Judy Mylne woke with a start and glanced at her bedside clock; it was 3am. She went to the window and looked out over the quiet street of terraced houses: in the middle of the road her son James was Rollerblading, up and down, up and down, between the rows of parked cars. He must have woken her as he went out, she thought, leaning her forehead against the cool glass, watching him, feeling sick and afraid.

At first when James had started behaving oddly, being difficult, she’d put it down to normal teenage moodiness, probably exacerbated by his parents’ divorce when he was 16. He’d always been very good at art, won prizes; you had to make allowances for artistic souls didn’t you? But now, two years later, it was a lot more worrying. He would rant at his mother obsessively about such things as the power of purple; friends avoided him, tutors on his art foundation course said they couldn’t teach him. “He’s a mess,” they told Judy.

One evening he dropped his Walkman on the floor, stamped on it and screamed: “I’m going to kill myself and take you with me.” He head-butted the wall, put his fist through a door and, with blood pouring from head and hand, ran out on to the street. The next day Judy took him to their GP, who referred him to a psychiatrist, to whom James admitted that he had been smoking cannabis regularly. By now he was hearing voices and thought people were following him. One night Judy came home from dinner to find James packing a few random objects into a bag inside a nest of twisted coat hangers. He said he was going to walk to Nepal in the morning.

“I thought: ‘My God, he’s really, really ill’,” Judy recalls. She closed the door quietly, fetched two sleeping pills, dissolved them in a glass of Coca-Cola and took it to him. Then she packed a bag and went to a friend’s house. In the morning she rang her GP, the psychiatrist, her older stepsons and her ex-husband. “I’m not going back to the house,” she told them, “you must go and get James and take him somewhere safe.”

James’s father, a barrister, was in court and asked leave to speak to the judge in his chambers. There he explained that his son had been taking drugs and was possibly psychotic. The judge looked at him: “My son has the same problem,” he said. “Go, and take as long as you need.”

There but for fortune, it seems, go any of us with teenage children. Though most people use cannabis without any obvious harm, most of us know of someone — our own child or a friend’s, a friend of a friend’s — who has got into trouble smoking weed, often skunk, which has higher levels of THC, the compound that gets you stoned.

The most extreme cases, such as James, develop a psychosis (schizophrenia or bipolar disorder) from which they may or may not recover. And it is no respecter of class, education or background. Dr Zerrin Atakan, apsychiatrist, sees severe cases at her clinic at London’s Maudsley Hospital: “Sadly many of these young people had been bright, sensitive, happy children,” she says. “Parents often feel dreadfully guilty for allowing them to smoke weed, because in their day, it was relatively harmless.”

Dr Atakan’s patients have usually been smoking from a young age, while the brain is still developing: “We know now that this is a significant risk factor in the development of psychosis. In an ideal world, no one would smoke before the age of 18.”

So, does cannabis cause psychosis? Almost certainly not by itself. Cannabis-related psychosis is a relatively new feature in the landscape of mental illness and there is little reliable data on it. One study found that people who use cannabis before the age of 15 are at least four times more likely to develop schizophrenia, but all of them probably had a predisposition for psychosis in the first place — sometimes, though not always, indicated by a family history of mental health problems.

A new Australian review of current evidence found that 42 per cent of patients with psychosis had used cannabis. Yet, despite much greater use of skunk during the Nineties, there has been no significant increase in the incidence of psychosis in the past 30 years. Why not? David Kavanagh, of the University of Queensland, is one of the authors of the review: “While cannabis may not cause psychosis, there is no doubt that it will trigger psychosis much earlier in vulnerable young people. This is very important because the period of late adolescence is critical for the completion of education and the development of social, emotional and sexual competence and a psychotic episode during this period is extremely disrupting.

“We also know that cannabis use tends to worsen subsequent symptoms and triggers further episodes.” British researchers believe that, because cannabis use by children is a recent phenomenon, the effects have yet to show in the figures and that there will be an increase in schizophrenia in this current decade.

In one study of 2,500 young people, the effect of cannabis use was much stronger in those with a predisposition for psychosis (23.8 per cent) than in those without (5.6 per cent). But even when there is no known family history of mental illness, some children may be genetically more vulnerable than their peers, or have a personality that does not handle cannabis very well, and the Government has ordered a review of the evidence for this. There are genetic tests, but they are expensive and unlikely to be ordered until the damage is done. So how do you tell? “Well, it’s not written on the forehead,” says Dr Atakan, wryly.

Marjorie Wallace, founder of the mental health charity SANE, agrees that you cannot know who is vulnerable: “It’s like watching children playing Russian roulette; one of them is going to be a victim.”