Plasma levels of methotrexate (MTX) after oral administration of 6.4 to 11.2 mg/m2/week (mean 8.5 mg/m2/week) were studied in 33 children with severe juvenile rheumatoid arthritis (JRA). MTX concentrations were measured by a fluorescence polarization immunoassay (TDx) at 1, 2, 3, and 24 h after administration. The maximum level was observed in most patients after 1 h. No significant correlation was found between MTX dosage and the 1, 2, and 3-h plasma levels. No patient showed values in the range of probable toxicity 24 h after administration. A stepwise multiple regression analysis on 1, 2, and 3-h MTX levels, selected clinical features, dosage and duration of MTX therapy, and concomitant drug treatment showed that MTX concentrations at the different time points tend to be closely related; among the other variables, only concurrent treatment with salicylates was found to affect significantly the 3-h level. Serial determinations performed in 20 patients at the same oral dosage showed a wide interindividual and intraindividual variability of the plasma levels from the first dose to the next. Variable and unpredictable levels were observed also in most of the 8 patients studied after one or more increases of MTX dosage. No difference in MTX concentrations was observed between patients who responded to treatment and those who failed to respond, and between patients who had serum transaminase elevation and those who did not. Our results suggest that, until the pharmacokinetics of low dose MTX is clarified, routine therapeutic monitoring of MTX has a limited value in the clinical management of children with JRA.