Problem: To develop a new measurement of caloric-induced nystagmus that is more physiologic and less variable than the maximum slow phase speed. Methods: The study design consisted of signal analysis of intra- and intersubject normal caloric responses. The setting was the tertiary medical Otology/Neurotology referral center. Data were collected from 10 normal subjects and from 10 caloric tests on the same subject. The main outcome measure was the maximum slow phase speed and the acceleration of the nystagmus response. The latter was measured by linear fitting of the slow phase speed data of each response. The mean, standard deviation, and the coefficient of variation of maximum speed and acceleration measures were calculated. The intervention was the standard caloric test using water irrigation at 44 and 30 degrees Celsius. Data were digitally processed and analyzed both automatically and under visual control. Results: The variability of nystagmus acceleration is 50% less than that of nystagmus velocity in both inter- and intrasubject results. Conclusion: A new caloric test parameter, namely nystagmus acceleration, is introduced. This parameter measures the rate of slow phase velocity increase and therefore is more physiologic as it reflects endolymph acceleration. The degree of nystagmus acceleration variability is 50% less than that of nystagmus velocity. Most current ENG analysis algorithms can be easily modified to generate this new parameter. Significance: The study provides increased accuracy in measuring the caloric response. Support: None reported.