Many anticancer drugs are myelotoxic and cause bone marrow depression; however, generally, the marrow/blood returns to normal after treatment. Nevertheless, after the administration of some anti-neoplastic agents (e.g. busulphan, BU) under conditions as yet undefined, the marrow may begin a return towards normal, but normality may not be achieved, and late-stage/residual marrow injury may be evident. The present studies were conducted to develop a short-term mouse model (a 'screen') to identify late-stage/residual marrow injury using a brief regimen of drug administration. Female BALB/c mice were treated with BU, doxorubicin (DOX), cisplatin (CISPLAT) or cyclophosphamide (CYCLOPHOS) on days 1, 3 and 5. In 'preliminary studies', a maximum tolerated dose (MTD) for each drug was determined for use in 'main studies'. In main studies, mice were treated with vehicle (control), low and high (the MTD) dose levels of each agent. Necropsies were performed, and blood parameters and femoral/humeral nucleated marrow cell counts (FNCC/HNCC) were assessed on six occasions (from days 1 to 60/61 post-dosing). Late-stage/residual changes were apparent in BU-treated mice at day 61 post-dosing: RBC, Hb and haematocrit were reduced, mean cell volume/mean cell haemoglobin were increased and platelet and FNCC counts were decreased. Mice given DOX, CISPLAT and CYCLOPHOS, in general, showed no clear late-stage/residual effects (day 60/61). It was concluded that a brief regimen of drug administration, at an MTD, with assessment at day 60/61 post-dosing was a suitable short-term method/screen in the mouse for detecting late-stage/residual marrow injury for BU, a drug shown to exhibit these effects in man.