Throughout this period, the research in laboratory animals was equally unsupportive of Mayer’s hypothesis. Male rats might actually limit their food intake after running for hours on a running wheel, as Mayer had suggested was possible, but they ate more on days when they didn’t exercise. They also made up for the exercise by moving less at other times. Moreover, these rats had to be forced to exercise to suppress hunger even temporarily; it did not happen voluntarily. In Mayer’s experiments, the rats were put on a motorized treadmill; they ran because they had no choice. This suggested that any decrease in appetite observed in these less-than-voluntary exercise experiments might have been induced by either stress or exhaustion rather than the exercise itself, and particularly by the use of what are technically known as shock grids to “motivate” the rats. In those experiments that relied on voluntary physical activity, the more the rats ran, the more the rats ate, and weights remained unchanged. When the rats were retired from forced-exercise programs, they ate more than ever and gained weight “more rapidly” than those rats that had been allowed to remain sedentary. With hamsters and gerbils, voluntary running activity produced “permanent increases” in body weight and adiposity—exercising made these rodents fatter, not leaner.

If Mayer’s hypothesis was true, if physical activity played a meaningful role in weight regulation, then researchers’ increasing interest in demonstrating this fact should have led, over the decades, to an unambiguous demonstration that this was the case. On the contrary. “When surveying the scientific literature on the treatment of obesity one cannot help but come away…under whelmed by the minor contribution of exercise to most weight-loss programs,” University of California, Davis, nutritionist Judith Stern, who had obtained her doctorate at Harvard with Mayer, wrote in 1986.

In the past few years, a series of authoritative reports have advocated ever more physical activity for adults—now up to ninety minutes a day of moderate-intensity exercise—but they have done so precisely because the evidence in support of the hypothesis is so unimpressive. No substantial evidence in fact supports this recommendation for weight loss or maintenance.