This article reviews the available epidemiologic evidence for relationships between the development of cancer in childhood and environmental agents, specifically chemicals, ionizing radiation, low-frequency electromagnetic fields, and infectious agents. Chemical exposures include medications and other drugs, components of diet, and second-hand exposure to industrial chemicals and to environmental carcinogens. Ionizing radiation is the only well-established risk factor for childhood cancers. Suggestive associations with excess childhood cancer risk have been found with exposures to paints, petroleum products, solvents, pesticides and metals. An excess risk of brain tumours has been reported with ingestion of n-nitroso compounds, and there have been positive findings relating leukemia risk and infections. Studies of risk of childhood cancer with parental use of alcohol and with parental smoking have been generally negative, while the inconsistencies and generally low risks reported in studies of power-frequency electromagnetic fields do not support a causal relationship.