An ice age is any part of several periods or epochs of time when glaciers, especially in the form of great ice sheets (see \TGlacier\t AND GLACIATION), covered more of the Earth's surface than they do today. The term "ice age" has been used in two senses. First, it may refer to whole glacial epochs (the Pleistocene, for example), 2.5 million to 60 million years long, when the climate in the middle latitudes fluctuated wildly from warm to cold and glaciation ranged from today's small polar ice caps to massive, semiglobal ice sheets. Alternatively, it may describe a single glacial stage (such as the Wisconsin or Illinoisan) within these epochs, lasting for approximately 19,000 to 100,000 years each, when glaciers covered perhaps 20 to 50% of the then-existing continents (see \Tpaleoclimatology\t). Evidence of extensive repeated glaciation has been found for at least five stretches of geologic time: the middle of the Huronian Era in Precambrian time, beginning about 1,700 million to 2,300 million years ago; the end of the Proterozoic Era, in Precambrian time, starting about 670 million years ago; the middle of the Paleozoic Era, between the Ordovician and Silurian Periods, about 420 million years ago; the late Carboniferous and early Permian Periods, late in the Paleozoic Era, beginning 290 million years ago; and the Pleistocene Epoch of the Quaternary Period (Cenozoic Era), beginning at least 1.7 million years ago. Each ice age lasted more than a million years; most of the earlier ones lasted more than 10 million years. Some of the great ice sheets, such as the Ordovician-Silurian and the Permo-Carboniferous sheets, appear to have migrated back and forth repeatedly across a large paleocontinent over a period of 100 million years. The combined length of glacial stages of the five aforementioned ice ages was probably between 50 and 200 million years, or only 1 to 4% of the Earth's long history (4.6 billion years). Ice ages are thus unusual and very short episodes in the Earth's climatic history. EVIDENCE The existence of the five ice ages is verified by clear evidence coming from widely scattered parts of the Earth's surface. The knowledge that such evidence dates from the same part of one geologic period depends upon critical fossils of short duration above and below the glacial layers or upon intrusions of igneous rocks datable in years by isotopes. Beyond 10 million years ago, these dates may vary by a million years; between 290 million and 650 million years ago, by as much as 10 million years. From evidence of the last ice age, which lasted more than 1.7 million years, it can be inferred that the comings and goings of the great ice sheets were simultaneous on several continents. The proof of glaciation over continent-sized areas lies first in the widespread deposition of a unique kind of sediment or dirt called till (see \Ttill and tillite\t) that can be observed under all glaciers today. The till and related deposits additionally contain a wide variety of rock types (stones and boulders called glacial \Lerratic\ls) such as would be derived from widely disparate areas. Third, the till and erratics lie on one of the unique erosional surfaces of glaciation, such as grooved, striated, or polished bedrock pavement. Many other features associated with glaciers today are taken as indications of ancient ice ages: "streamlined" elongated ridges on the underlying rock called roche moutonnee; crescent-shaped nesting fractures in that rock; striated or faceted stones in the till; interspersed \Tgravel\t layers or lenses containing many rock types or all well-rounded pebbles; fine clay and silt laminae with dropstones from icebergs deposited in glacier-dammed lakes (see \Tvarved deposit\t); and uniformly sorted massive silt (not layered) deposited by glacial winds (see \Tloess\t). The older the rock record of an ice age, the less it is preserved, because rocks are altered by metamorphism over long periods of time. The Quaternary deposits of the last 2.5 million years are still easy-to-dig dirt, loose and slumping except in the case of certain ice-pressed and water-dried subglacial tills. But till deposits 420 million to 290 million years old have been contracted by wetting and drying, compressed by overlying layers, and cemented by chemicals carried in groundwater. Till, gravel, and varves from these earlier ice ages are found today in the form of tillite, \Tconglomerate\t, and laminated \Tshale\t, respectively. Buried and at times tilted, layers of these glacial rocks can be found at the surface today at scattered locations in mountain and valley regions. Beds of glacial materials 1,800 million to 650 million years old have been altered further to \Tmetamorphic\t \Lrock\ls (\Tphyllite\t, \Tquartzite\t, and \Tslate\t) by the heat and pressure of mountain building. These materials are even harder to identify as glacial till, except for the fact that they often rest on