Bone marrow aspiration is an old and well-established test procedure, the use of which may rest more on diagnostic tradition than on scientific documentation. Its use was evaluated in a Norwegian regional hospital by registering indication, presence of pathology and clinical "utility" judged from a given definition, in 200 reports on bone marrow aspiration. The result was pathological in 53%, and judged to be clinically useful in 61% of the tests. The diagnostic yield was highest for indications related to bone marrow function, and low for unspecific indications such as fever of unknown origin and elevated SR without M-component. The large number of results with pathological findings that are judged to be clinically useful raise the question of whether use of this test should be increased.