The pharmacokinetics of serum methotrexate were studied in 45 bladder cancer patients receiving 250 mg. per m.2 as part of the initial cycle of combination chemotherapy. Serum methotrexate was determined routinely 43 to 49 hours after administration. If the methotrexate levels remained at more than 80 nmol. per l. measurements were repeated daily until the serum levels decreased below this point. The patients were classified into group 1-23 with a bladder in situ and no ureteral obstruction, group 2-11 with a bladder in situ and unilateral hydronephrosis, and group 3-11 who had had cystectomy and ileal conduit diversion before chemotherapy. Of the patients in groups 1 and 2, 5 and 6, respectively, had serum methotrexate levels of 80 nmol. per l. or more 43 to 49 hours after administration, which decreased to below this level on the next day. Of the 11 patients in group 3, 8 had elevated methotrexate levels at the initial determination. Daily methotrexate analyses showed a delayed elimination in 4 of 7 patients and levels of more than 80 nmol. per l. for 3 to 9 days. Low creatinine clearance but, in particular, the previous performance of an ileal conduit predicted high methotrexate levels on day 2 after treatment. The most likely explanation for this observation is the resorption of methotrexate by the small bowel mucosa in the ileal conduit. Patients with an ileal conduit performed 2 years or less before chemotherapy and/or those with a long ileal segment seem to have a particularly high risk for delayed methotrexate elimination. Bladder cancer patients with an ileal conduit who receive methotrexate-containing chemotherapy have a high risk of delayed methotrexate elimination and increased clinical methotrexate toxicity. Leukovorin rescue should be used liberally in these patients together with other prophylactic means (intensive hydration and alkalization of the urine).