nder the cold fluorescent lights of his laboratory at Fordham's Bronx, New York, campus, Robinson hands me a wad of ancient leaves dried to a parchment-like brittleness -- the gut contents of a mastodon found more than a century ago at a study site in Orange County. Holding this half-digested food in my hand brings the Pleistocene giant to life in a way that mounted museum skeletons cannot; I hear the snuffling of an elephantine trunk, and feel a hot, phantom breath on my shoulders.
Robinson's lean face glows in the dim light cast by his microscope. With a glance at a slide, he can tell what kinds of plants grew near Lake Fairchild thousands of years ago. Alder pollen resembles a squashed pentagon; oak pollen is shaped like the blade of a window fan, its surface bumpy as an orange peel.
The study of preserved pollen makes it possible for modern researchers to explore long-extinct ecosystems. And because the shifting patterns in dominant plant communities in the Northeast -- from conifer to oak forest, for instance -- are well known from fossil pollen data, Robinson can use pollen to place his samples in time.
He has found two dramatic shifts in the microfossil record that may explain the fate of New York's megafauna. First, Sporormiella spores vanish. Then, within a few hundred years, levels of microscopic charcoal -- the legacy of landscape fires -- increase more than tenfold. As Robinson reads this evidence, local populations of mastodons and other big grazers crashed when the first people arrived and found the naive animals easy prey. With most of the monster vegetarians wiped out, fuel built up on the landscape. So when the newcomers lit fires, as hunter-gatherer peoples have always done, they burned hotter and spread farther than ever before.
Both these shifts took place before the Younger Dryas event and the subsequent heat wave that closed the Pleistocene, supporting Robinson's argument that people, not weather, destroyed the megafauna. At the time the big herbivores vanished from southeastern New York, the pollen record shows no significant shift in local vegetation, as would be expected during a dramatic climate change.
When Paul Martin first presented his blitzkrieg theory, he assumed that people had arrived not long before the last of the megafauna died out about 11,000 years ago. That was a reasonable assumption: Radiocarbon dating puts the youngest megafauna bones and the artifacts of Clovis culture in that window of time. But Robinson's microfossils alter that scenario. Dating the soil layer in which spore numbers plummet puts the population crash at 12,500 years ago. "The populations collapsed well before the last of the animals died out," he says. "The data are telling me that the process didn't happen as quickly as in Paul's original blitzkrieg model."
Robinson believes that most of the big beasts did die out fast, hunted by humans. His controversial claim is that this happened more than a millennium before the last of the megafauna died, finally succumbing to continued human assaults. The youngest known fossils, he says, are the bones of a lonely few creatures that survived in slowly dwindling groups. Long before these last individuals expired, their species were doomed. "It's like Javan rhinos today," he says. "Their populations are functionally extinct. But there still happen to be a few individuals around."