Various hypertonic solutions were infused in healthy human volunteers to determine their effect on thirst and vasopressin secretion. Hypertonic saline and mannitol produced prompt and parallel increases in plasma osmolality and vasopressin concentration. For both of these solutes, there was a high degree of correlation between these measurements. The slope describing this relationship varied considerably between individuals, but the same subjects showed similar slopes with either saline or mannitol. Both solutions stimulated thirst. Hypertonic urea infusions produced a comparable rise in osmolality but produced a smaller increase in plasma vasopressin and stimulated thirst in only one of the subjects. With urea, the correlation between plasma osmolality and vasopressin was significantly lower. Within individuals, the slope describing this relationship was significantly correlated with that seen during hypertonic saline. Hypertonic glucose significantly increased plasma osmolality but decreased plasma vasopressin and had no detectable effect on thirst. We conclude that osmoregulation of vasopressin in humans is mediated by a selective osmoreceptor that is located primarily outside of the blood-brain barrier and that individual differences in osmoregulatory sensitivity are not solute specific.