The caloric test is still the most useful vestibular test by far. Rotatory tests may allow more accurate control of the stimulus and the response parameters may vary less in normal individuals, but only the caloric test allows individual testing of each ear. By various means described in this article, it is possible to minimize potential sources of error and thus maximize the reliability of the test. Although the test allows a good measure of horizontal semicircular canal function and, by inference, an estimate of labyrinthine disease, the fact that the reflex arc traverses the vestibular nerve and brain stem renders this pathway liable to damage by central vestibular disease. Thus the electronystagmographic recording of the caloric nystagmus response may show a variety of changes, depending on the site and nature of the central vestibular lesion. Some of these changes are discussed.