It began with some marshmallows (棉花软糖). In the 1960s Walter Mischel, a psychologist then working at Stanford University, started a series of experiments on young children. A child was left alone for 15 minutes with a marshmallow or similar treat, with the promise that, if it remained uneaten at the end of this period, a second would be added. Some of the children, who were aged four or five at the time, yielded to temptation before time was up. Others resisted, and held out for the reward. 
    Dr. Mischel then followed the children's progress as they grew up. Those who had resisted, he found, did better at school than those who had given in. As adults they got better jobs, were less likely to use drugs and got into trouble with the law less frequently.  Moreover, children's family circumstances suggested thatimpulsive behavior was as much learned as inherited. This suggested that it could be unlearned-improving the child in question's ch" nces in life. 
    Study after study has confirmed Dr. Mischel's insight. However, recent observations, however, raise the possibility that developing self-control is not always an unalloyed (纯粹的) good. 
    Work published two years ago by Gene Brody of the University of Georgia, who looked at a group of young black Americans, showed that those who exhibited serf-control as teenagers did indeed get the expected benefits. But if such serf-controllers came from deprived backgrounds, they developed higher blood pressure, were more likely to be obese and had higher levels of stress hormones than their less-serf-controlled peers. 
That correlation did not apply to people who started farther up the social ladder. 
    Dr. Brody and his colleagues have followed this study with one that comes to an equally astonishing conclusion: for people born at the bottom of the social heap, seif-control speeds up the process of ageing. 
Dr. Brody and his colleagues followed almost 300 black American teenagers of different backgrounds as they aged from 17 to 22. For the first few years the researchers assessed their volunteers' levels of self-control, and also looked for signs of depression, aggression and drug use. They assessed, too, those volunteers' socioeconomic backgrounds. But the last examination, when participants were 22 years old, was different. 
    Then, the researchers took a blood sample, recorded the DNA-methylation ( DNA 甲基化) patterns of cells in it, and worked out how much these deviated from the pattern expected at that particular age. As the study shows, for people from high-status backgrounds, higher serf-control meant lower cellular ages. For those whose background was low-status, the reverse was true. 
    Dr. Brody's findings are both intriguing and worrying. The research into gene methylation suggests changing methylation patterns are a common response to changing circumstances as well as changing age, as the body's physiology strnggies to keep up. Fortunately, people can change their circumstances in rational ways: the lesson of the marshmallows shows that. If Dr. Brody's result is confirmed, the challenge it poses will be to work out how to avoid the adverse effects of serf-control.