In Denmark one surgical team, during the last 7 years, has performed about 80 per cent of all acoustic neuroma surgery. Because of this centralization, in such a limited population as that of Denmark, we have attempted to make a epidemiological survey of all diagnosed tumours in the period from 1976 to 1983. Systematic and prospective records were made by the authors of all patients with translabyrinthine operations, and data on patients with suboccipital operations were collected retrospectively from the six neurosurgical departments in Denmark. The average annual incidence was 8 tumours per million inhabitants, with the highest incidence of approximately 13 tumours per million occurring in Copenhagen County. The incidence reported in previously published autopsy series is 800-900 times higher and the following may serve as an explanation for this enormous difference: Autopsy series are in all probability based on highly selected cases; they are predominantly based on elderly people and the incidence is not directly applicable to the population at large. Several of the silent tumours from the autopsy series were located in the cochlea or in the labyrinth and not in the internal ear canal. The conclusion from our study is that both the knowledge and data available at present are insufficient to serve as a basis for an actual calculation of incidence and prevalence of acoustic neuromas.(ABSTRACT TRUNCATED AT 250 WORDS)