They do what they like in the most literal sense: rational people do not go into cosmology or dendrochronology or functional genomics for the money, they go into it because they are driven by a passion for the subject. They also do what they like in a capricious sense: they may take a university post or a research council grant to work on one thing, and then find themselves beginning to follow the logic of their labours in quite another direction. This, too, is on the whole no bad thing. If 55 years ago, Francis Crick and James Watson had done precisely what they were told, instead of exactly what they liked, the first of a whole procession of Nobel prizes might have gone to Berkeley, California, or Cambridge, Massachusetts but they certainly would not have gone to Cambridge, England.

All of which is the backdrop to a very sorry drama being played out in institutes funded by the Biological Sciences and Biotechnology Research Council. Two agricultural institutes are under sentence, and an estimated 500 jobs could disappear across five research stations, all in the cause of paying for hardware that nobody has thought of, and laboratories that have not yet been built, decades into the future. This is because a government that invaded Iraq without the least idea of what it might cost even six weeks into occupation, has insisted that universities and research councils factor in the full economic costs of science that nobody has yet dreamed of. So the wealth of now - the accumulated expertise of 500 thoughtful, enthusiastic or downright obsessive people - is to be discarded because the doctrine of full economic costs means that any research that survives must be sustainable.

Words such as "baby" and "bathwater" keep springing to mind. Lord May of Oxford, the famously forthright, outgoing president of the Royal Society, may have been thinking of them when he spoke out recently about the "appalling, obsessive bureaucracy" hampering British science. "Today, Crick and Watson's work on DNA would have been blocked before they had got started," he told to Robin McKie in the Observer. "Crick would have been sacked for being idle and Watson would have been told to piss off and stop messing about with his grant."

Scientists for Labour (not a body openly committed to this government's downfall, or even to its embarrassment) has already angrily said that "such swingeing cuts, particularly in the areas of food safety and public health, make a mockery of the government's frequently stated aims of increasing support for science." The Campaign for Science and Engineering (it used to be called Save British Science) is more sad than angry. The determination to put the management of both science and its infrastructure on a sound and sustainable footing is fine, says Peter Cotgreave, its director. Once this golden future is secured, research will be so much the healthier. Under successive Conservative governments, the infrastructure of science slowly collapsed.

"We won't get back into that situation in 10 or 15 years, as we did before. That's a good thing. But clearly, as we move from where we were to where we are, if you put money into mending the roof instead of paying scientists, you lose scientists," he says.

"What has happened is the budget hasn't increased fast enough to avoid pretty significant redundancies. It is not a strategy to get rid of people. These are accidental victims of a thing which in principle is a very good idea and will avoid many of the issues that we were complaining about when this government first came in."

But, he agrees, there seems to be no way of measuring the full economic cost of not doing the research that those 500 scientists might have achieved. The betting is that research in fields that are unfashionable now will be cut, but of course this is precisely what may turn out to be vital five years on. And anyway, if we live in a knowledge economy, this a bizarre way to finance it. If Britain is to become the best place in the world to do science, as the chancellor of the exchequer keeps promising, then saying goodbye to 500 scientific posts purely for accounting tidiness is not a great way to start. If the research councils want more children to take up science at school, then telling 500 scientists that they cost a bit too much may be sending the wrong signal.

And if the science bosses want future glory for British science, they won't get it by playing safe. "If Alexander Fleming had been made redundant five years earlier in a budget cut we'd never have known about penicillin," says Cotgreave.