In 1927 a Soviet expedition trekked to the Tunguska River in Siberia to discover a scene of peculiar devastation: an area twice the extent of that caused by the Mount St. Helens eruption, with flattened trees and branchless, standing trunks at the center. There was no crater, and no meteor fragments were found.

The "Tunguska event" was apparently caused by a meteor that entered the atmosphere at a sharp angle and evaporated after bursting into tiny particles. The explosion, which occurred in mid-air, was comparable to a large nuclear bomb. (A recent article in Science magazine says the explosion may have generated 30 million tons of nitric oxide, and depleted up to 40 percent of the ozone in the northern hemisphere.)

Sir Fred Hoyle, British astrophysicist (and science fiction novelist), believes that similar events caused the Ice Ages, and could occur again at any time, which presumably justifies the word "catastrophe" in his subtitle. If that giant meteorite had been even bigger, he writes, its heat (or impact, had it struck the ocean) could have vaporized and thrown into the stratosphere vast amounts of water. The water would have frozen into "diamond dust," which are tiny ice crystals whose brilliant reflections cause the dazzling optical effects -- halos, mock suns, arcs, coronas, iridescent fog -- that have long intrigued and bewildered explorers in the polar regions, where they are most common.

The heat balance and climate of the Earth depend greatly on the planet's reflectivity. If Sir Fred is right, and a cosmic accident were suddenly to fill the stratosphere with diamond dust, the Earth would act like a mirrored ball, bouncing back into space most of the sun's light and heat. We might be plunged into an instant Ice Age. "The switch could occur tomorrow," he says, "and no human being would escape its appalling consequences."

Our planet is constantly bombarded by galactic debris, most of which burns up in the atmosphere, causing "shooting stars." The pockmarked surface of the moon shows what Earth might look like were it not for erosion smoothing out most of the craters and erasing the scars. So it is not unreasonable to imagine extra-terrestrial forces might account for events that are otherwise difficult to understand. For example, Hoyle has proposed that interstellar dust may be the residue of microbes, and one of his books is "Diseases From Space." Luis Alvarez, citing evidence in geological strata, believes a giant asteroid kicked up a dust cloud that blocked out the sun for almost five years. With no sunlight penetrating the gloom, plants died and nearly half the Cretaceous-era animal and plant life, including the dinosaurs, became extinct. (A recent article in Nature magazine presents evidence challenging this, however.)

A first-class writer with Hoyle's vast knowledge of science can make a fascinating case for almost anything. And this book is certainly a page-turner.

But even brilliant hypotheses must be supported by evidence. Studies of deep sea sediments show that ice age cycles fit very well with the Milankovitch theory that the advance and retreat of the glaciers is related to variations in the Earth's orbit and plane of rotation (and thus to the intensity of solar radiation in the polar latitudes). Hoyle dismisses this idea with contempt. Evidence from the deep sea cores does not fit at all with his speculation about an essentially random series of events, such as collisions of Earth with extraterrestrial objects. Additionally, if Hoyle is right, we should find evidence that ice ages begin abruptly. The evidence, however, seems to square not at all with meteors and diamond dust: the cores show cycles of gradual cooling over about 10,000 years.

Hoyle's colleagues will likely accord a skeptical reception to the iconoclastic theory presented here. The clarity of his writing only accentuates his guesswork and assumptions -- untainted by scrutiny fropm scientific peers.

Right now, climatologists are concerned about man-made warming. The buildup of atmospheric carbon dioxide continues; two recent reports claim we can already discern evidence of a global warming caused by "the greenhouse effect." Readers interested in the ice age story might profit from John Imbrie's "Ice Ages: Solving the Mysteries" -- not as flashy as Sir Fred's deus ex machina, just a solidly sourced account of the patient unraveling of one of the exciting scientific puzzles of our time.