a smooth, striated surface--the clearest glacial evidence of all. Ancient Ice Ages Evidence for the oldest known glaciation occurs in basal middle-Huronian rocks, 2,300 to 1,700 million years old, of the Precambrian early Proterozoic Era. The best tillites from this ice age are in the Gowganda Series in Canada; they rest on striated pavement and have been dated as being 2,288 million (plus or minus 87 million) years old. Three altered tillites indicate three continental ice advances. This formation, which contains varied subangular erratics up to 5 m (16 ft) long and interbedded varved slates, extends across a 130,000-sq km (52,000-sq mi) area (only 20,000 sq km/8,000 sq mi is exposed) under west central Ontario near Sudbury. Other widespread formations of the same age (Kalevian) lie under Finland and South Africa. A similar formation has been found in upper Michigan, in the Fern Creek Formation, but it may date from an even earlier, Lower-Huronian interval. These deposits were laid down when all these land masses were at the North Pole. Another drastic ice age occurred 680 million years ago, and may or may not have triggered an enormous diversity of marine invertebrates whose fossil remains are found in the rocks of this time. On the west side of the Wasatch Mountains in Utah there is a 100-m-thick (330-ft) layer of sooty-gray phyllite rocks that were once tillites; they contain various erratic boulders up to 6 m (20 ft) long and lie on laminated or varved slates. Glacial sediments have been mapped in detail over an area of 250,000 sq km (100,000 sq mi) in the Adelaide geosyncline running roughly southeast to northwest through central Australia. The best tillites are preserved in eastern Greenland, Scotland, and Scandinavia and date from the late Proterozoic Era. Other tillites have been found in Belorussia and China. All have giant erractics or varied lithology, but the Bigganjargga in northern Norway has basal pavement to clinch the glacial origin. The most recently authenticated major glaciation is the one that took place at the Ordovician-Silurian boundary, about 420 million years ago. A layer of sedimentary marine tillites and beautifully grooved pavement is found all around the Sahara. Under it is a striated pavement of Ordovician sandstone containing \Ttrilobite\t fossils, and over it are dropstones in silt containing the remains of Silurian \Lgraptolite\ls. Central Africa was the Earth's south pole then, and evidence of glaciation from this period extends well up through Spain and west into eastern Brazil. The outstanding worldwide sediment record of the Pennsylvanian (Upper Carboniferous) Period shows cycles of rise and fall in sea level. It is not clear whether or not these cycles were glacially induced, but no better explanation has been put forth. Also, the evidence for glaciation simultaneously or just into the Permian Period (about 280 million years ago) is good. A series of marine or coal-swamp (shore) beds covered by several tillites containing coals or silts and dropstones, and being overlain by varved slates, lie under areas totaling thousands of square kilometers in five widely separated continents: the Buckeye and Pagoda Formations in Antarctica, the Dwyka Series in South Africa, the Itarare Group in Brazil, the Talchir Beds in India, and certain localities in Australia. Only in Australia does there appear to have been extensive mountain glaciation. Each area shows clear evidence of tillite resting on a fine striated and polished older rock surface. Although this evidence comes from far-flung parts of the globe, in Permo-Carboniferous time what are now today's continents then formed the paleocontinent \TGondwanaland\t, which was then centered about the South Pole. One gigantic ice sheet may have covered 50% of Gondwanaland. The Permo-Carboniferous rocks are the first to clearly show the effects of an ice age upon higher forms of life. Most notable were changes in plant life all over the world. Pennsylvanian trees, including the Calamites, nearly disappeared, and two new tree genuses, Glossopteris and Gangamopteris, proliferated far and wide in the Permian Period. Faunal changes were not so extensive, because no mammals existed then. Development and change in the dog-sized amphibians and reptiles were rapid in the following Triassic Period, but their migration relates mostly to the expansions and contractions of the seas. Quaternary Ice Age In the Pleistocene Epoch, beginning 1.7 million years ago, ice sheets developed on highlands in North America and Europe and spread over the northern half of North America and a quarter of Eurasia, dominating the Northern Hemisphere. As sea level fell and the ice covered more land area, glaciers spread over what is now the shallow seafloor of Hudson Bay and the Barents Sea. This glacial spread comprised 52% of the Pleistocene ice-covered area. The older Antarctic and Greenland ice sheets grew