Although President Bush has yet to submit a detailed budget, the document that he released Feb. 28 shows that he plans to cut the research and development investments of some agencies and freeze the budgets of others. This has set off alarm bells in the research community.

Although the administration is committed to a hefty $2.8 billion increase in the National Institutes of Health budget, support for the physical sciences and engineering is likely to be flat or down. The National Science Foundation, which has the responsibility of supporting research and education in all science and engineering disciplines, will receive a 1 percent, $50 million increase, not sufficient to keep up with inflation.

If enacted by Congress, this approach would exacerbate a troublesome trend in federal R&D policy -- the growing imbalance in support between biomedical research and the physical sciences and engineering. Congress has been very generous with the NIH budget, which has increased from roughly $10 billion in FY93 to more than $20 billion in FY2001. Unfortunately, support for many other scientific fields has been stagnant.

For a variety of reasons, allowing this imbalance to continue would be a serious policy mistake. First of all, the research enterprise is becoming increasingly interdependent. Medical breakthroughs depend on advances in the physical sciences and engineering. Physics led to medical imaging technology such as MRI and CAT scans, computer science is reducing the time needed to develop life-saving drugs through sophisticated simulations, and nanotechnology could lead to much earlier detection of cancerous tumors.

Second, federal support for research plays an important role in the development of new ideas and innovative technologies, the engine of our knowledge-based economy. Many of the technologies that are driving today's economy have their origins in federally sponsored research.

Although companies invest billions of dollars to develop and market new technologies, they find it very difficult to justify to their shareholders making investments in long-term, risky research that may or may not have any direct payoff for the company. The payoff to our economy as a whole, however, is enormous, given the contribution that new technologies have made to increased productivity, faster economic growth and the creation of high-wage jobs.

Third, federal R&D supports the education and training of the next generation of scientists, engineers and entrepreneurs. The university professors who receive federal grants often use it to provide stipends for graduate students, who are integrally involved in performing the research.

Many companies report that a shortage of workers with technical skills is their No. 1 constraint on growth, and they have been lobbying Congress to increase the number of skilled immigrants under the H1-B temporary visa program. Increasing university-based research would help expand the pool of Americans who can compete for these high-tech jobs.

It's no accident that many of America's booming high-tech clusters, such as Silicon Valley, Austin, Boston's Route 128, and North Carolina's Research Triangle Park, have grown up around world-class research universities. University professors and recent university graduates are often involved in launching high-tech start-ups, and companies locate near universities to take advantage of the skilled workforce that universities help create.

Silicon Valley is ``Exhibit A'' of this phenomena. A number of Silicon Valley's most successful companies have their roots in government-sponsored research at world-class universities such as Stanford and University of California-Berkeley.

While Europe and Japan were promoting a computer networking standard drafted by a U.N. committee, the Defense Advanced Research Projects Agency (DARPA), NASA, and the NSF were nurturing the Internet. This allowed U.S. companies like Cisco Systems (founded by Stanford's Leonard Bosack and Sandra Lerner) to establish a commanding presence in domestic and international markets.

The work of DARPA-sponsored researchers such as Jim Clark, Bill Joy, Forest Baskett, Andy Bechtolsheim, John Hennessy and David Patterson led to the creation of Sun Microsystems and Silicon Graphics. The first graphical Web browser was developed by Marc Andreessen and Eric Bina at the NSF-funded National Center for Supercomputing Applications, who together launched Netscape. More recently, Internet companies such as Google and Inktomi were founded by NSF and DARPA-sponsored researchers at Stanford and Berkeley such as Larry Page, Sergey Brin and Eric Brewer.

Study the origins of many of today's information and communications technologies (design tools for computer chips, inexpensive data storage, relational databases, optical networks), and you'll find that government-supported university research played an important role in their creation.

It's critical that the Congress provide funding for a balanced research portfolio. Increasing the NIH budget is important, but it should not be the sole objective of U.S. science and technology policy. Expanded investment in the physical sciences and engineering, such as physics, computer science, mathematics, electrical engineering, materials science and nanotechnology, is essential to maintaining America's economic and technological leadership in the 21st century.

If Congress and the administration decide to act to correct this imbalance, the National Science Foundation and the Defense Advanced Research Projects Agency would be good places to start.

The NSF budget is woefully underfunded. The average grant size at the National Science Foundation is $70,000 a year, which is not enough to support the teams of researchers that are often required to make progress on the toughest scientific and technological challenges.

Currently, NSF is only able to fund a fraction of the meritorious proposals it receives. In response to a $90 million solicitation for long-term information technology research, for example, NSF received $3 billion in proposals.

Over the years, DARPA-funded research has had a tremendous payoff for America's military and technological edge. DARPA needs additional funding so that it can continue to invest in high-risk research while assuming new responsibilities for developing defenses against cyber-terrorism and biological warfare.

President Bush plans to increase defense R&D by $2.6 billion next year, and by $20 billion over the next five years. Although much of this will go toward missile defense, Congress and the administration should devote some of it to increase DARPA's budget and support for long-term defense research more generally.

This is an issue that can and should attract bipartisan support. Silicon Valley executives should devote time and energy to educating Washington policy-makers about the importance of federal funding for long-term research. If we don't have the national will to increase R&D budgets now, when we are enjoying budget surpluses and the economic payoff from far-sighted federal investments in the 1960s and 1970s -- when will we?