A cross-sectional study is reported in which 53 patients with Holmes-Adie syndrome have been subjected to a battery of tests of autonomic nervous function referable to the cardiovascular system, to two objective tests of sweating function, and to subjective assessment of sweating by application of quinizarin powder followed by body heating. The majority of patients were consecutive referrals; none was selected because of clinical indications of autonomic dysfunction. Eighty three per cent of these patients had at least one, 57% at least two, and 40% at least three objective test abnormalities, as defined by values lying outside 95 percentiles of healthy subjects who were matched for age and subjected to the same tests. In the context of multiple testing, the probability of finding outside values was such that a minimum of 3 was required to define abnormality. On this basis 40% of patients were found to have significant evidence of autonomic dysfunction. The most frequent abnormalities were impaired digital vasoconstriction to cold (23%), a reduced heart rate response to the Valsalva manoeuvre (17%), and excessive variability in sweating between test sites (in one of the tests, 43%) which is consistent with patchy loss. Abnormal quinizarin test appearances were seen in 10 patients and in a further five patients the appearances were thought to be suggestive of abnormality. Though assessment of the results of this test are subjective, the observations are consistent with the findings obtained from the objective tests which were applied. Cardiovascular and sweating abnormality did not concur significantly and only the former was found to increase progressively with known duration of the pupillotonia. It is concluded that Holmes-Adie syndrome is commonly accompanied by progressive mild but widespread autonomic involvement but rarely is this symptomatic. If symptoms suggestive of autonomic neuropathy are found in a patient with tonic pupils, a careful search for some other generalised disorder is recommended.