ON MAY 18, 1980, MOUNT ST. Helens exploded, the first major eruption of any volcano in recorded U.S. history. Obvious portents—big quakes, melting snow, a visible bulge—had led to partial evacuation of the area. But the eruption's scale and suddenness caught geologists off guard: 230 square miles were scoured, forests flattened like matchsticks, without any new precursors in the final hours. A young USGS geologist whom Newhall knew, David Johnston, was among the 57 dead.

Months later, with the mountain still firing occasional salvos, Newhall was appointed "hazards coordinator," charged with synthesizing data from seismologists, geochemists, and other specialists into practical advice. It was a breakthrough moment in his career. He invented simple equations based on things like the frequency and force of past eruptions, detection of active fissures, and visitors' proposed locations and length of stay. With these he estimated the risk of visiting St. Helens versus, say, being a wartime soldier or smoking cigarettes. According to his figures—perhaps off by ten times or more, he admitted—loggers salvaging blown-down trees had a one-in-a-thousand chance per year of being killed, provided they complied with evacuation alarms. This was about equal to routine work in a stone quarry, so in 1981 the loggers went in.

"This kind of cold calculation at first met with skepticism, and maybe horror," Newhall says. "But people need some way of making decisions."

Not that they'll always make smart ones. Active craters often exert an almost erotic magnetism for volcanologists. Many like whiffing toxic gases, feeling the ground shake, and generally peering into death's open mouth. Newhall is not like that; he avoids personal risk whenever possible. When he was on St. Helens, he argued with colleagues who he felt made unnecessary visits to the mile-and-a-half-wide crater, where boulders rattled down vertical walls and a dome of congealed lava unpredictably shot out searing gases and basketball-size rocks.

"Right up at the action, it's the little things that will get you—you don't need an eruption," says Newhall.

Of course, some tasks can be done only up close. These include collecting samples of volcanic gases directly from vents in vials, for detailed studies of magma chemistry and evolution. Still, when it comes to immediate hazard assessment, Newhall believes remote sensing now provides most of the essential data.

"Volcanologists take risks in order to reduce the risks to others," he says. "Before you go in, you should ask yourself: If I die, will my family and colleagues be able to say it was worth the data I got?"

Since the 1970s, about 35 volcanologists have been killed on the job. Many were Newhall's buddies or friends of friends. He counted them off to me one day over breakfast. His Dartmouth mentor, Dick Stoiber, had a grad student, Gary Malone, who fell off a cliff on Sicily's Stromboli volcano. There was Russian volcanologist Igor Menyailov, who along with five others was incinerated in 1993 when a lava dome blew at Colombia's Galeras volcano—an event later chronicled in two books. Two Indonesian friends, Asep Wildan and Mukti, were killed in a similar blast in 2000 at Semeru volcano; Newhall helped evacuate four battered survivors.

In the late eighties, Newhall co-wrote the script for a documentary designed to scare the hell out of civilians living near volcanoes so it would be easier to evacuate them. The footage—raging lahars and pyroclastic flows, corpses with burned-off faces, a tiny mud-spattered girl shaking with terror—was supplied by Maurice and Katia Krafft, a French husband-and-wife team famous for jetting to every eruption. On June 3, 1991, the Kraffts were in Japan filming Mount Unzen when a pyroclastic flow carbonized them, along with former USGS geologist Harry Glicken, who had narrowly escaped St. Helens. Some 40 others were also killed.

Most recently, on April 28, 2005, a helicopter carrying five PHIVOLCS staffers clipped a tree and crashed, exploding into flames. Among the dead was former PHIVOLCS director Ray Punongbayan—one of Newhall's closest friends.

"It's been a bad few decades," said Newhall, staring into his coffee cup.