This report introduces quantitative tests for the evaluation of sweating and the results obtained in 81 diabetic and 30 control subjects. The tests rely on the ability of pilocarpine, introduced into the skin by iontophoresis, to stimulate sweating from fully or partially innervated sweat glands but not from denervated glands. Many diabetic patients had a reduced number of excitable sweat glands and a low volume of sweat per square centimeter of skin. The results of the sweat tests correlated best with the clinically determined perception of pain from pinprick. The similar degree of involvement of sudomotor axons and pain-conveying axons may be related to the known similarity in size and reinnervation patterns. There was poor correlation of the sweating deficiency with alpha motor conduction velocity and with denervation of foot muscles as determined by the evoked muscle action potential. The number of excitable sweat glands was usually normal if the muscle action potential was above 0.5 mv, and often normal even when a muscle action potential was unobtainable. Every diabetic patient with abnormal sweating and several with normal sweating had reduced heart rate fluctuation during a standard Valsalva maneuver or during slow respiration.