Unfortunately, just as Agassiz’s theory was at last beginning to find converts in Europe, he was busy taking it into ever more exotic territory in America. He began to find evidence for glaciers practically everywhere he looked, including near the equator. Eventually he became convinced that ice had once covered the whole Earth, extinguishing all life, which God had then re-created. None of the evidence Agassiz cited supported such a view. Nonetheless, in his adopted country his stature grew and grew until he was regarded as only slightly below a deity. When he died in 1873 Harvard felt it necessary to appoint three professors to take his place.

Yet, as sometimes happens, his theories fell swiftly out of fashion. Less than a decade after his death his successor in the chair of geology at Harvard wrote that the “so-called glacial epoch . . . so popular a few years ago among glacial geologists may now be rejected without hesitation.”

Part of the problem was that Croll’s computations suggested that the most recent ice age occurred eighty thousand years ago, whereas the geological evidence increasingly indicated that Earth had undergone some sort of dramatic perturbation much more recently than that. Without a plausible explanation for what might have provoked an ice age, the whole theory fell into abeyance. There it might have remained for some time except that in the early 1900s a Serbian academic named Milutin Milankovitch, who had no background in celestial motions at all-he was a mechanical engineer by training-developed an unexpected interest in the matter. Milankovitch realized that the problem with Croll’s theory was not that it was incorrect but that it was too simple.

As Earth moves through space, it is subject not just to variations in the length and shape of its orbit, but also to rhythmic shifts in its angle of orientation to the Sun-its tilt and pitch and wobble-all affecting the length and intensity of sunlight falling on any patch of land. In particular it is subject to three changes in position, known formally as its obliquity, precession, and eccentricity, over long periods of time. Milankovitch wondered if there might be a relationship between these complex cycles and the comings and goings of ice ages. The difficulty was that the cycles were of widely different lengths-of approximately 20,000, 40,000, and 100,000 years, but varying in each case by up to a few thousand years-which meant that determining their points of intersection over long spans of time involved a nearly endless amount of devoted computation. Essentially Milankovitch had to work out the angle and duration of incoming solar radiation at every latitude on Earth, in every season, for a million years, adjusted for three ever-changing variables.

Happily this was precisely the sort of repetitive toil that suited Milankovitch’s temperament. For the next twenty years, even while on vacation, he worked ceaselessly with pencil and slide rule computing the tables of his cycles-work that now could be completed in a day or two with a computer. The calculations all had to be made in his spare time, but in 1914 Milankovitch suddenly got a great deal of that when World War I broke out and he was arrested owing to his position as a reservist in the Serbian army. He spent most of the next four years under loose house arrest in Budapest, required only to report to the police once a week. The rest of his time was spent working in the library of the Hungarian Academy of Sciences. He was possibly the happiest prisoner of war in history.

The eventual outcome of his diligent scribblings was the 1930 book Mathematical Climatology and the Astronomical Theory of Climatic Changes. Milankovitch was right that there was a relationship between ice ages and planetary wobble, though like most people he assumed that it was a gradual increase in harsh winters that led to these long spells of coldness. It was a Russian-German meteorologist, Wladimir Köppen-father-in-law of our tectonic friend Alfred Wegener-who saw that the process was more subtle, and rather more unnerving, than that.

The cause of ice ages, Köppen decided, is to be found in cool summers, not brutal winters. If summers are too cool to melt all the snow that falls on a given area, more incoming sunlight is bounced back by the reflective surface, exacerbating the cooling effect and encouraging yet more snow to fall. The consequence would tend to be self-perpetuating. As snow accumulated into an ice sheet, the region would grow cooler, prompting more ice to accumulate. As the glaciologist Gwen Schultz has noted: “It is not necessarily the amount of snow that causes ice sheets but the fact that snow, however little, lasts.” It is thought that an ice age could start from a single unseasonal summer. The leftover snow reflects heat and exacerbates the chilling effect. “The process is self-enlarging, unstoppable, and once the ice is really growing it moves,” says McPhee. You have advancing glaciers and an ice age.

In the 1950s, because of imperfect dating technology, scientists were unable to correlate Milankovitch’s carefully worked-out cycles with the supposed dates of ice ages as then perceived, and so Milankovitch and his calculations increasingly fell out of favor. He died in 1958, unable to prove that his cycles were correct. By this time, write John and Mary Gribbin, “you would have been hard pressed to find a geologist or meteorologist who regarded the model as being anything more than an historical curiosity.” Not until the 1970s and the refinement of a potassium-argon method for dating ancient seafloor sediments were his theories finally vindicated.

The Milankovitch cycles alone are not enough to explain cycles of ice ages. Many other factors are involved-not least the disposition of the continents, in particular the presence of landmasses over the poles-but the specifics of these are imperfectly understood. It has been suggested, however, that if you hauled North America, Eurasia, and Greenland just three hundred miles north we would have permanent and inescapable ice ages. We are very lucky, it appears, to get any good weather at all. Even less well understood are the cycles of comparative balminess within ice ages, known as interglacials. It is mildly unnerving to reflect that the whole of meaningful human history-the development of farming, the creation of towns, the rise of mathematics and writing and science and all the rest-has taken place within an atypical patch of fair weather. Previous interglacials have lasted as little as eight thousand years. Our own has already passed its ten thousandth anniversary.

The fact is, we are still very much in an ice age; it’s just a somewhat shrunken one-though less shrunken than many people realize. At the height of the last period of glaciation, around twenty thousand years ago, about 30 percent of the Earth’s land surface was under ice. Ten percent still is-and a further 14 percent is in a state of permafrost. Three-quarters of all the fresh water on Earth is locked up in ice even now, and we have ice caps at both poles-a situation that may be unique in Earth’s history. That there are snowy winters through much of the world and permanent glaciers even in temperate places such as New Zealand may seem quite natural, but in fact it is a most unusual situation for the planet.

For most of its history until fairly recent times the general pattern for Earth was to be hot with no permanent ice anywhere. The current ice age-ice epoch really-started about forty million years ago, and has ranged from murderously bad to not bad at all. Ice ages tend to wipe out evidence of earlier ice ages, so the further back you go the more sketchy the picture grows, but it appears that we have had at least seventeen severe glacial episodes in the last 2.5 million years or so-the period that coincides with the rise of Homo erectus in Africa followed by modern humans. Two commonly cited culprits for the present epoch are the rise of the Himalayas and the formation of the Isthmus of Panama, the first disrupting air flows, the second ocean currents. India, once an island, has pushed two thousand kilometers into the Asian landmass over the last forty-five million years, raising not only the Himalayas, but also the vast Tibetan plateau behind them. The hypothesis is that the higher landscape was not only cooler, but diverted winds in a way that made them flow north and toward North America, making it more susceptible to long-term chills. Then, beginning about five million years ago, Panama rose from the sea, closing the gap between North and South America, disrupting the flows of warming currents between the Pacific and Atlantic, and changing patterns of precipitation across at least half the world. One consequence was a drying out of Africa, which caused apes to climb down out of trees and go looking for a new way of living on the emerging savannas.