Parts of Greece have been exposed to fallout radiation from the Chernobyl accident as much as any of the countries boardering with the former Soviet Union, because of the direction of the prevailing winds after the accident. Although fallout radiation did not reach levels expected to be associated with measurable effects, there is widespread concern in Greece that the incidence of childhood leukemia may be rising in the more heavily affected parts of Greece. Patient discharge data from all Greek hospitals treating childhood leukemia were used to calculate the annual incidence of the disease from January 1980 to June 1986 (preaccident period), from July 1986 to June 1988 (immediate postaccident period) and from July 1988 to June 1991 ("relevant" post-accident period, that accommodates the presumed latent period of the disease). Fallout radiation measurements (in Bq/kg Cs-137) were used to create 17 regions of similar (within regions) but highly variable (between regions) levels of fallout deposition. Background radiation (in Bq/kg Ra-226) and annual incidence of childhood leukemia by region were also estimated. There was no evidence of increased incidence of childhood leukemia during the immediate or the "relevant" post-Chernobyl period in any part of the country. Furthermore, regression analyses did not show any significant or suggestive association of childhood leukemia by region with either background or fallout radiation. These results indicate that the Chernobyl accident did not affect noticeably the incidence of childhood leukemia in Greece during the five-year post accident period.