n the course of three expeditions to the remote South Pacific island where the real-life Robinson Crusoe was washed ashore, Gregory Anderson has roamed steep-sided, eroded volcanic rock ridges in search of a rare botanical jewel known as Lactoris fernandeziana.

Once thought to be nearly extinct, says the conservation biologist, this small herbaceous shrub - which grows only on the slopes of the island named after literature's most renowned castaway - may hold critical biological data for scientists racing to salvage ecosystems around the world threatened with extinction.

With more than 300,000 species of plants on Earth in jeopardy - including an estimated 50,000 that will be extinct in 25 years - conservation research has focused on preservation and learning how to contain forces causing the mass extinction of the world's plant biodiversity.

Now, renewed attention is being paid to islands as storehouses of biological information.

"On most remote volcanic islands, isolation and time combine to create a vast diversity of unique plant species with small population sizes that are genetically homogeneous," explains Anderson, a professor of ecology and evolutionary biology.

"These are natural laboratories to study the ways organisms deal with rarity, fragmenta- tion, small populations and isolation - and perhaps they can provide insight into how continental species now faced with similar kinds of pressures might respond."

Anderson has completed research on the reproductive and breeding system of the L. fernandeziana and presented his research last month at the International Botanical Congress in St. Louis, Mo.

Habitat fragmentation - the piecemeal rending and splitting of habitats that plants and animals occupy - is now the primary threat to the world's current conservation and preservation efforts. Even carefully protected nature preserves, isolated from each other by agriculture, urban sprawl and other human-dominated landscapes, may not be sufficient to support certain species.

In light of this threat, conservationists want to determine the minimum critical size needed to preserve biological diversity and the associated reproductive system of the plants, and services required of insect or bird pollinators.

Or, as Anderson puts it, "How few plants can there be and still expect reproduction, or how low can a species population go until reproduction stops?"

Crusoe Island is part of the three-island archipelago discovered by Spanish conquistador Juan Fernandez in 1547, some 650 kilometers directly west of mainland Chile.

Between 1704 and 1709, a Scottish sailor named Alexander Selkirk was marooned on one of these then uninhabited islands. His story inspired Daniel Defoe to write Robinson Crusoe, which he set in the warm Caribbean rather than the cool South Pacific.

Close to 65 percent of the 158 extant flowering plants on Crusoe Island are endemic, or found nowhere else, says Anderson who visited the island in 1991, 1996 and 1997.

"It offers a unique opportunity to study the narrow genetic base of the islands species, which is determined by the reproductive system, which is essential to successful conservation efforts."

David Bauman