In a recent article in this journal, Francesco Bosello, Roberto Roson, and Richard Tol make the remarkable prediction that one degree of global warming will, on balance, save more than 800,000 lives annually by 2050. They introduce enormous, controversial monetary valuations of mortality and morbidity, varying with income; they then focus primarily on modeling the much smaller, indirect economic effects of the changes in health outcomes. Their calculations, large and small, are driven by the huge projected reduction in mortality-an estimate that Bosello et al. fail to substantiate. They rely on research that identifies a simple empirical relationship between temperature and mortality, but ignores the countervailing effect of human adaptation to gradual changes in average temperature. While focusing on small changes in average temperatures, they ignore the important health impacts of extreme weather events. They extrapolate the effects of small changes in average temperature far beyond the level that is apparently supported by their principal sources, and introduce arbitrary assumptions that may bias the result toward finding net health benefits from warming. (C) 2007 Elsevier B.V. All rights reserved.