The drinking and lysine vasopressin (LVP, porcine vasopressin) responses were measured in minipigs given simultaneous intracarotid (i.c.) (1.28 ml/min) and intracerebroventricular (i.c.v.) (50 μl/min) infusions of solutions of differing osmolality and sodium content. Observations were made during, and for a further 15 min after, the combined infusions which lasted 15 min. Drinking in response to i.c. infusion of 2.0 Osm NaCl started with a latency of 7 min and was unaffected by simultaneous i.c.v. infusion of 0.15 M NaCl, was additive with the drinking produced by i.c.v. 1.4 Osm NaCl, sucrose or mannitol, and inhibited by i.c.v. infusion of 1.4 Osm urea. LVP release following i.c. hyperosmotic NaCl was observed as early as the first blood sample, at 2 min, and was attenuated by i.c.v. hyperosmotic urea and attenuated, then slightly augmented, by i.c.v. infusions of the hyperosmotic non-electrolyte solutions. These results show that drinking following peripheral administration of an osmotic stimulus in the minipig can add to that induced by central infusions of solutions that of the cerebrospinal fluid (CSF) but may have increased or decreased CSF sodium concentration. By contrast, the stimulated LVP release was attenuated by solutions that decreased CSF sodium concentration. Thus, in minipigs, separate osmoreceptors appear to mediate osmotically induced drinking and LVP release.