Sir

In these times of tight funding, it is surprising that some Principal Investigators (PIs) at the US National Institutes of Health (NIH) have seven or more research grants. (In a recent casual search of the NIH database, I found one investigator with 12 grants.) This distribution has not undergone major changes over time (see Nature 363 578; 1993).

In general, the more grants one has, the more each is worth. Rather than the infrastructure of one grant assisting another, the opposite appears true. The science done in big laboratories may be of good quality but, for every additional grant given to a big laboratory, several independent investigators are given no chance.

In 1996, three investigators had seven research grants, each averaging almost $900,000. For the cost of funding these three investigators, almost 80 PIs were unfunded. If NIH were to limit the number of research grants to two per PI, approximately 3,000 new RO1 grants for young investigators could be funded with no change in the NIH budget and without eliminating any currently funded laboratories.

An examination of the Nobel prizes in biology and medicine makes it clear that the work for which the prize was awarded was done while the investigators were not holders of multiple grants. Nobel prizewinners are obviously bright, well funded and have excellent students, but they don't get more Nobel prizes. Can one buy creativity in science? In the competitive field of molecular biology, the average number of NIH grants per cited author (of the top 10) is 2.1. Apparently more NIH funding is associated with less interesting results.

Grant collectors deprive young people of independent research opportunities and also keep them out of academic institutions. If an applicant doesn't have a continuing line of NIH funding, tenure approval is unlikely. Much discussion has been devoted to bringing young investigators, and particularly minorities, into science. But how can a newcomer compete with established scientists with many grants and dozens of students and technicians?

The NIH should examine, by some objective measure, whether channelling vast sums into big laboratories does produce more productivity per dollar than grants to smaller laboratories. If it doesn't, then NIH needs to revise the rules by which grants are awarded. One possibility would be to limit the number of grants per PI. About 1,200 PIs (5 per cent of the total grant recipients) would be affected if awards were limited to two per PI.

If limits were placed on the number of grants per PI, the biggest laboratories would get smaller, and more junior faculty would become independent PIs and generate the key ideas that determine the future path of science. Also, on obtaining a maximal number of grants, PIs would be relieved of the demand (self-inflicted and from administrators) for more proposals.

There are obviously many ways to set limits on grant support and none is immune from abuse and error. But we, the research community, should tell NIH, Congress and the American public that we are willing personally to sacrifice funding opportunities to obtain a more equitable distribution. If we are then rewarded with increased funding by Congress, the increase will not be squandered.