(FORTUNE Magazine) – No other corporate R&D operation matches Bell Labs in size and scope, but some do as impressive a job in a number of areas. IBM's research division -- often % referred to by connoisseurs as ''the other lab'' -- has been coming along fast and boasts four Nobel Prize winners on its staff, vs. Bell Labs' seven. GE is third, with two Nobelists. Another big corporate research player, Du Pont, got its first Nobel in December. In basic research the best labs all offer their people considerable freedom and diversity in scientific activity. All compete for the same young scientists, but treat them somewhat differently once hired. While Bell Labs offers minimum supervision and encourages individual or two- or three-man efforts, IBM researchers at the central lab in Yorktown Heights, New York, typically work in larger teams on specific projects and are more closely supervised. NEVERTHELESS, IBM must be doing something right, especially at its lab in Zurich. Two researchers there shared a Nobel Prize in 1986 for inventing the scanning tunneling microscope, and two others won a 1987 Nobel for discovering that superconductivity can occur at temperatures much higher than earlier recorded -- thus setting off the current boom in that field. Of the 3,000-employee IBM worldwide research team, close to a third do basic research. As Ralph E. Gomory, senior vice president for science and technology, explains it, basic research at IBM consists of two parts. One connects closely to product development; the other ''faces into the world of science, the university world, the world of the American Physical Society.'' Thus, an IBM astrophysicist can be doing highly theoretical work without visible connection to the business, apart from his use of computers. Besides the work of the recent Nobelists, IBM's research contributions include the computer languages Fortran and APL; relational database software, which allows fast access to related items; and reduced instruction set chips (RISC), which permit faster computation. IBM's current work concentrates on computation, chips, and materials. In all, about 10% of corporate sales go to R&D; of the $5.2 billion companywide R&D budget last year, some 10% went for research. GE, which maintains a big 2,000-employee R&D lab in Schenectady, New York, is proud of introducing advanced computerized tomography (CT) scanners and nuclear magnetic resonance (NMR) imagers. Both resulted from close interaction between the Schenectady lab and GE's medical systems division. That interaction didn't just happen. Back in 1974 Walter L. Robb, now senior vice president for corporate R&D, took command of the company's medical systems group in Milwaukee. Primitive CT scanners were just appearing on the market, threatening GE's conventional X-ray machines. Robb recalls asking the manager of engineering if he had conferred with the R&D center. The manager replied: ''Hell, no -- we never talk to them.'' Robb suggested they go back to Schenectady to see what ideas for more advanced scanners were available. The ensuing collaboration eventually helped create a $400-million-a-year business for GE. The Schenectady center sometimes practices what Robb's predecessor, Roland W. Schmitt, calls ''chuck wagon technology -- where you cook it up in your lab, dish it out, and yell, 'Come and get it!' That just doesn't work.'' Schmitt married technological ideas to market needs by bringing marketing specialists into the development organization. Among the results: important new products like an adjustable-speed motor controlled by semiconductor chips and a programmable controller for factory automation. In 1986 GE spent $2.9 billion on companywide R&D. Managers won't reveal the budget of the Schenectady center but say basic research accounts for 5% of it; 45% goes for applied research and 50% for development. GE's past breakthroughs include the first commercially feasible process for making synthetic industrial diamonds, development of high-performance plastics, and the discovery of the superconductive tunneling phenomenon, which brought a GE scientist a Nobel Prize. Key research thrusts today include work on materials such as plastics, metals, alloys, and composites, along with artificial intelligence and medical diagnostics, advanced electronics, and design of next-generation jet engines. Change is afoot at Du Pont's campuslike Experimental Station on the outskirts of Wilmington, Delaware, which houses 2,300 researchers. The emphasis is shifting from bulk polymers and fibers, the revenue-producing mainstays of recent decades, to electronics, life sciences, agriculture, and materials for such new technologies as superconductivity. Says Richard K. Quisenberry, director of research: ''Our strengths are in chemistry and materials, and they are growing in biology. Our objective in the next few years is to strengthen our position in physics.'' LAST YEAR Du Pont spent $1.2 billion on R&D; 7% went for basic research. Total R&D spending amounted to 4.4% of sales. Some recent research accomplishments: development of a new, economical way to make hydrogen peroxide directly from hydrogen and oxygen, and a DNA sequencing machine that puts fluorescent tags on the different components of basic genetic material for easier identification in diagnostic and other work.- G.B.