Epidemiologic and experimental research on the potential carcinogenic effects of extremely low frequency electromagnetic elds (EMF) has now been conducted for over two decades. Cancer epidemiology studies in relation to EMF have focused primarily on brain cancer and leukemia, both from residential sources of exposure in children and adults and from occupational exposure in adult men. Because genotoxic effects of EMF have not been shown, most recent laboratory research has attempted to show biological effects that could be related to cancer promotion. In this report, we brie y review residential and occupational EMF studies on brain cancer. We also provide a general review of experimental studies as they relate both to the biological plausibility of an EMF-brain cancer relation and to the insuf ciency of such research to help guide exposure assessment in epidemiologic studies. We conclude from our review that no recent research, either epidemiologic or experimental, has emerged to provide reasonable support for a causal role of EMF on brain cancer. Neuro-Oncology 1, 212-220, 1999 (Posted to Neuro-Oncology [serial online], Doc. 99-11, June 3, 1999. URL ) B ecause of the relatively poor survival probability, and the substantial risk of long-term morbidity among survivors, brain cancer is a particularly ominous disease. In the United States alone, an estimated 16,800 primary CNS malignancies will be diagnosed in 1999 (Landis et al., 1999). Since an estimated 45% of primary CNS tumors are of nonmalignant pathology (benign or uncertain pathologic behavior) (Davis et al., 1999), the total number of people diagnosed in the U.S. with malignant and nonmalignant CNS tumors in 1999 should exceed 30,000. Although the numbers are sobering, CNS malignancies represent only about 1.4% of the total burden of malignant neoplasms in the U.S. Proposed epidemiologic studies of CNS cancer have not, as a result, enjoyed the relatively high funding priority of lung, breast, prostate, colon, or several other common adult epithelial neoplasms. Additionally, since there is no comprehensive nationwide cancer registry, the relative rarity and the high degree of histologic heterogeneity of CNS tumors make epidemiologic studies inherently difficult to conduct. Etiologic studies of CNS cancer in human populations, therefore, have been quite limited.From studies that have been conducted, no causal factor has been identi ed that explains even a modest proportion of CNS cancer occurrence (Davis and PrestonMartin, 1998; Kuijten and Bunin, 1993). The few known risk factors, most of which are related to genetic diseases that impart a CNS tumor susceptibility, likely account for not more than 10% of incident cases. Great interest, therefore, was generated when a ubiquitous environmental exposure, power frequency electromagnetic elds, was rst reported in 1979 to be a possible cause of brain cancer (and leukemia) in children (Wertheimer and Leeper, 1979). A few years later, an occupational study found leukemia mortalit...