Thursday, September 23, 1999

Botanists Turn to Data Base to Resolve Issues of Plant Taxonomy By VINCENT KIERNAN Cambridge, Mass. As far as Shakespeare was concerned, "a rose by any other name would smell as sweet." To botanists, however, making sure that a plant is correctly named is a matter of utmost concern: "Researchers could spend less time in the library and more time with the organisms themselves," says an organizer of the data-base project. The name indicates the plant's relationship to other species and varieties of plants. Now three groups of botanists are collaborating to compile a free data base that would be an authoritative record, on the Internet, of the names of every flowering plant on earth. The effort is one of several in which scientists hope to use information technology to cut through the confusion that surrounds plant and animal taxonomy. Eventually, many of the projects -- including the botanical data base -- could be merged into a single, electronic reference source for all biological sciences, a goal that was endorsed by senior science officials of the United States and 28 other industrialized countries in June. An initial version of the botanical data base, called the International Plant Names Index, is scheduled to begin tests in November, with modifications and refinements expected to be made for at least three years. In concept, the index is simple -- for every known flowering plant, the data base will provide the plant's official scientific name as well as a citation of the journal article or book in which the discovery of the species was recorded. But having that information on line is of enormous interest to botanists. It could speed the pace of scientific discovery and give scientists new ways to address hot public-policy issues such as biodiversity, says Donald H. Pfister, one of the organizers of the project. He is curator of non-vascular-plant collections at the Harvard University Herbaria and a professor of systematic botany at Harvard. The other participants are the Royal Botanic Gardens, in Kew, England, and the Australian National Botanic Gardens, in Canberra. "Researchers could spend less time in the library and more time with the organisms themselves," says Mr. Pfister. "This really will exploit the advantages of the Net in a new way," says Peter H. Raven, director of the Missouri Botanical Garden, in St. Louis. "It's a big advance." Researchers currently can spend hours in the library trying to locate previously published information on plants. For example, a botanist who comes across an unusual plant might need to find out whether it has been described before, or where it has been found in the past. But searching the botanical literature can involve techniques so arcane that some researchers rely on librarians who specialize in such tasks. There already are some computerized indexes that can assist botanical researchers, but they have limitations. For example, Harvard's Herbaria has a catalogue of information on its World-Wide Web site, but it covers only 325,000 species of flowering plants, all from the Western Hemisphere, and all discovered after 1885. Kew has produced a series of printed indexes that contain information on more than a million flowering plants from around the world that have been discovered since 1753, but those indexes are available in a searchable, computerized format only as an expensive CD-ROM. The new, free index will combine the Harvard and Kew data bases with an index of information about Australian plants that has been published since 1973 by the Australian botanic gardens. Workers from the three institutions will verify the accuracy of bibliographic information in the new data base by checking the actual books and articles that are listed as having originally described the 1.2 million names of flowering plants. The researchers also will reconcile discrepancies in the information from the three institutions and devise standardized formats for the names of plants and authors, so that researchers will not have to try out variations on a name to be sure that they have searched the new data base exhaustively. The index's being free will make it useful to scholars in developing nations -- who might not be able to afford a subscription -- as well as to those in industrialized nations, Mr. Pfister says. The prospect of a single, authoritative source of information is an appealing one, says Richard K. Rabeler, a collections manager at the University of Michigan Herbarium, in Ann Arbor. He says he often has to locate information about the discovery of plants that are stored in the herbarium. "It should be a very useful tool. I'm looking forward to it." Moreover, the data base of plant names could serve as the foundation for even more-comprehensive data bases on the characteristics and geographic distribution of plants, says Mr. Raven, of the Missouri Botanical Garden. By analyzing such data bases, scholars might be able to identify which endangered species of plants are most in need of conservation efforts, or unearth clues about plant species yet to be discovered, he says. With a $616,000 grant from the National Science Foundation, Mr. Pfister and his colleagues have already produced a test version of the data base. It combines information about thousands of authors of articles about flowering plants from the Kew and Harvard data bases, but not bibliographic information on the plants themselves. The final version of the data base will be maintained jointly by the three institutions, with each monitoring various scientific journals for reports of new discoveries of plants, and then adding that information. Users will be able to suggest changes and revisions through a "reply" function that will dispatch the suggestions to editors. Such corrections are more common that one might think. Michigan's Mr. Rabeler, for example, has identified multiple errors in Harvard's data base. Indeed, David E. Boufford, the Harvard Herbaria's assistant director for collections, identifies Mr. Rabeler as among the most prolific suggesters of corrections. In one case, Mr. Rabeler says, he pointed out that three species of peanuts found in the Michigan collection were not listed in the Harvard data base; Harvard's subsequent investigation led to the addition of information on 60 species of peanuts. All proposed revisions -- even those that editors reject -- will be accessible to users of the data base. "We hope to prevent people from making the same mistakes over and over again," says Noel Cross, a systems administrator at the Harvard Herbaria who is working on the project. Other data bases of plant names also could be combined with the new index. The Missouri Botanical Garden, for example, has its own data base of information on 840,000 plants, including names and locations of about 250,000 specimens, says Mr. Raven. The garden plans to link its records to those in the new index, to make them easy to use together, he says. "They will become a living index to all the work people might want to do about plants. By allowing researchers to easily add and correct information, he adds, the system "perpetually will become more useful and richer." The International Plant Names Index and the Missouri data base themselves could become part of the proposed Global Biodiversity Information Facility, a worldwide biological data base. Its creation was endorsed in June by science officials of the nations that belong to the Organization for Economic Cooperation and Development. The global facility would consist of a set of international data bases, created according to one set of rules, and comprising a list of the names of all living organisms. "Although approximately 1.8 million species of plants, animals, and micro-organisms have been given scientific names, an even larger number of names exists in the scientific literature, as some species have received different names at different times or in different locations," says the proposal endorsed by the science officials. "There currently is no repository that contains the accepted names and resolves incompatibilities, so that there is much uncertainty about exactly which organism may be referred to in a scientific report or even a national law." The International Plant Names Index could eventually be expanded beyond bibliographic information to include images, such as digital photographs of each plant, says the Harvard Herbaria's Mr. Boufford. That would be a boon to researchers who are not certain whether a plant they have found is a new discovery. Currently, a researcher who wants to determine if a plant specimen belongs to a known species faces a big task in resolving the issue. The researcher must compare the newly found plant with a "type specimen," a preserved sample of the older plant, collected by its discoverer. But that sample could be in a herbarium halfway across the globe. The researcher either has to travel to the herbarium or persuade officials there to send a sample by mail. Once the Web site is available, however, a researcher might be able to use digitized photos of the type specimen to decide whether it matches the unidentified specimen, without having to examine a physical sample. "Having the images could save a lot of time and effort," says Mr. Boufford, and make for "less wear and tear on the specimens." Easy-to-print version E-mail this article