THE editor of Nature - the world's premier scientific journal - highlights a hallmark of modern science. As he notes in a recent editorial, "publications more than 25 years old are likely to be forgotten, which is a shameful waste."Classics of scientific literature, such as Einstein's relativity papers, are safe. But lesser works may well be overlooked as today's working scientists confront yesterday's literature. Nature itself provides two examples in the issue that carries this editorial. One involves the ancient question of why the night sky is dark. The other concerns an arcane aspect of fluid dynamics. Why should the night sky be dark if the universe contains what, for practical purposes, is an infinitude of stars? The sky should blaze with light wherever we look. A traditional answer holds that the universe is expanding so fast that light from the distant stars is degraded and thinly spread. Last February, astrophysicist Paul S. Wesson from the University of California at Berkeley and the University of Waterloo at Toronto in Canada suggested that the darkness is better explained by the simple fact that the universe is of finite age. Galaxies have not had time to flood the sky with starlight. His paper in the Astrophysical Journal intrigued Nature's editor, John M addox, who commented on it. This has drawn a mild complaint from physicist Edward Harrison of the University of Massachusetts at Amherst. Dr. Harrison notes that neither Dr. Wesson nor Dr. Maddox referred to his similar conclusion published in Nature in 1964. A search of my own files shows that Harrison also discussed the age question in 1984 in reviewing the dark-sky riddle in the journal Science. So it isn't just a matter of Harrison's work being too old to be remembered. It apparently also was lost in the flood of recent scientific publications that can overwhelm the most conscientious literature searches. "In the frenetic activity that seems to characterize modern science, the same discovery may well have more than one parent and, sometimes, more than one life," observes mathematician Renzo L. Ricca of the University of Cambridge in England. In his Nature paper, he traces how a set of hydrodynamic equations was discovered at least three different times in this century, beginning with a 1906 paper by Luigi Sante Da Rios at the University of Padua in Italy. Other workers rediscovered the equations in 1965, and again in 1977. Full recognition of Da Rios's priority seems to have come only within the past decade. As more scientific publication finds its way into computer-based data banks, the ability of scientists to search the literature mountain should be greatly enhanced. But the flood of current papers will strain even these sophisticated archives. They are unlikely to hold much material from earlier decades, let alone earlier centuries. The American Association for the Advancement of Science has begun the trend with its announcement last month of the first all-electronic peer-reviewed scientific journal. Literature-burdened researchers will likely welcome the the speed and easy access of electronic publication. They should remember that the hot "new" idea they are eager to broadcast may already lie in a musty old paper journal.