MEN who start smoking as children are likely to have fatter sons, research has suggested.
A large study found that sons whose fathers smoked regularly before the age of 11 grew into chunky teenagers.
Between the ages of 13 and 17, they carried five to 10 kilograms (11 to 22 pounds) more body fat on average than those whose fathers never smoked or took up the habit later.
Scientists said the result from the Children of the 90s study was clear evidence that the effects of lifestyle could cross generations.
They believe toxic chemicals in tobacco smoke may have caused changes to inherited DNA that triggered a metabolic reaction in the boys.
One of the researchers even speculated that the weight gain might be an adaptive response to counteract tobacco effects, since the fathers who started smoking young tended to be thin.
Daughters of early-smoking fathers also gained body fat, but to a much smaller degree that was not considered significant.
Lead scientist Professor Marcus Pembrey, from the University of Bristol, said: "It seems that we haven't got the whole story for why we have the rise in obesity.
"Diet and lack of exercise may not be the whole story from the environmental side of things.
"We have to entertain seriously the fact that there are effects coming from a previous generation."
In total, 9,886 fathers were recruited fro the study, 54% of whom smoked at some time in their lives.
Of these, 166 (3%) were smoking regularly before the age of 11.
Assessments of Body Mass Index (BMI), a measurement of weight in relation to height, were made of the men's sons and daughters at ages seven, nine, 11, 13, 15 and 17.
They showed that the average BMI of boys whose fathers smoked as children rose at each time point until it reached 25.9, just putting them in the "overweight" category.
More significant were measurements of body fat made by whole body low-dose X-ray scans.
They revealed that teenage sons of young-smoking fathers had markedly higher levels of fat mass than their peers.
Between the ages of 13 and 17, they gained an extra five to 10 kilograms on average, or up to 1.57 stone in fat.
Genetic changes caused by the environment that are handed down to future generations are known as epigenetic effects.
They have been demonstrated in animals exposed to toxins and drugs, with their influence emerging over three or four generations.
But evidence of their impact on humans is controversial.
Historical studies in Sweden have linked men's diets in mid-childhood with longevity and deaths from diabetes in their grandchildren.
Prof Pembrey added: "The discovery of trans-generational effects has big implications for research into the current rise in obesity and the evaluation of preventative measures.
"It is no longer acceptable to just study lifestyle factors in one generation."
The research forms part of a large health and lifestyle investigation called the Avon Longitudinal Study of Parents and Children (Alspac), also known as the Children of the 90s study.
Findings from the smoking data were published in the European Journal of Human Genetics.
In their conclusions, the scientists stated: "Smoking by boys in mid-childhood may contribute to obesity in adolescent boys of the next generation."
Why smoking before the age of 11 was so important, and why sons and not daughters were significantly affected, remain unanswered questions.
"It may be that the lead-up to puberty is a particularly sensitive time," Prof Pembrey said.
"We just don't know at this stage."
The different roles played by male and female chromosomes in early development may be involved, the researchers believe.
In addition microRNAS, short lengths of genetic material that can switch off genes, are carried in the sperm.
Which of the 2,000 chemicals in tobacco smoke might be having the epigenetic effect is another mystery.
"We know very well that paternal smoking can lead to breaks in DNA that can be observed in newborn's cord blood," Prof Pembrey said.
Animal studies had suggested trans-generational epigenetic changes could be the result of cells responding and adapting to damage to their DNA, he said.
It was possible that in some cases they could even be beneficial.
Colleague Professor Jean Golding, also from the University of Bristol, said accumulating more body fat might be just such an adaptive response.
"These fathers who started smoking early are thin, and the next generation might feel they need to put on weight to counteract the nasty things that are going to happen to them," she said.
"It is only speculation, but there are mechanisms to do with being primed for harsh conditions that you then try to overcome."
The research was criticised by leading epidemiologist Professor Sir Richard Peto, from Oxford University, who alleged that reliance on what could be chance findings had led to a "false positive result".
"The main claim in this report is unjustified," he said.
"This new study does not provide good evidence that fathers who started to smoke before age 11 have fatter teenage children as a result."
Prof Pembrey pointed out that Sir Richard had overlooked the fact that the study was testing a prior hypothesis.
"We don't buy that criticism at all," he said.
"It seems he hasn't read the paper very carefully."
Professor Wolf Reik, head of the Epigenetics and Chromatin Programme at the Babraham Institute in Cambridgeshire, said: "The paper makes a fascinating observation which further supports the idea that trans-generational epigenetic effects can occur in humans.
"These are also increasingly observed in animal models, showing that while there is a mechanism for 'erasure' of epigenetic information between generations, this can fail on occasion. The inheritance effects reported here (and those in some animal models) occur with transmission from father to son, which means that the information needs to be carried in the sperm."
Tim Spector, professor of genetic epidemiology at King's College London, said: "Trans-generational effects of environment have been clearly shown in rodents but not in humans. This is a rare study showing trans-generational effects in humans.
"The data are persuasive but not yet definitive, as we need to confirm the same smoking-related epigenetic changes in the kids' DNA. We urgently need more studies like this."
Statistician Professor Patrick Wolfe, from University College London, also thought the way the study was conducted increased the chances of a false positive result. He did not think the science was "very compelling".
"Even if the results can be considered as meaningful, this study cannot rule out hidden associations because one cannot control for all confounding factors," he said.