New efforts to decloak 'invisible' science.

In the late 1970s, Iceland's director of geothermal energy programs looked over a couple of just-published reports and lamented that U.S. scientists were continuing to "reinvent the wheel." He then pulled out several documents--one a decade old--that he said described what the U.S. geothermal studies reported as new.

The documents he pulled had all been written in Icelandic--a language spoken by barely a quarter of a million people. Why? Because the Icelandic government required the scientists it funded to publish in the country's native language, the director told Science News.

Though Iceland has abandoned this policy, the anecdote illustrates the role a scientist's native tongue can play in shrouding science--and possibly in retarding research advances. In the August Scientific American, reporter W. Wayt Gibbs outlines the extent to which this and related factors continue to bury the contributions of many scientists.

Gibbs reports that "[a]lthough developing countries encompass 24.1 percent of the world's scientists and 5.3 percent of its research spending, most leading journals publish far smaller proportions of articles . . . from those regions." To illustrate the underrepresentation of scientists from developing countries, the magazine mapped the residences of authors appearing in last year's Science Citation Index (SCI), a commercial service that abstracts some 3,300 journals.

Citing more than 100 interviews with scientists and journal editors, Gibbs examines the roots and fallout of this bias against research from developing countries.

For instance, commercial indexing services ignore the vast majority of the world's journals. In addition, libraries tend to subscribe only to the more popular, frequently cited journals, contributing further to the invisibility of scientists who publish in nonindexed ones.

Though some abstracting services cover non-English journals, the editor of one Mexican medical journal noted that it had to provide English abstracts for its articles, publish on time, and buy a $10,000 subscription to SCI in order to qualify for inclusion. In 1982, hard times hit and the journal could no longer meet those conditions. Since then, it has struggled unsuccessfully to get back into SCI--despite the fact that it now publishes solely in English, has a U.S. editor to avoid translation errors, and has even recruited an editorial board of the top-cited Mexican scientists in the field and an international review board.

Even in the former Soviet Union, non-English-speaking scientists face many of these obstacles, Marjorie M.K. Hlava of Access Innovations in Albuquerque told Science News. But several firms are working to change that. Moreover, their technologies and strategies might later be adapted to build an English-speaking audience for research in less economically developed nations.

Interlock Systems Group in Lanham, Md., for example, is beginning to market English-language compact discs of Russian patent filings, medical abstracts, and other Cyrillic databases. By year's end, its president says, Interlock hopes to offer on-line English abstracts of the holdings in Russia's National Public Library for Science and Technology, a primary repository for research documents in the former Soviet Union.

Hlava's firm is testing an on-line "browser" that translates English queries into Russian, scouts Russian-language journal abstracts, then converts any finds back into English. And the American Association for the Advancement of Science in Washington, D.C., hopes to raise the visibility of major scientists from the former Soviet Union by posting on the Internet the addresses and phone numbers of those who recently won major competitive research awards.