Parents' use of marijuana and cocaine was evaluated in a national (United States) case-control study of childhood rhabdomyosarcoma (RMS). Subjects were 322 RMS cases, aged 0-20 years, from the Intergroup Rhabdomyosarcoma Study, and 322 matched controls identified by random-digit telephone dialing. Parents of subjects were interviewed by telephone using a structured questionnaire. Mothers' marijuana use during the year before their child's birth was associated with a 3.0-fold increased risk of RMS in the child (95% confidence interval [CI] = 1.4-6.5) and maternal cocaine use was associated with a 5.1-fold increased risk (CI = 1.0-25.0). Risk was increased 3.1-fold (CI = 1.4-6.7) with use of any recreational drug. Fathers' marijuana use was associated with a 2.0-fold increased risk (CI = 1.3-3.3), cocaine use with a 2.1-fold increased risk (CI = 0.9-4.9), and use of any recreational drug with a 2.0-fold (CI = 1.3-3.3) increased risk. Case mothers' cocaine use and both parents' marijuana use were associated with their children being diagnosed at a significantly younger age. It was not possible to determine whether cocaine and marijuana have independent effects, since use of the two drugs was materially correlated. Similarly, mothers' and fathers' use of these drugs was highly correlated. In summary, parents' marijuana and cocaine use during the year preceding their child's birth may increase, by twofold to fivefold, the risk of RMS in their children.