Residential proximity to electrical power lines of different voltage in relation to childhood leukemia was investigated through a case-control study undertaken in Greece during 1993-1994. The study comprised 117 incident cases of childhood leukemia and 202 age-, gender- and place-of-residence-matched controls. Four measures of exposure to magnetic fields were developed, using data provided by the Public Power Corporation of Greece: Voltage (V) divided by the distance (d), V/d2, V/d3 and an adaptation of the Wertheimer-Leeper code. Conditional-logistic-regression modeling was used to adjust for potential confounding influences of 18 variables. No significant trends of childhood leukemia risk with increasing exposure levels were noted, nor were there statistically significant elevations of disease risk at the higher exposure levels in each measure of exposure. These results do not support a causal link between residential proximity to electrical high-voltage wires and childhood leukemia risk, but in themselves do not refute a weak empirical association.