somewhat larger as well. Mantles of ice developed in the group of mountain ranges stretching from southern Alaska to Colorado and California, in the European Alps, in the Ural and Caucasus Mountains, and in the Himalayas. Mountain glaciers in the Southern Hemisphere (in the South American Andes, the New Zealand Alps, and western Tasmania) extended down onto the plains. Altogether the mountain glaciers made up 16 percent of the total ice area. The continental glaciers became very thick. Today more than half of the Antarctic ice sheet is more than 2,500 m (8,200 ft) thick; in central Greenland the ice is 2,000 m (6,600 ft) thick in most places. These have been measured by two drill holes, by seismic (wave) sounding, by radio-echo sounding, and by precise gravity measurements. Thickness has also been calculated by using a formula involving primarily the temperature of the basal ice and the slippage at the bottom. Similar calculations applied to the much larger Pleistocene ice sheets indicate that these were over 3,000 m (10,000 ft) thick in North America and 2,500 m (8,200 ft) thick in Europe. These thicknesses are independently confirmed by the isotopic ratios of oxygen, which differ between rainfall and snowfall ice; the oxygen isotopes of microfossil sediments that fell to the seafloor during the colder stages of Quaternary time indicate that the thickness of the ice averaged 2,500 m (8,200 ft) much of the time. The ice-age seas were lower than today's by 100 to 140 m (330 to 460 ft). Many wave-cut terraces and sea-cliffs are found that far below today's sea level. Elevated coral reefs and shore terraces, however, indicate that sea level was also 5 to 50 m (16 to 160 ft) higher than it is now during some interglacial intervals. By comparison, if all the Greenland and Antarctic ice melted today, sea level would rise nearly 65 m (215 ft). Ice-age sea levels are generally lower than those in the nonglacial 99 percent of geologic time. They certainly fluctuate wildly over the few tens of thousands of years during which the ice sheets grew and melted away. During the last ice age sea level rose or fell over 1 to 2 m (3 to 7 ft) per century. During the same period the ice crept over the land at 50 to 150 m (160 to 490 ft) per year. These Quaternary events occurred because of changes in climate. Cirques produced by local mountain glaciers that could have existed only during glacial stages are found 1,200 to 1,700 m (4,000 to 5,600 ft) lower than today's mountain glaciers; this indicates that the temperature in mid-latitude mountain ranges during these stages was 7 C degrees (13 F degrees) cooler than it is today. Careful study of temperature and salinity requirements of hundreds of species of living ocean-surface plankton (\Ldiatom\ls, \Tforaminifera\t, and \Tradiolaria\t) can be applied by regression curves to the undisturbed cores of ice-age ooze on the deep ocean floor. These measurements show that the surface ocean fauna and flora moved repeatedly 600 to 1,100 km (360 to 660 mi) north and then back south. Most ocean surface waters at the height of the glacial stages were at least 2 to 5 C degrees (4 to 9 F degrees) cooler and slightly less saline than today. The oxygen-isotope ratios of snow falling during the last ice age indicate that on the still-remaining ice sheets (Antarctica and Greenland ice at 1,300 and 1,150 m/4,300 and 3,800 ft depth), ice-age temperatures were as much as 6 to 8 C degrees (11 to 14 F degrees) colder, and the atmosphere was ten or more times dustier than it is today. The Quaternary ice age had great effects--far more than previous ice ages had--upon life forms, because more advanced life forms and more numerous species of fauna and flora were involved. The stresses of rapidly changing temperature, precipitation, and winds must have been great. Heavily insulated animal species--such as the woolly rhinoceros, the woolly mammoth, and the musk ox--developed, while slowly evolving Cenozoic species, such as the Pliocene horse and several fish, disappeared early in this ice age. A few new species of birds, amphibians, and field mice (microtines) died out. For some reason, giant mammals developed; for example, the ancestral \Tbison\t, the imperial \Tmammoth\t, and a giant bear and dog. The list of extinct Pleistocene mammals is long. The large ones--the mastodon, the great beavers, the sabertooth cats, the ground sloth, and the glyptodon--disappeared everywhere. Llamas, camels, tapirs, horses, yaks, and other species became extinct only in North America. Lower animal forms, such as mollusks and beetles, did not change much by adaptation, but they did migrate. River valleys became natural migratory highways; while some animal forms were frozen into extinction, others were able to escape the ice by migrating toward the equator. Plant life changed little during the 1.7 million years of the Quaternary ice age. Plant pollen and spores, being quite