In anaesthetized cats exogenous vasoactive intestinal peptide failed to elicit any secretion of saliva from the submandibular and parotid glands. However, protein release from both glands occurred in response to VIP in the presence of alpha- and beta-adrenoceptor blocking agents and was dose-dependent. This response was revealed by means of a subsequent washout flow of saliva evoked by intravenous injections of methacholine or stimulations of the parasympathetic innervation. The submandibular glands responded to vasoactive intestinal peptide at a lower dose than the parotid glands. In the presence of atropine (but in the absence of adrenoceptor blockers), stimulation of the parasympathetic chorda-lingual nerve, which of itself elicited no secretion of saliva, contributed to the release of protein within the submandibular gland, since the output of protein in response to a subsequent stimulation of the sympathetic innervation was increased. Vasoactive intestinal peptide administered in combination with methacholine or during ongoing parasympathetic nerve-induced salivary secretion revealed positive interactions, particularly with respect to protein release. In-vitro protein release in response to vasoactive intestinal peptide was also demonstrated by perfusing small pieces of the two glands in the presence of muscarinic and adrenoceptor blockers. As in vivo, submandibular tissue responded at a lower concentration of vasoactive intestinal peptide than the parotid tissue. One to two weeks after combined parasympathetic and sympathetic denervation of the parotid glands, the glands were sensitized to vasoactive intestinal peptide when tested in vitro. It is concluded that vasoactive intestinal peptide or a structurally related peptide is a potential transmitter in the parasympathetic control of protein secretion in salivary glands of cats.