Both boys and girls in public school are now more likely to opt for science than their state school counterparts, according to Professor Alan Smithers, from Brunel University.

A regime of narrow A-levels might also be to blame for the collapse of female interest in A-level physics - the only subject in which girls’ performance is declining. The latest figures show that since 1988 the ratio of boys to girls taking the subject has risen from 3.41 to 3. 59.

Speaking yesterday at the Association for Science Education Annual Meeting in Birmingham, Professor Smithers said that “piecemeal” reform of the education system and a dramatic fall in industrial research spending are to blame for the collapse of recruitment to science A-levels and degree courses.

He said that the three A-level regime is also making a return of serious teaching shortages in the sciences more likely.

In a speech sponsored by the Save British Science organisation, Professor Smithers called for a “major re-think” of science education in England and Wales with more emphasis on academic breadth from 14 onwards, and a five A-level system. “We need fundamental decisions about the purpose of science education and whether the emphasis at particular ages should be on generally informing, or on educating the professionals of the future. ”

The present, narrowly-based A-level programme, he said, ensures that highly motivated generalists are unlikely to become scientists or science teachers, even if they have the necessary understanding.

Professor Smithers called for higher salaries for science teachers and the introduction of regular sabbaticals to aid recruitment.

He also attacked the Government for low spending on industrial research and development. The lack of posts for full-time scientists, he said, is another reason for the low number of science graduates.

He said the quality of science and engineering departments has now been put at risk by the rapid expansion of higher education, and suggested cutting some of the 105 different engineering courses.

The proportion of A-level students specialising in science has fallen from 44 per cent in 1962 to less than 17 per cent in 1994. Since 1988, the numbers taking A-level physics have dropped from 45,716 to 32,801, and the total taking chemistry is down by 6 per cent.

At degree level the sciences have stagnated. Total admissions have risen from 112,000 in 1962 to 240, 000, but chemistry, maths and physics have each kept to a plateau of around 2,000-3,000. To make matters worse, engineering and technology courses attract fewer of the best A-level students.

Yet this is despite a major success story at GCSE. In 1980, only 30 per cent of boys and 10 per cent of girls entered O-level physics whereas last year more than 80 per cent of each sex entered GCSE physics or double-award science.

GCSE and A-level now point in different directions, said Professor Smithers. “There is a disjunction at age 16. Although participation may have been increased to GCSE it does not feed though into A-level. This may be because GCSEs and A-level really belong to different systems. GCSE is about science for all, a threshold of scientific literacy; A-levels . . . are a high-level selection device which has enabled us to identify students who could be educated to degree standards.”

The narrow A-level system, he said, could also turn students away from becoming science teachers.

“A-levels tend to pick out particular personalities and the science specialists tend not to be person-oriented. In other words, they are not likely to enjoy working with children. It is perhaps not surprising that it is difficult to attract science teachers, especially when teaching is potentially very draining and we do not have good arrangements for teachers to revitalise themselves in their own subjects. ”

There have been 30 to 40 per cent shortfalls of science teacher recruits at all times except in 19911992.