Only in Sweden is skin disease reputedly caused by working at visual display units (VDUs) a considerable problem. In the present work facial skin complaints in VDU-exposed persons have been studied by means of epidemiological, clinical and histopathological methods. The results can be summarized as follows. Half the patients referred for dermatological examination because of facial skin complaints they attributed to working at VDUs had rosacea. Most of them claimed that they had pronounced symptoms, but their skin lesions were mild. When they were followed up most of the patients reported improvement without pharmacological therapy in spite of continued VDU work. Individuals exposed to VDUs had significantly more self-reported skin complaints than a nonexposed control population of office employees, but no more visible signs. Skin tumours with malignant potential and unilateral facial skin signs were found in the same frequencies in both categories. Rosacea was the most common facial skin condition (10%) in the 809 randomly selected office employees (mean age 43 years) examined, and this condition as well as the occurrence of telangiectases (found in 57% of the examined population) was no more common among the exposed persons. The answers in the self-administered questionnaires (n = 3745), used when studying the office employees, corresponded poorly to skin status (only 46% correlated), but were reasonably accurate as regards the occurrence of more clear-cut symptoms, such as in atopic or seborrhoeic dermatitis. Patients with skin complaints they said were associated with VDUs had no characteristic histological changes in facial skin. Histological changes of varying degree were also common among the control subjects without skin complaints or exposure to VDUs. This is particularly true of the degree of degenerative change in elastic fibers, which increased with age. Rosacea was more common in women, and there was a strong association between rosacea and migraine, as well as between migraine and patients with symptoms thought to be related to work at VDUs. In contrast to previous assumptions, most subjects with rosacea in this study experienced no impairment due to exposure to sunlight. This work shows that people exposed to VDUs had subjective skin symptoms more often than a nonexposed control population. The design of the study does not permit an analysis of the cause of this phenomenon. The occurrence of skin signs in persons exposed to VDUs is, however, fortuitous from an epidemiological viewpoint.