Arthur Judson pulls on the steering lever and angles the Snowcat up the side of Colorado Mines peak. Outside, a bellicose wind buffets the tracked vehicle, and shards of sunlight poke through the Engelman spruce boughs, laden with a fresh snow.
''This is a good avalanche day - very good,'' says Mr. Judson, well bundled in a plaid wool shirt and tundra-proof boots.
He and a friend are venturing up this 12,500-foot peak to watch avalanches. But nature has something else in mind. A 50-mile-an-hour wind produces near white-out conditions on the treeless summit. The temperature is -70 F.
Despite the conditions, Judson and Ed Henion struggle out of the Snowcat for a better look.
Mr. Judson and Mr. Henion don't venture beyond arm's reach of each other, to avoid getting lost. Judson sidles up to a snowy cornice, where it drops several thousand feet on the other side. Then he retreats. ''I think I want to make that dinner with my wife tonight,'' he says.
Judson doesn't challenge one of nature's most ferocious, if fascinating, expressions just for the thrill of it. He and Edwin Henion are part of a small scientific community - fewer than 400 worldwide - who are trying to unravel the mysteries of avalanches: when they will occur, and what mischief they might bring about.
Judson is a researcher with the US Forest Service's Mountain Snow and Avalanche Research Project, run out of Fort Collins. He spends a lot of time with computer printouts, but periodically travels to this wind-swept pass to confront his subject face to mountain face.
Henion, stout and unassuming, is a project technician who works out of an instrument-filled field station at the foot of Mines peak here, about 70 miles west of Denver. ''On good days you can see a hundred avalanche paths from the summit,'' says Judson.
Avalanche research doesn't involve all the derring-do it once did. Judson remembers the day in 1961 when, still a cocky young ''snow ranger'' with the Forest Service, he ventured too far out on a snowfield on skis. He ended up buried up to his neck in snow - an uncomfortable experience that has given him a sound appreciation for avalanches: ''Snow sets up like concrete within a matter of minutes. You could be buried under a few inches and not get out.''
Such occurrences weren't unusual in the early days. The study of avalanches was less scientific, and one way to glean information was to take a ''ride.'' One legendary ranger, still around to spin yarns, buried himself some 50 times in a year in the pursuit of knowledge.
For centuries, avalanches baffled man (the Swiss once thought they were caused by devils). But today, research is leading, slowly, toward better avalanche prediction and, at least on paper, wiser construction in slide-prone areas.
Well it should. Avalanches are not nature's most violent outlet, but they have claimed 252 lives in the United States since 1950. Worse, the long-term trend has been on the upswing: Since 1970, an average of 13 people have perished in snow slides each year - up from three a year in the 1950s. Property damage usually tops $400,000 annually in the US. It is much higher in countries like Switzerland and Japan, where larger populations are shoe-horned into mountainous areas.
Meanwhile, a growing recreation ethic, reflected in the popularity of sports like cross-country skiing, is putting more people in remote areas. More condominiums are going up at the foot of snowcapped peaks. These facts add urgency to avalanche research.
Nowadays, the cause of these powerful but often eerily silent occurrences is no longer a matter of folklore. The villain behind most slides is an icy layer of snow known as ''depth hoar.'' In early winter, when the ground is still warm, heat percolates up into the top layers of snow to form a granular sheet of poorly bonded ice crystals. As snow piles up, usually helped by wind, the pack becomes unstable.
''Wind is the architect of avalanches,'' says Judson. ''Some 95 percent of all avalanches occur during snowstorms, when it's blowing and you get a tremendous transport of snow in a short time.''
Scientists also know avalanches emit sounds (''talk'') before they release. But they have had only limited success in using this warning siren to predict slides. Microphones have been planted in the snow. But these picked up noise from planes and trucks as well as the snow. They haven't been able to isolate the right symphony, Judson says.
For now, avalanche prediction remains mainly an alchemy of numbers - temperature, snow accumulation, wind speed - and human judgment. Forecasters don't do badly: At times they can predict avalanche cycles in an area with 80 percent accuracy. But Judson is working on a new, more precise computer model that may allow scientists to forecast slides for single mountain ranges instead of just for regions.
Most of the 10,000 avalanches reported in the US each year are considered ''small'' - roughly 50 feet long and 10 feet wide. But even these pack enough punch to knock a person down. The big ones, in effect a tobogganing mountain face, can topple a train and kick up a snowy ''dust cloud'' that will blacken a valley like a solar eclipse.
A chief problem in avalanche research is that scientists can't measure the stability of snow directly. One new tool will be tried this winter: US researchers will test a Swiss-developed radar system that measures snow depth, density, and water concentration.
Other high-tech tools have been tried - but not always with success.
To help control avalanches, for instance, scientists once beamed a laser at a snowfield to precipitate a slide. It didn't work, and Judson sees little merit in the idea. ''We'd probably end up cutting a ski lift in half,'' he quips. Instead, howitzers and hand-thrown explosives are usually used to trigger them.
Also being probed is the physics of avalanche flows. At Montana State University in Bozeman, scientists want to pinpoint what factors regulate the speed of snow slides, which would lead to more knowledge of the forces they unleash. Most travel at about 50 miles per hour, but a few have been clocked at speeds in excess of 120 m.p.h.
The key lies in the thin, granular layer at the bottom of the snow pack. It hugs the slope during a slide, forming a ball-bearing-like bed that the top layers glide over. As the snow picks up speed, the friction among the crystals drops. But why? And by how much?
Sometime this winter, in the name of science, either Theodore Lang, a professor of engineering mechanics, or his Montana State colleague Jim Dent will careen down a mountain slope on skis with sand glued to the bottom. The sand will represent the granular layer. Instruments will clock the skier's speed, in theory leading to clues about the frictional resistance of snow.
''It's a relatively unsophisticated test,'' admits Dr. Lang, who hasn't been on skis in three years. But he thinks it will solve puzzles about avalanche mechanics as well as mud and rock slides.