Using a recently developed technique for the determination of the pure bone marrow cell count per mm3 bone marrow, two different induction regimens for AML were compared for their efficacy, i.e. the therapy induced cytoreduction. In both protocols cytosine arabinoside, daunorubicin and 6-thioguanine were applied at comparable doses but different schedules. In both regimens a blast cell reduction of more than 2.0 log10 within the first 5 to 6 days of therapy strongly correlated (p less than 0.001) with the achievement of an adequate blast clearance in the marrow and complete remission (n = 18), while an inadequate blast cell reduction was revealed in all patients with a cell kill of less than 2.0 log10 (n = 9). Total cytoreduction on day 5 or 6 of therapy was comparable for both protocols. However, on one protocol a continuous logarithmic blast cell reduction was found while the cell kill followed a biphasic function on the other regimen.