19 children in their first remission “fromstandard risk” lymphoblastic leukaemia had the urinary concentration of 17 oxogenic steroids (OGS) measured after two scheduled 5-day courses of prednisolone, only for the second of which were the parents aware of what the urine would be examined for. Despite a carefully standardised dose of the drug, there was a very wide scatter of OGS excretion which changed from time to time with each patient, but not in a way related to when the parents knew what was being measured. 6 low-excreting patients were reassessed in hospital with controlled compliance but only 1 produced her highest level under such circumstances. These results suggest variable bioavailability of prednisolone, but do not indicate the reason. Non-compliance does not appear to be the sole explanation.