LIKE many girls in my year, I approached maths and the science labs with something close to dread. When it came to algebra and trigonometry, I was in that unhappy league who had to forfeit lunch-break for extra tuition, and during class were placed in the front row, under the teacher's eye.

Horribly exposed, and in the firing line, I had as much chance of understanding a word as if I'd been parachuted into Beijing. Maths, like physics and chemistry – such a sickening gassy smell as you walked in the door – were foreign languages. Unlike French and German, they were not ones I wanted to learn.

It was not just girls who were in the dunce category, but there was a marked gulf between most of the bright boys, who seemed to take these subjects in their stride, and most of the clever girls, who approached the classroom with a heavy step. At the time, I never wondered if this apartheid had less to do with innate ability, or lack thereof, and more with cultural pressures. In my case, my incapacity with numbers was all of my own doing. I'd happily have been a maths whizz, defying expectations of what girls should and should not be good at.

Now, Professor Polly Arnold, Crum Brown Chairwoman of Chemistry at the University of Edinburgh, has suggested that, for a while at least, school girls could benefit from being taught science and maths on their own. The number of women making careers in science and technology might be slowly rising, but slow is not good enough. Meanwhile some, after a promising start, begin to fall by the wayside. Even in those areas where women have caught up, notably in medicine, Arnold says that the top posts are still overwhelmingly held by men. She is thus deeply concerned about the "leak" of women out of these areas, whether at school, university or the workplace.

Yet is segregation the answer? For most of us, it sounds like a retrograde step, a retreat from received wisdom. One friend, who went to an all-boys' school, immediately resisted the idea, knowing too well the problems that same-sex education can create. Yet within coeducational schools, sport is still regularly divided into girls' and boys' games and teams. So why not – within strict limits – any other subject? After all, as Arnold discovered, research suggests it works.

Given the poor track record of women in so-called Stem subjects – science, technology, engineering and maths – it has to be worth considering. There is something about all of them that reeks of masculinity. Going into a physics or chemistry lab is like entering a car mechanic's shed. It's not that women cannot do well in such a place, but before they even open their jotters they have first to overcome the sense – and the stark reality – of being in a man's world.

Consequently, at the first hint of difficulty, it is easy for them to give up. As they are well aware, nobody really expects them to do well. Since the advent of universal education, we have not been expected to fly in these subjects. Indeed, it is a truth universally dinned into us that girls are hopeless at maths. Even now it is seen as a little startling if a girl excels in these areas, whereas if she were a whizz at Spanish or art, no eyebrows would be raised. The brilliant early 19th-century mathematician Mary Somerville is an inspirational role model, yet I never once heard her name in class.

An all-girl maths or chemistry lesson would have a very different tone. For a start, it would show the students how seriously they were being taken. It would also encourage a mood in which they could gain confidence, making mistakes or asking "stupid" questions without the fear of boys laughing at them, or – worse – worrying that being better than them will make them less popular or attractive. As well as segregated tuition, these groups could also be encouraged to discuss careers in science and technology, and encouraged to realise they have as much chance of making it in this environment as the boys along the corridor.

The segregationist idea, as I see it, is of propagation: taking the seedlings most likely to wither in a hostile or cold climate and putting them into a greenhouse, under expert care, until they are strong enough to thrive on their own. It should not be the start of increasingly splintered education, but a springboard to get able girls to a level they might otherwise not reach. Also, to help give them the mental resources to prevail in later life. Because unfortunately, the higher up the ladder they climb, the more chance of encountering the testosterone tyranny. By then, one hopes, they will be better placed to resist it.

As the army of excellent women doctors shows, there is no reason why girls can't excel in science. The issue is how to help them see that they can. If that means a return to Malory Towers for a term or a year, what possible harm can that do?