The brilliant thing about science is that when something is disproved once, it's disproved forever. The not-so-brilliant thing about public health policy is that it has little to do with science.

Everyone in the developed world knows what's causing our obesity epidemic. BBC nailed it: "We eat too much, and too much of the wrong things," and Michelle Obama tells us "We have to move more."

Clearly what we need is a balanced diet of lean meats, some good fats, and complex carbohydrates like fruit, vegetables and whole grain bread, and exercise of 30 to 90 minutes per day. Their prescription is completely reasonable and makes intuitive sense.

It is neat, plausible, and wrong. It has in fact been disproved, as nearly as "disproof" can exist in nutrition science.

In his previous book, Good Calories Bad Calories, respected science journalist Gary Taubes exhaustively researched and cited two centuries worth of research in nutrition. He came to the conclusion that none of those recommendations is supported by science, because the fundamental theory on which they're based is wrong.

Why We Get Fat is an updated summary of that earlier work, much quicker and easier to read, with some significant points clarified.

The most important point of the book is that all those public recommendations -- the food pyramid, the "eat food, not too much" approach, everything we know about a balanced lifestyle -- is founded on the premise of Calories In vs. Calories Out.

That we get fat because we eat too many calories, or we don't burn enough of them through movement. But this is nonsense. It's not just wrong, it is actually not a statement about what causes obesity at all (or heart disease, cancer or diabetes, for that matter.)

It is, in Taubes' words, a "junior high level mistake," because it tells us nothing about fat accumulation. If we get fat, by definition we have taken in more calories than we've put out -- but WHY we took in those calories, or didn't burn them, is the key point.

Taubes reviews the scientific literature (rather than the popular press) and presents a conclusion that was common knowledge before WWII, and heresy afterward: we get fat because our fat cells have become disregulated and are taking nutrients that should be available to other tissues.

Like a tumor, the cells live for themselves rather than in balance with the rest of the body. And since those nutrients aren't available, we become hungry and tired. Therefore we eat more, and move less.

For the chronic dieters among us, one passage about animal models will explain decades of frustration. Rodents with a particular part of the hypothalamus destroyed would become obese and/or sedentary *as a consequence* of their bodies putting on more fat. "After the surgery, their fat tissue sucks up calories to make more fat; this leaves insufficient fuel for the rest of the body...The only way to prevent these animals from getting obese is to starve them...they get fat not by overeating but by eating at all." Sound familiar?

The problem isn't one of gluttony and sloth, as Taubes refers to it, but of hormone balance. Simply put, some people are more sensitive to the hormone effects of insulin, cortisol, and a few other -ols, than other people are. The more sensitive you are, the more you're likely to get fat, and the more fat you're likely to get, in the presence of even small amounts of carbohydrate -- and in the absence of enough fat.

That's right, this book advocates eating fat. Not just moderately, but as much fat as possible, up to 78% of calories. Not lean meats, not Jenny-O 99.6% fat-free turkey, not skinless chicken breasts, but lard. Yes, lard. The healthy way of eating, according to Taubes, is moderately high protein and high fat. Yes, high fat.

About a 3:1 ratio of fat to protein, and almost no carbohydrates. (Telling people to eat a balanced diet containing carbohydrates is, he says, equivalent to telling smokers to include a balanced serving of cigarettes.)

And he demonstrates exactly why a high-fat, low-carbohydrate diet is the most heart-healthy approach, as borne out by several dozen recent studies.

While Taubes acknowledges that exercise seems to be good for us for a variety of reasons, weight control isn't one of them. Study after study conducted by proponents of exercise have admitted that they see no compelling evidence for exercise as a weight-loss tool.

And it makes sense if you throw out the calories in/calories out model of why we get fat. If we're fat because our fat tissues are starving the rest of our cells of fuel, exercise is just going to make us hungrier and more tired, not leaner and more fit. (It's worth noting that according to Taubes, in the 1930s obese patients were treated with bed rest.)

[This review was edited to clarify the following point.] The main thrust of Taubes' argument, however, surrounds sugar and to a lesser extent any carbohydrate. Insulin is the primary hormone that fixes fat in the fat cells. This is why Type I diabetics lose weight: they're not producing enough insulin.

Since insulin is manufactured in direct response to carbohydrates, if you don't eat them, you won't have a mechanism by which to store fat. (Taubes notes that this mechanism is not controversial; it simply hasn't had an impact on nutrition policy.)

Taubes argues that any success in standard diets can be attributed directly to the dieter's reduced intake of carbohydrates, especially sugars and particularly fructose.

Once the underlying cause of obesity is understood (hormone balance, not gluttony/sloth) the recommendations on what to do about it are surprisingly simple and therefore brief. This is a book about the science of nutrition, not a diet book, but there is a list of recommended foods in the Appendix.

The book does not tell you how to eat in a restaurant. But it does tell you that the issue isn't in your brain, your willpower, your character, your job, your environment or even (except to the extent that you're sensitive to carbohydrate) in your genes. The problem with fat is in your fat cells.

For a lay audience, this book is as good as it gets if you want to read actual science about health and nutrition. If you're of scientific or technical bent, read Good Calories Bad Calories first, then give Why We Get Fat to your parents.