Four hundred and fifty-four unselected consecutive patients who underwent fine-needle aspiration biopsy between 1978 and 1983 were reviewed restrospectively. The overall sensitivity of the technique in the diagnosis of malignancy was 77%. When pancreatic and biliary masses were excluded, the sensitivity was 84%. The predictive value of a positive was 100% and of a negative, 65%. Reasons for false negative biopsies are analysed. Ways of increasing the sensitivity of the procedure and improving biopsy techniques are discussed. No significant complications were encountered. We conclude that the technique is valuable, safe and can be performed in any hospital with basic ultrasonographic and radiological equipment, provided that there is a cytology service.