The child survival hypothesis is immensely popular with politicians, religious leaders, and executives of organizations engaged in foreign philanthropy, because it justifies the anti-Malthusian and tender-minded belief that reducing infant mortality will automatically bring about a reduction in fertility. The belief easily converts into policy, because saving babies is something we know how to do. Nonetheless, the term child survival hypothesis is not widely known outside professional circles. By contrast, the theory of the "demographic transition" has been extensively popularized over several decades. Its meaning can, however, stand a bit of clarification. The theory was born French: in 1934 Adolphe Landry wrote of the revolution demographique.' A decade later this was translated into the familiar English form. By 1969 a widely used population textbook expressed the common, if not the predominant, opinion of demographers when it identified the theory as "one of the best documented generalizations in the social sciences." Documented it certainly is: the literature is appallingly large. But documented does not mean proved. Ironically (in the words of demographer Michael Teitelbaum), "its explanatory power has come into increasing scientific doubt at the very time that it is achieving its greatest acceptance by nonscientists. In scientific circles, only modest claims are now made for transition theory." That was said in 1975. Ten years later Teitelbaum and Winter put the matter more forcefully: "It is doubtful whether this theory was ever truly a theory at all (that is, a set of hypotheses with predictive force)." Before we look into its predictive abilities we need to find out exactly what the theory asserts. This is not easy because the theory is almost never carefully and rigorously described. We need once more to call upon the art of graphing. Transition theory assumes a finite world. For most of the world, most of the time, both birth rate and death rate have been in the neighborhood of 40 per thousand population per year.5 When the two rates are equal, ZPG (zero population growth) prevails. Despite perennial fluctuations in population size at different locations, the average growth rate of the entire human population for the past million years has been very close to ZPG, namely 0.02 percent per year. At this growth rate the population doubles every 3,500 years—hardly a population explosion!