In a prospective evaluation of percutaneous Trucut needle biopsy for localised intrathoracic lesions in 89 patients histological specimens were obtained in 81. Malignancy was diagnosed in 66 cases. Subsequently definitive histological reports were available in 18 of these patients with complete concordance of malignant cell type. Sixteen patients had non-malignant histological appearances, which were later confirmed objectively in six. In three patients there was no follow up information, but in the remainder the clinical course was entirely consistent with the histological appearances of the biopsy specimens. Adequate specimens were obtained from only two of the five lesions less than 2 cm in diameter. Lesions deeper than 8 cm from the site of biopsy were associated with significantly more haemorrhagic complications than more superficial lesions. Comparison with other series indicates that Trucut needle biopsy which produces histological specimens has greater diagnostic accuracy than cytological techniques for both malignant and non-malignant localised lesions. It is concluded that this technique has a definite place in the investigation of this common problem in carefully selected patients provided that strict attention is paid to the details of the technique.