The effect of pain on plasma AVP concentration in man has previously been studied only during major surgery with general anaesthesia. Plasma AVP concentration (pAVP) and plasma osmolality (pOsm) were measured in thirty-six patients seen in a surgical emergency department complaining of pain and in fifty-one control subjects. No significant difference in pOsm was found, but pAVP was significantly higher in the emergency room patients in pain (M +/- SEM = 4.94 +/- 0.98 pmol/1 compared to 2.31 +/- 0.32 pmol/1 in control subjects, P less than 0.01). In the control subjects, age was found to have a low but significant inverse correlation with pAVP (r = 0.37, P less than 0.01). Chronic smoking was associated with significant elevation of pAVP (3.81 +/- 0.99 pmol/1 in smokers vs. 1.89 +/- 0.28 pmol/1 in non-smokers, P less than 0.02). Neither smoking nor age could account for the difference in pAVP between the pain and control groups. Thus, pain is a non-osmolar factor capable of elevating AVP in conscious man.