthesia.2. Continuous stimulation caused a flow of saliva at 5-10 Hz, but not at 2 Hz. In contrast, the same total number of impulses delivered intermittently in bursts elicited a prompt secretion at a frequency as low as 20 Hz for 1 s at 10 s intervals (corresponding to 2 Hz continuously) and a significantly higher rate of secretion at 50 Hz in bursts than that obtained in response to 5 Hz continuously.3. Continuous stimulation also caused a rise in submandibular vascular resistance (s.v.r.), which persisted throughout the period of stimulation, and was followed immediately thereafter by an intense but transient fall in s.v.r. During stimulation in 1 s bursts, each burst was followed first by a brief rise in s.v.r. and shortly after by a fall. The balance between these two components varied widely between individual animals but often led to an overall fall in s.v.r. during stimulation i.e. complete reversal of the mean vascular effect. A further fall in s.v.r. was then recorded when the stimulus was discontinued.4. Propranolol (1-0 mg/kg) reduced but failed to abolish the secretary response. It also altered the balance between the two phases of the vascular response slightly in favour of a rise in s.v.r. during stimulation, without apparently affecting the size of the after-dilatation.5. Pre-treatment with dihydroergotamine (1-0 mg/kg) invariably blocked secretion and revealed a small vasodilator response during sympathetic stimulation with either pattern of stimulation; it also blocked the after-dilatation.6. Following combined pre-treatment with propranolol and dihydroergotamine, to produce total adrenergic blockade, there was a small residual vasoconstrictor component which amounted to an increase in mean s.v.r. of about 20% during stimulation at 10 Hz continuously. This may have been due to release ofneuropeptide Y (NPY).7. Small but significantly greater amounts of NPY were released into the effluent S. R. BLOOM, A. V. EDWARDS AND J. R. GARRETT blood during stimulation of the ascending cervical sympathetic nerve at 70 Hz in bursts than during continuous stimulation. No significant release of vasoactive intestinal peptide (VIP), somatostatin, bombesin, substance P or calcitonin generelated peptide (CGRP) was observed during stimulation at any frequency.8. It is concluded that, in addition to a vasodilator component due to ,-adrenoceptor activation, stimulation of the sympathetic innervation to the submandibular gland in the cat elicits a vasodilator response by some pathway that depends upon a-adrenoceptor activation. This is sufficiently potent to override aadrenergic vasoconstriction during certain patterns of stimulation and is mainly responsible for the transient dilatation which occurs on cessation of stimulation. These vasodilator effects may invol...