"Geologically, the further you go back, the less Earth you have to look at today," explains John P. Grotzinger, a geologist at MIT. Although entire mountain ranges date from the Permian—in western Texas, northern England, British Columbia, and Japan—much of this rock is not scientifically useful.

			To study such an ancient age, geologists need rocks that can be radiometrically dated based on the known rate of decay of uranium to lead; that contain fossilized animals or plants; or that provide specific information, usually trace elements that can be determined in ratios of one to another, about the nature of the paleoceanic environment. Because the Permian was so long ago, such data are subject to destruction and inaccessibility caused by more recent geologic events.
	Late Permian haline mode ocean circulation scenario in which convection is triggered by evaporation from the subtropics. The color map is the sea surface temperature, the overlapped contour is the stream function showing the pattern of horizontal flow. This haline mode is unstable and periodically—every few thousand years or so—flips to the thermal mode before slowly reestablishing itself. Click on the image to enlarge.

What's more, to figure out what might have happened to so many marine species, "We need a much better understanding of how late Permian oceans may have worked. That's why John Marshall's work is so important," says Andrew Knoll, a Harvard University paleobotanist and expert on mass extinctions. "He's building on the kind of circulation models that have been constructed to understand circulation in today's deep oceans. That's no mean problem in itself."