AMSTERDAM - The problem of overweight children won't be solved by piling on exercise in school, according to new research presented Thursday.

Youngsters who have a lot of physical activity during school hours tend to wilt when they get home. Children who have less action in school are more active after the final bell rings, says the study.

In the end, a child will expend the same amount of energy, whether in school or out, suggesting that his level of activity is set by some kind of internal meter in the brain, said the lead researcher, Terry Wilkin, professor of endocrinology and metabolism at Peninsula Medical School in the British city of Plymouth.

During humanity's evolution, "energy expenditure is a very precious commodity," Wilkin told The Associated Press. Energy output would be automatically controlled to maintain body mass, "because that would be a major pressure on the survival of the species," he said.

The research was conducted on 206 children aged 8-10 over two years beginning in 2003. It was presented to the European Congress on Obesity in the Dutch capital.

It is the latest round in a scientific battle between biologists and social scientists on what determines obesity, and the report was by no means going uncontested.

Tim Lobstein, director of childhood obesity program for the International Association for the Study of Obesity, said Wilkin's study was "an interesting phenomenon" worthy of further research, but does not tell the whole story. Nor does it cover a wide enough range of subjects to be conclusive, he said.

Lobstein said it was important to encourage physical activity at school so that children develop an affinity for it later on - "to set the thermostat, as it were, the activity-stat, at a high level."

He said he was concerned the report "will influence policymakers who are looking at squeezing the school curriculum."

Wilkin's team studied children in three Plymouth schools that had widely different activity levels, from a preparatory school for wealthy students that included more than nine hours a week of physical education, to an inner-city school with less than two hours of training per week. Prep school students had 40 per cent more activity during school hours, he said.

The subjects were monitored by an acclerometer, which he said recorded their movements 600 times a minute. Each child wore the device all day for one week during four consecutive school terms. In addition, blood samples and measurements of their weight and body mass were repeatedly taken. The data was adjusted for variable weather conditions and for socio-economic factors, he said.

The study was funded by Bright Future Trust, Child Growth Foundation, Nestle, Kirby Lang Foundation, London Law Trust, and the Early-Bird Diabetes Trust.

Alissa Fremeaux, a biostatistician at the Peninsula Medical School, who analyzed the data, said in a paper prepared for the conference that the children showed a range of activity at school, "but the range and its average were the same regardless of what school they went to."

She added: "We discovered that the children who go a lot of PE (physical education) time at school were compensating by doing less at home, while those who got very little PE time compensated by cranking up their activity at home, so that over the week they all accumulated the same amount."

The conclusion, Wilkin said in an interview Wednesday from Plymouth, was that "intervention with children ... doesn't work." Children who are forced to do more activity will slip back to their old habits quickly, he said.

"That is absolutely characteristic of a biological control system," he said.

Lobstein, who is a trained psychologist, called the study "a snapshot" of children of that age, all sharing a British cultural background even if they came from different social classes. Different results might be obtained in different countries or at different times, he argued.

"I wouldn't say there is a genetic or natural construction here. I think it's very socially constructed," he said at the Amsterdam conference centre.