Following oral administration considerable variation in the bioavailability of etoposide has been reported between patients and with different formulations of the drug. The variation within patients following repeated doses is unknown and has therefore been studied in seven patients receiving therapy on three successive days for relapsed small cell lung carcinoma. Etoposide was administered at a dose of 400 mg orally and plasma concentrations were measured using high-performance liquid chromatography. Within-patient coefficients of variation over three successive days ranged over 19-45% for peak plasma concentrations and 16-53% for the area under the plasma concentration-time curve. There was no evidence of a trend to suggest improving or worsening absorption and accumulation did not occur. Urinary excretion was less than 25% and showed no increase over the 3 days. These data indicate that etoposide bioavailability is not constant and oral therapy may lead to unsuspected underdosing or unexpected toxicity in schedules extending over several days. Monitoring blood concentrations for a single day following oral therapy may give a misleading idea of the total bioavailability of etoposide during a course of therapy. Studies of the relationship between the pharmacokinetics of prolonged schedules of etoposide and disease outcome may lead to unreliable conclusions unless intravenous etoposide is used.