OBJECTIVE To evaluate success in obtaining adequate bone marrow trephine biopsy cores from children. METHODS Sections of trephine biopsy cores submitted by 25 centres from children with neuroblastoma over a five year period were reviewed centrally. In cores containing no tumour adequacy was defined as 0.5 cm of well preserved bone marrow after processing. Occasional smaller cores containing obvious tumour were also considered adequate. RESULTS Of 822 biopsy specimens, 139 (17%) were inadequate. In 13 centres submitting at least 20 cores failure rates ranged from 2.6 to 50%. There was no improvement over the five years of the study. There was no practically important correlation between the numbers of cores submitted and success in obtaining adequate specimens. Although a lower rate of inadequate biopsy specimens was found when haematologists rather than paediatricians (13 v 29%) were the predominant operators this should not be overinterpreted, not least because of the potentially confounding association between haematologist operators and larger numbers of biopsy specimens, and because the arbitrary subdivision of centres according to operator specialty was crude. The skill of individual operators could not be assessed. CONCLUSIONS Many operators do not obtain adequate bone marrow biopsy specimens from children. Improvement is necessary because this is an invasive investigation, often performed under general anaesthesia. Reporting pathologists are well placed to influence practice by pointing out inadequacies in the specimen and suggesting retraining or even a change in operator. Improvement would almost certainly occur if this investigation was restricted to locally recognised successful operators, whatever their specialty. Most centres should review their practice and devise strategies to improve their ability to obtain adequate cores.