SAM FELTHAM, A qualified master personal trainer, has worked in the U.K. health-and-fitness industry for more than a decade. Not accepting the caloric-reduction theory, he set out to prove it false, following the grand scientific tradition of self-experimentation. In a modern twist to the classic overeating experiments, Feltham decided that he would eat 5794 calories per day and document his weight gain. But the diet he chose was not a random 5794 calories. He followed a low-carbohydrate, high-fat diet of natural foods for twenty-one days. Feltham believed, based on clinical experience, that refined carbohydrates, not total calories, caused weight gain. The macronutrient breakdown of his diet was 10 percent carbohydrate, 53 percent fat and 37 percent protein. Standard calorie calculations predicted a weight gain of about 16 pounds (7.3 kilograms). Actual weight gain, however, was only about 2.8 pounds (1.3 kilograms). Even more interesting, he dropped more than 1 inch (2.5 centimeters) from his waist measurement. He gained weight, but it was lean mass.

      Perhaps Feltham was simply one of those genetic-lottery people who are able to eat anything and not gain weight. So, in the next experiment, Feltham abandoned the low-carb, high-fat diet. Instead, for twenty-one days, he ate 5793 calories per day of a standard American diet with lots of highly processed “fake” foods. The macronutrient breakdown of his new diet was 64 percent carbs, 22 percent fat and 14 percent protein—remarkably similar to the U.S. Dietary Guidelines. This time, the weight gain almost exactly mirrors that predicted by the calorie formula—15.6 pounds (7.1 kilograms). His waist size positively ballooned by 3.6 inches (9.14 centimeters). After only three weeks, Feltham was developing love handles.

      In the same person and with an almost identical caloric intake, the two different diets produced strikingly different results. Clearly, something more than calories is at work here since diet composition apparently plays a large role. The overfeeding paradox is that excess calories alone are not sufficient for weight gain—in contradiction to the caloric-reduction theory.

       OVERFEEDING EXPERIMENTS: UNEXPECTED RESULTS

      THE HYPOTHESIS THAT eating too much causes obesity is easily testable. You simply take a group of volunteers, deliberately overfeed them and watch what happens. If the hypothesis is true, the result should be obesity.

      Luckily for us, such experiments have already been done. Dr. Ethan Sims performed the most famous of these studies in the late 1960s.^{1, 2} He tried to force mice to gain weight. Despite ample food, the mice ate only enough to be full. After that, no inducement could get them to eat. They would not become obese. Force-feeding the mice caused an increase in their metabolism, so once again, no weight was gained. Sims then asked a devastatingly brilliant question: Could he make humans deliberately gain weight? This question, so deceptively simple, had never before been experimentally answered. After all, we already thought we knew the answer. Of course overfeeding would lead to obesity.

      But does it really? Sims recruited lean college students at the nearby University of Vermont and encouraged them to eat whatever they wanted to gain weight. But despite what both he and the students had expected, the students could not become obese. To his utter amazement, it wasn’t easy to make people gain weight after all.

      While this news may sound strange, think about the last time you ate at the all-you-can-eat buffet. You were stuffed to the gills. Now can you imagine downing another two pork chops? Yeah, not so easy. Furthermore, have you ever tried to feed a baby who is absolutely refusing to eat? They scream bloody murder. It is just about impossible to make them overeat. Convincing people to overeat is not the simple task it first seems.

      Dr. Sims changed course. Perhaps the difficulty here was that the students were increasing their exercise and therefore burning off the weight, which might explain their failure to gain weight. So the next step was to overfeed, but limit physical activity so that it remained constant. For this experiment, he recruited convicts at the Vermont State Prison. Attendants were present at every meal to verify that the calories—4000 per day—were eaten. Physical activity was strictly controlled.

      A funny thing happened. The prisoners’ weight initially rose, but then stabilized. Though at first they’d been happy to increase their caloric intake,³ as their weight started to increase, they found it more and more difficult to overeat, and some dropped out of the study.

      But some prisoners were persuaded to eat upwards of 10,000 calories per day! Over the next four to six months, the remaining prisoners did eventually gain 20 percent to 25 percent of their original body weight—actually much less than caloric theory predicted. Weight gain varied greatly person to person. Something was contributing to the vast differences in weight gained, but it was not caloric intake or exercise.

      The key was metabolism. Total energy expenditure in the subjects increased by 50 percent. Starting from an average of 1800 calories per day, total energy expenditure increased to 2700 calories per day. Their bodies tried to burn off the excess calories in order to return to their original weight. Total energy expenditure, comprising mostly basal metabolic rate, is not constant, but varies considerably in response to caloric intake. After the experiment ended, body weight quickly and effortlessly returned to normal. Most of the participants did not retain any of the weight they gained. Overeating did not, in fact, lead to lasting weight gain. In the same way, undereating does not lead to lasting weight loss.

      In another study, Dr. Sims compared two groups of patients. He overfed a group of thin patients until they became obese. The second group was made up of very obese patients who dieted until they were only obese—but the same weight as the first group.⁴ This resulted in two groups of patients who were equally heavy, but one group had originally been thin and one group originally very obese. What was the difference in total energy expenditure between the two groups? Those originally very obese subjects were burning only half as many calories as the originally thin subjects. The bodies of the originally very obese subjects were trying to return to their original higher weights by reducing metabolism. In contrast, the bodies of the originally thin subjects were trying to return to their original lower weights by increasing metabolism.

      Let’s return to our power plant analogy. Suppose that we receive 2000 tons of coal daily and burn 2000 tons. Now all of a sudden, we start receiving 4000 tons daily. What should we do? Say we continue to burn 2000 tons daily. The coal will pile up until all available room is used. Our boss yells, “Why are you storing your dirty coal in my office? Your ass is FIRED!” Instead, though, we’d do the smart thing: increase coal burning to 4000 tons daily. More power is generated and no coal piles up. The boss says, “You’re doing a great job. We just broke the record for power generation. Raises all around.”

      Our body also responds in a similarly smart manner. Increased caloric