The paleobots say we've got it all wrong, you know. Over the past several decades the scientists who study the history of plant life — known as paleobotanists, or paleobots for short — have been building up a picture of how groups of plants have responded to climate changes. In doing so, they've created a picture of how ecosystems evolve that may change how we think about the multibillion-dollar business of conservation.

Back in 1889 a young explorer and biologist named C. Hart Merriam climbed 12,633-foot Humphreys Peak in the San Francisco Peaks in Arizona. He noted what many mountaineers had noted before: that as he climbed, the plant life changed radically. Starting among the cacti of the Sonoran Desert, he climbed into a pine forest at 7,000 feet and a treeless alpine tundra at the summit. It seemed to him that the plants at a given altitude were associated in what came to be called "communities" — groupings of interacting species. The idea was that over the course of time, plants that require particular climate and soil conditions come to live in the same places, and hence are frequently to be found together. An oak-hickory forest might be one such community, a spruce-fir forest another. What the paleobots are asking is whether these associations, real in the present, are permanent.

A great natural experiment took place on this planet between 25,000 and 10,000 years ago, when small changes in the earth's orbit and axis of rotation caused great sheets of ice to spread from the poles. These glaciers covered much of North America and Europe to depths of up to two miles, and then, as the climate warmed, they retreated. During this retreat, they left behind newly uncovered land for living things to colonize, and as those living things moved in they laid down a record we can read now. Today the study of tree migration is a growth industry; in the past couple of years workshops on the subject have been held in California and Australia.

Think about a lake in New England, a few miles south of the retreating ice. As plants started to grow nearby, they would release pollen into the air. Some of these grains would fall into the lake, sink to the bottom and be incorporated into the sediments. As the climate warmed and the ice sheet moved farther north, different plants would grow near the lake and their pollen grains would land in new layers of sediment. Today scientists can drill into that lake bottom, bring up a core and read the record of nearby plant life from the first colonizers at the bottom to the current tenants at the top.

If communities are defined as associations of plants that migrate together as well as simply live together, you would expect to see ecosystems marching north in lockstep. The fossil record seems clear, however; there is little or no evidence that entire groups of plants moved north together. Things that lived together in the past don't live together now, and things that live together now didn't live together in the past. Instead of stable collections of plants and animals following the glaciers north, we find each individual organism moving at its own pace. At any given time during a migration, particular organisms happen to be grouped together, but those groupings are just snapshots of a continuously changing reality.

What does this lack of community during migrations say about the webs of interaction we see in current ecosystems? And what does it tell us about how we should act to conserve the natural world? In an Italian restaurant near the University of Chicago, I raised these questions with two old friends, David Jablonski and Sue Kidwell, husband and wife paleontologists who, although they normally work in different areas, have come to remarkably similar conclusions on the subject of communities. Kidwell notes that "the fossil record really makes you question how much glue there is holding communities together. The plants and animals don't seem to care who they're living with, and all the different combinations seem to work equally well." To which Jablonski adds: "There seems to be something like a backlash developing against the idea of communities. There's a lot more interchangeability among members of an ecosystem than people had thought."

The two are quick to point out that as with most rules relating to living systems, there are exceptions to this one. Biologists can sometimes identify a so-called keystone species — a pollinating insect, for example — that plays such a crucial role that its removal can result in the upheaval of the entire ecosystem.

Nevertheless, interchangeability over geologic time seems to be much more the rule than the exception.

If the relationships between members of an ecosystem are indeed as malleable as this story suggests, then the lack of community in the northward movement of life after the glaciers is easy to understand. As each new association formed, an essential role that had been played by one species would be taken over by another. "Ecosystems can stand a little loss and invasion," says Jablonski. "Some give and take need not be fatal; in fact, it's natural."

To nail down the past, paleontologists are compiling databases of everything they have on past ecosystems. The palynologists, who study pollen, are key players, as are their colleagues who study phytoliths, microscopic glass-like secretions of plants long gone. The work involves more than academic curiosity. Policymakers have to take the best information that scientists can come up with and make difficult decisions.

As a people, we Americans have decided that we want to preserve at least a part of our natural environment. We can't preserve everything, however, and that means we have to have some way of choosing between competing conservation strategies. This choice depends, in turn, on what we think it is important to conserve. When we thought the most important thing to do was to save a species, we concentrated on how to save a particular sand dune or prairie that was home to that species. Today conservationists are more likely to talk about preserving entire ecosystems: the Greater Yellowstone Region, for example, or the Everglades.

It may be time, however, to focus on a different goal. We should concentrate on preserving the ability of plants and animals to respond to environmental changes. "We should be thinking about establishing reserves that cross environmental gradients," argues Jablonski. "We should think of ecosystems as sets of living organisms that need to expand and contract and shift around."

Such shifting already occurs in nature, of course. The Hawaiian Islands have been created by progressively newer volcanoes rising above the sea surface. As the older islands slowly sink again, the corals that are able to migrate move upward so that they remain at the same depth. Not all of them move fast enough. (Other organisms, for various reasons, have had an even harder time: Hawaii has probably lost a greater percentage of its flora and fauna than any other place in the United States.)

The fossil record seems to be telling us that we should be thinking about preserving species by giving them room to maneuver. Jablonski suggests, for example, that if we make an environmental preserve in the Amazon, we should include some foothills of the Andes to protect this kind of flexibility. That philosophy could guide the creation of new parks and preserves in other parts of the world, too.

In the United States, unfortunately, our parks and preserves are often surrounded by expensive development. It is very difficult to change boundaries that have existed for decades, even when we find out we should have done things differently. Had we known a century ago what we know now, we might, for example, have extended the boundaries of places like Yellowstone Park to include greater ranges of elevation. We might have bought more inland acres when establishing coastal preserves, to maintain the ability of coastal ecosystems to migrate landward if sea levels rise in the next century. Even if our current choices are limited, I expect that these new ideas will have an impact on how we spend our conservation dollars in the future.

The implications go further. If the climate does warm, natural ecosystems are not the only ones that will be moving north: those associated with agriculture will be doing the same. We normally think of an Iowa cornfield as a monoculture, but as every farmer knows, all sorts of varmints, weeds and insects live in even that simplified system. If the Corn Belt moves north to Minnesota, the paleobots are telling us, the rest of the "community" may not move with it like one big unhappy family. Knowing what will happen in advance could be worth billions of dollars to American agriculture.

As I left that restaurant and walked out into the cold Chicago night, I felt strangely elated. For one thing, it seemed to me that this new understanding could move the entire environmental debate to a new and more realistic level. Perhaps even more important, though, it confirmed my belief in the value of basic research, which is sometimes portrayed as the pursuit of useless knowledge. We have much to learn from something as seemingly "useless" as ancient pollen grains in the mud at the bottom of an ordinary lake.

By James Trefil