AT THE TURN of the last century, the Smithsonian was largely a scientific research institution, with a few art exhibitions on the side. Nowadays, the "nation's attic" houses everything from Jackson Pollocks and first ladies' dresses to moon rocks and pandas. Along the way, the Smithsonian's original scientific research mission has become diluted, to the point where many people don't realize it still has one. Universities, the National Institutes of Health and others now spend billions of dollars on scientific research, dwarfing the Smithsonian's $111 million research budget. So large is the gap that some now question whether the Smithsonian should use any of its ever-tighter funding on scientific research. Indeed, when he first took office, Smithsonian Secretary Lawrence M. Small ruffled feathers by suggesting that some Smithsonian research divisions should close.

Now an independent commission, set up in the wake of that controversy, has taken a harder look at that question and come up with some answers. Diplomatically, it chose to focus not on what kinds of research the Smithsonian does badly but on what it does well: research that makes use of the Smithsonian's world-class tropical research institute and its unique specimen collections, as well as the astrophysics program the institution developed some years ago with Harvard University. The scientists on the commission thought the Smithsonian could focus more sharply on these areas. While they didn't quite say so, it is clear that the Smithsonian should bravely prepare to cut some of its other programs, on the grounds that they duplicate work that is better funded elsewhere.

Yet it is not just its rare plants that make the Smithsonian special. Because of its reputation, its location and its millions of visitors, the Smithsonian is also uniquely positioned to play a strong role in the delivery of scientific education. This suggests that in addition to focusing more sharply on what they do best, Smithsonian scientists should make sure they are transmitting the fruits of their research to the public. For whatever reasons -- lack of funding, lack of personnel, lack of vision -- some of the Smithsonian's science exhibits, and the Museum of Natural History in particular, are starting to look tired. Yet the museum's diverse collections -- plant, mineral and animal, not to mention live animals at the National Zoo -- contain possibilities so far insufficiently explored: exhibitions on biodiversity, environmental education, links between different aspects of the natural world, programs aimed variously at children, teenagers, scientists, the general public. Given that many policy arguments in Washington revolve around disputed questions of science, the Smithsonian would be a logical forum for discussing them. Smithsonian scientists are not, in the end, quite the same as scientists at MIT or CalTech: They are part of a federal institution, and they could make their presence more strongly felt in public life here, the home of the federal government.