It has been postulated that variations in methotrexate absorption may influence the outcome of treatment in lymphoblastic leukaemia. One hundred and forty four children with acute lymphoblastic leukaemia not of the T cell type were randomised to receive continuing treatment with daily 6-mercaptopurine, vincristine, and prednisolone six weekly and methotrexate once weekly, either as a single oral dose or an intramuscular injection. Analysis of results with a minimum follow up of three and a half years has shown that the route of administration of methotrexate has had no influence on relapse at any site, but more children receiving intramuscular methotrexate died in remission. These results indicate that oral methotrexate is as effective as intramuscular methotrexate in continuing treatment of lymphoblastic leukaemia.