On the whole they believe the system still works and as one put

On the whole they believe the system still works, and, as one put it, they would "home in" on any journal that they felt was not maintaining standards. But they know that the strain on their editors and the academics who referee papers is increasing too. Quality, and the detection of fraudulent or sub-standard work, depends upon this network of unpaid peer reviewers, who are themselves teachers and researchers seeking publication and grants and complaining of ever-increasing stress.Publishers are aware that the integrity of the whole system depends on rigorous peer review. So the competition for space increases.Publishers are trying to meet demand for space by increasing the numbers of pages in their journals. That speeds up delivery to referees, although it does not shorten the time needed to read and assess the work. Where they can't save money at their end, publishers are having to explore new deals with the universities to try to keep prices down and journals available.But while the number of papers published per British academic went up between 1980 and 1990 from 1.5 to 6 per university teacher, the number of new journals launched to accommodate them has been declining steadily since 1970. Authors are now expected to submit their papers on disc, and referees increasingly receive them over the Internet.

At the same time cash-strapped university libraries are 'deselecting' journals and cancelling subscriptions."To some extent publishers have been able to keep journals going by making use of new technology. "The new universities without a long history of research are under particular pressure to perform. Bob Campbell of Blackwell, one of the major players in Britain's pounds 50m journal business, is at the sharp end. "There is a huge amount of 'me too' research being done by people who want to keep their jobs," he says. The RAE itself is looking for quality rather than quantity of published material, but several universities are already threatening staff who do not appear to be performing well enough to meet their employers' grade requirements.Unfortunately for the academics who desperately need to be published, the market is against them. In the UK it has been exacerbated by the Research Assessment Exercise, which is grading departments on a one-to- five scale.

But almost any academic knows of instances of plagiarism, piracy - that is stealing someone else's work - or cooking of results to make them look better." Increased pressure of all kinds in academic life, Dr Lock thinks, makes the risk of suspect work slipping through the checks and balances of the system all the greater.And the pressure is growing. The Campaign for the Advancement of State Education (Case) campaign for a legal maximum on primary class size - "No More Thirtysomethings in Primary Schools" - began more than five years ago. Although over the succeeding years we have been able to prove our case time after time, class sizes have continued to rise. Now (and Case can claim some credit) we have a government pledged to do something about primary class sizes. Until we see how much is needed to reduce class size in each primary school, the cost can be only roughly estimated Not all schools will need new classrooms; some will.

But I believe that merely shifting resources will not be enough. In the years of campaigning over class size I have spoken to many parents, concerned about class sizes in their primary schools, who have been told by their local education authorities that there is enough money in the budget to reduce class size, and by their governors that there isn't. If we are to avoid this claim and counter-claim we need legislation to fix a statutory maximum class size If there is a legal limit, the money will have to be found. A recent survey by the Audit Commission showed that the majority of OECD countries have either a nationally recommended minimum space allocation, or a maximum class size for primary pupils.Legislation based on the Scottish model of agreed maximum class sizes in the teachers' pay and conditions of service could be introduced. The nature of the agreement there gets over the problem of one maximum number of, say, 30, which could result in child No 31 not getting a school place The agreement has a normal maximum and an upper limit. For Scottish primary schools the range is a normal maximum of 33, to the upper limit of 39. Child No 34 is thus admitted by agreement between teacher and head. Although the limit is much higher than we want to see, there are fewer children in classes over 30 in Scotland than in England.

The government could stay with a normal maximum for infant classes of 30, up to an upper limit of 35, and set targets to bring these two numbers down, to a normal maximum of 25 with an upper limit of 30, as soon as possible.The existence of the national limit would influence decision making at school, local and national level. As the Audit Commission showed in 1993, two primary schools with similar funding and pupil numbers employed 12.7 teachers and eight teachers respectively. Ofsted recently quoted a range of 55 per cent to 83 per cent of primary school budgets spent on teachers. These are decisions made at school level.At local level, spending formulae are based on an expectation of a particular class size.