With the sweet temptations of the season beckoning from every dining room, Americans are engaging in a mass food binge that has become almost an annual tribal ritual. Next, of course, comes the tribal penance -- the diet that heads everyone's list of New Year's resolutions.

But this cycle of on-again, off-again dieting is dangerous and probably even unnecessary, according to Dr. Wayne Callaway, director of the Center for Clinical Nutrition at the George Washington University School of Medicine.

In the first place, it implies that excess weight is bad for everybody, which just isn't true. Some people, because of the distribution of their fat, the stage in life in which they gained weight and their family histories of fat-related diseases, do just fine, medically speaking, even with some extra poundage.

In the second place, the cycle encourages the dangerous habit of yo-yo dieting, which can actually create medical problems and, ironically, can make the weight harder to take off next time around.

Excess fat on the hips and below is harmless, Callaway told participants at last week's seminar on the genetics of obesity, sponsored by the National Institutes of Health's Office of Disease Prevention. "Other than the cultural stigma in our society," he said, "scientists have not been able to identify any metabolic abnormalities in thigh fat." Big hips, thighs and buttocks, he concluded, have no relationship to the medical complications of obesity.

But big bellies, shoulders and bosoms, however, do. The easiest measure of body fat distribution is the waist-to-hip ratio; the higher the ratio, the more the fat is concentrated in the upper body. And this ratio is a useful predictor of just how important it is to get off those extra pounds.

"The risk of developing diabetes," Callaway said, "is 15 times higher in individuals with the highest waist-to-hip ratio than in those with the lowest," even if their overall weight is the same.

This is true for other diseases associated with obesity, too, he said: the risks of coronary artery disease, high blood pressure, hyperlipidemia and endometrial cancer are highest in persons whose excess weight concentrates above the waist.

Because of the importance of fat distribution in predicting the health of an obese individual, Callaway said the height-weight tables now in use can be "misleading."

"The man at highest risk of coronary artery disease is the guy with no buns at all and a big gut hanging over his belt -- he runs a risk roughly 20 times that of the general population," Callaway said. "But he looks at the height-weight table and sees that his weight is normal, and he thinks he has nothing to worry about."

A physician's role, according to Callaway, should be to differentiate those overweight patients whose excess weight is dangerous from those whose excess weight is merely displeasing to them. "In those with a significant amount of abdominal fat, who are sort of apple-shaped, there is good reason to restrict their food intake, get their weight to normal and calculate ways to sustain that weight loss," said Callaway.

"But in those who can be considered the healthy obese, mostly those who are pear-shaped, the most reasonable medical intervention is allowing them to deal with the fact that they are simply bigger than other people."

Most overweight Americans, especially women and especially those with high incomes and high education, have trouble simply accepting a bigger body. According to Callaway, the National Center for Health Statistics has reported that half of adult American women say they go on a diet two or more times a year. Paradoxically, the diets themselves can disrupt their metabolism, and their appetite, so badly that the weight actually becomes progressively harder to lose.

"An under-fed human," said Callaway, "is stimulated to overeat by sugar and alcohol," both available in great abundance during the holiday season. He cited a classic study of college-age women, some dieting and some not, who were asked to sample three flavors of ice cream. The amount of ice cream they tasted depended on how hungry they were -- but not in the way one might expect.

If they were tasting on an empty stomach, both the dieters and the non-dieters took small samples of the ice cream. But if they were first "pre-loaded" by drinking a milk shake, meaning they were rather full when they got to the ice cream, a surprising observation was made. The non-dieters, feeling full, ate less ice cream; the dieters, their appetites whetted by the sweet milk shake, ate more.

And when they were pre-loaded with two milk shakes, the non-dieters ate even less ice cream, but the dieters ate even more.

Because dieting can so distort appetite cues, Callaway proposed restricting diets only to those people who can truly benefit from weight loss. "Physicians can't do anything in a rational way until they know when the weight was put on, its distribution, and the patient's metabolic rate and eating pattern."

Someone who has been overweight since adolescence, and with no family history of heart disease or diabetes, may do just fine "with a weight goal of her weight at age 18," he said. Someone with fat thighs and hips "should deal with it as a cosmetic issue."

And someone with abdominal fat should try to reduce his or her weight, but in a very slow and careful way -- by increasing exercise levels and keeping to a diet that provides a full complement of calories and nutrients.

"What's most important is to match the treatment or intervention with the subtype of obesity, to look at it in functional terms," Callaway said. "How can you compare two people of equal weight when one is immobile and nonfunctional, and the other is playing tackle for the Redskins?" Obviously, he said, the number of pounds alone does not tell the whole story.

Robin Marantz Henig, a Washington free-lance writer, is the author of "How a Woman Ages."