At the commencement of UKALL XI, a national MRC trial for childhood lymphoblastic leukaemia (ALL), the therapy included a bolus of daunorubicin (DR) on the first 2 d of the protocol. This component of treatment was subsequently withdrawn because of concern about long-term cardiotoxicity. All children both before and after this change of policy had their marrow status at the end of the first week assessed by central review as part of the trial to examine the clinical importance of the rate of disease clearance. This also afforded an opportunity to observe the effect of DR on gross residual disease at an early stage of therapy. 1419 children were studied: 342 received DR ('recipients'), 1077 did not. 44% of the recipients completely cleared their marrow of blast cells after 8 d compared with 13% of the non-recipients (chi2 = 158.2, P < 0.0001). The difference in the proportion with massive residual disease (>80% blasts) was less impressive but there was still a difference in favour of DR recipients (DR 9%, no DR 15%; chi2 = 7.7, P = 0.006). The rate of disease clearance correlated with disease-free survival for both recipients and non-recipients, but there was no significant difference in outcome when comparing the two groups with each other, either in terms of disease-free or relapse-free survival. DR accelerated the rate of blast cell disappearance from the marrow but the difference this made to disease free survival is small or non-existent. It appears to be the relative speed of response to a given therapeutic regimen that is prognostically important rather than the absolute rate of response when comparing one treatment with another.