The electronystagmograms on 400 sequential patients who completed bithermal caloric tests were examined. It was found that taking +/- 25 per cent as the limit for bithermal caloric testing and +/- 15 per cent for monothermal hot caloric testing, the hot caloric screen detected almost all patients who gave abnormal bithermal caloric test results without producing an excess of false positive results. Cold caloric screening tests were shown to be inaccurate enough to be unsatisfactory. When the speed of the slow phase was less than 11 degree per sec on either hot caloric test, cold calories were done even if the monothermal hot tests were the same, as a bilateral caloric reduction was sometimes detected.