Add one more theory to why criminals behave as they do. You already know about social disadvantage, economic hard times, televised violence or TV's touting of the good life. Now add food. Overrefined, oversugared food. Junk food.

"Food can make an individual aggressive, hyperactive, violent and antisocial," the health magazine Let's LIVE was saying nearly a year ago. "Food can set up the biochemistry for crime."

It isn't such a startling notion, when you think about it. We've known for some time that there is a connection between junk food and hyperactivity in children. For some reason, however, most of us haven't associated diet with behavior in adolescents and adults.

But some researchers have.

Diana Fishbein, professor of criminology at the University of Baltimore, ran an experiment in which a group of inmates at a Florida correctional facility were taken off refined carbohydrates--virtually eliminating white flour and white sugar from their diets. The result: a strikingly low rate of disciplinary problems and violence as compared with a control group. According to Let's LIVE:

"When the experimental group was returned to regular prison chow, the facility officials and the inmates themselves noticed the difference in terms of behavior, outlook, sleep patterns, anxiety states and aggression. The inmates, according to Dr. Fishbein, requested to be returned to the low refined carbohydrate diet again."

At the Tidewater Detention Home in Chesapeake, Va., researchers conducted a double-blind experiment in which a group of juvenile offenders were given a diet that excluded refined sugar. Soft drink machines were taken out and replaced with machines that dispensed unsweetened fruit juice. Table sugar was replaced with honey. Presweetened breakfast cereals were eliminated. High-sugar desserts and snacks were replaced with nuts, fresh fruit, peanut butter and popcorn. Again the result: "The number of continually trouble-prone juveniles dropped 80 percent."

Other experiments have marked canned soda (high in phosphates), processed foods and even excessive amounts of milk as culprits in aggression and other antisocial behavior.

The evidence of these studies so convinced New York Assemblyman Ralph Goldstein that he will shortly be re-introducing legislation designed to help determine whether improved nutrition might be an effective--and cheaper--way of reducing crime. Under his bill, prison inmates would be screened for hypoglycemia (too little blood sugar) and, where indicated, given special diets and up to six meals a day. The proposal has the strong endorsement of Prevention, the health magazine.

It all sounds a little crazy--like the "television intoxication" defense offered on behalf of the Florida kid who murdered an 83-year-old woman a few years back: one more attempt to diminish our personal responsibility for the choices we make.

But suppose the theory proves to be correct. What do you do with a young man convicted of a violent offense if it later turns out that his behavior resulted from junk-food-induced chemical imbalance? Release him on condition that he eat his spinach and give up Sugaroos? And then what do you do with the thousands of others who throw themselves on the mercy of the courts as junk-food junkies? How could you hope to distinguish between the chemically confused and the merely rotten? And assuming it were possible to do so, what would you do about the people who manufacture and sell the crime-inducing junk? Force them to post a warning: "Caution: the Surgeon General has determined that eating this product can lead to criminal behavior"?

Goldstein hasn't opened that can of worms. His notion is much more modest: if it can be shown that improving diets signficantly improves behavior, isn't it worth a shot--in prison or out?