PHILADELPHIA -- The "French paradox" -- the perplexing disconnect between France's rich cuisine and slender population -- can be explained in part by portions that are significantly smaller in French restaurants and supermarkets than in their American counterparts, according to researchers at the University of Pennsylvania and the Centre Nationale de la Recherche Scientifique in Paris.

The researchers compared the size of restaurant meals, single-serve foods and cookbook portions on both sides of the Atlantic.

"The French paradox is only a paradox if one assumes that dietary fat is the major cause of obesity and cardiovascular disease," said Paul Rozin, professor of psychology at Penn and lead author of a paper in the September issue of the journal Psychological Science. "However, recent studies suggest that the importance of fat intake as a risk factor has been greatly exaggerated. While the French eat more fat than Americans, they probably eat slightly fewer calories, which when compounded over years can amount to substantial differences in weight."

The French paradox has long stymied American dieters and scientists, puzzled by the ability of the French to remain trim while downing buttery croissants, creamy brie and decadent pastries. Just 7 percent of French adults are obese, compared with 22 percent of Americans, and the mortality rate from heart disease is significantly lower in France.

Rozin and his colleagues weighed portions at 11 comparable pairs of eateries in Paris and Philadelphia, including fast-food outlets, pizzerias, ice-cream parlors and a variety of ethnic restaurants. They found the mean portion size across all Paris establishments was 277 grams, compared to a mean in Philadelphia of 346 grams -- 25 percent more than in Paris.

The Parisian portions were larger in just one of the 11 comparisons, at Hard Rock Cafes in both cities. Three other international restaurant chains consistently served larger portions in the United States., and Philadelphia's Chinese restaurants served meals that were, on average, 72 percent heftier than those served by Chinese restaurants in Paris.

The researchers also examined references to portion size in Philadelphia and Paris editions of the 2000 Zagat restaurant guide. Serving sizes were not only mentioned roughly three times as frequently in reviews of Philadelphia restaurants, but, of these mentions, fully 88 percent described large portions, compared to just 52 percent in Paris.

"Many studies have shown that, if food is moderately palatable, people tend to consume what is put in front of them and generally consume more when offered more food," Rozin said. "Much discussion of the 'obesity epidemic' in the U.S. has focused on personal willpower. But our study shows that the environment also plays an important role, and that people may be satisfied even if served less than they would normally eat."

Extending their approach to single-serve foods sold in supermarkets, Rozin and colleagues found 14 of 17 items studied were larger in American stores. For example, a candy bar sold in Philadelphia was 41 percent larger than the same product in Paris, a soft drink was 52 percent larger, a hot dog was 63 percent larger and a carton of yogurt was 82 percent larger.

Rozin's co-authors on the Psychological Science paper are Kimberly Kabnick and Erin Pete at Penn, who conducted the work as part of their senior Psychology Honors thesis, and Claude Fischler and Christy Shields at CNRS. Their work was sponsored by the National Institute on Drug Abuse.