A noninvasive method to evaluate autonomic nervous system (ANS) function in animals is needed for studies of diabetic autonomic neuropathy. These studies modified the RR-variation test, used to test diabetic ANS function in humans, and applied it to rats. Permanent wire electrodes were implanted in the chest wall of a rat. ECG complexes were obtained by connecting the electrodes to leads going to an impedence pneumograph and high gain coupler. This information was then converted into square waves by a trigger unit and recorded on magnetic tape for subsequent analysis by computer. Recordings were at least 60 seconds long, of which 30 seconds was used for analysis. In order to establish autonomic influence, RR-variation was measured before and after application of pharmacologic agents. Directly decreasing parasympathetic tone with atropine (20 mg/kg, n = 6) increased heart rate ( P < 0.001) and decreased RR-variation ( P < 0.05). Directly decreasing beta adrenergic tone with propranolol (10 mg/kg, n = 7) decreased heart rate ( P < 0.01) but had no effect on RR-variation (NS). Stimulation of the beta adrenergic receptors (isoproterenol, 0.1 mg/kg, n = 5) increased heart rate ( P < 0.01) but decreased RR-variation ( P < 0.01). Increasing parasympathetic tone reflexly with alpha-1 adrenergic receptor stimulation (phenylephrine, 1 mg/kg, n = 7) decreased heart rate ( P < 0.05) and increased RR-variation ( P < 0.025). The responses to phenylephrine could be blocked by parasympathetic blockade. Phentolamine (0.1 mg/kg, n = 7) caused an increase in heart rate ( P < 0.001) and a decrease in RR-variation ( P < 0.01). The responses to phentolamine could be blocked by beta adrenergic receptor blockade. These results show that changes in RR-variation are due to both direct and reflex changes of ANS activity. Since propranolol had no effect on RR-variation, we conclude that during nonstressed situations RR-variation represents predominantly parasympathetic activity. However, stress (isoproterenol infusions) and blood pressure changes can alter RR-variation. Thus, RR-variation appears to be a reliable noninvasive index of autonomic activity in rats and could be used in animal studies of diabetic autonomic neuropathy.