hardy, are preserved in bogs between glacial layers and provide complete records of the interglacial spasms and approaching glacial stages. Tree genuses in those times were similar to today's; although they failed to reproduce in some colder areas, they appeared and flourished anew in warmer areas as the ice slowly moved in. In North America, spruce grew best 1,200 to 1,600 km (745 to 990 mi) further south than usual--in North Carolina, Virginia, and Ohio--suggesting temperatures 7 to 10 C degrees (13 to 18 F degrees) cooler at that time. Trees lying in the path of the ice were crushed and swept up into the till. South of the ice, in temperate latitudes, a mixture of warm-climate species on hillsides facing south and cool-climate species on those facing north coexisted. The group of warmth-loving species returned in each succeeding warm interglacial time; in Europe these species were more diverse than in America, but all the trees of that time were similar to those of today. Today we live in a warm interlude during or just after the Quaternary ice age. All of the early development of humans came during this last ice age, and civilization has come into existence in its aftermath. According to current theory, the current warm interlude will end with a long-term cooling trend culminating in the next glacial stage, 23,000 years hence. CAUSES OF GLACIATION The only adequate source of water for such massive amounts of ice is the oceans. The creation of massive ice sheets on land thus necessarily depends on wind and weather patterns. To preserve snow from year to year, the summer climate must be cooler. There are so many ways these events could happen, and happen repeatedly, that the cause is locked into complicated \Locean-atmosphere interaction\ls. It is extremely difficult to know what is cause and what is effect. For example, the stratosphere was ten times dustier during glacial times than today. Dust not only absorbs some radiation in the upper atmosphere but also reflects some of the Sun's heat back out, thus cooling the Earth's surface. At first this was thought to be volcanic dust, but volcanism has not been correlated with any of the glacial stages. Dust may have been raised because much bare ocean shelf was exposed around receding seas. Evidence in dunes and loess blankets shows that in glacial times the wind was intensified, and the belts of westerlies were pushed toward the equator. Intensified heat exchange would also produce more clouds and precipitation; the increased cloudiness would be 80% effective in reflecting solar radiation back out into space; therefore the Earth's surface would become 1 to 2 C degrees (2 to 4 F degrees) cooler. A varying amount of carbon dioxide in the lower atmosphere may also have had some effect. Carbon dioxide lets in short-wave sunlight, but it also prevents long-wave heat radiation from passing out of the atmosphere, thus raising the atmospheric temperature between glaciations. In glacial stages, however, surface seawater would have been 2 to 6 C degrees (4 to 11 F degrees) cooler; this cooler water could absorb more carbon dioxide from the air and thus cool the latter by 1 to 2 C degrees (2 to 4 F degrees). When sea level was lower, many surface ocean currents no longer delivered heat to the far north (or south, in earlier ice ages); this would have dropped high-latitude temperature by 2 to 4 C degrees (4 to 7 F degrees). Once it had begun to form, the snow-ice surface itself would have then cooled the air masses contacting it by more than 5 C degrees (9 F degrees); this would have caused the further extension of sea ice that is recorded in the sandy sediments on the ocean floor. Sea ice, like clouds, is 80% effective in reflecting the radiant energy from the Sun, thus adding to the cooling effect. This also prevents free access of moisture into the air, the principal effect behind the Ewing-Donn theory of glaciation. The ice sheets, then, cut off from their source of moisture, begin to shrink, sea level correspondingly rises, and warm ocean currents begin to melt the sea ice. All these mechanisms are self-reinforcing, and each may contribute to the waxing and waning of any ice age. Continental glaciation requires either high mountain altitudes (over 4,000 m/13,000 ft) or polar land (high ground located above 60 deg latitude). \Tcontinental drift\t and shifting polar positions play a part here. Paleomagnetic analysis of ice-age sediments is now showing that today's continents were at one time united and near one pole or the other. Also, high mountains have arisen wherever continental plates have slid into each other, as India's current collision with Asia is raising the Himalayas. Unlike the self-reinforcing mechanisms, continental drift and mountain building occur so slowly that they could contribute only to whole ice ages, not to the drastic cycling of climate characteristic of the several glacial stages. Some trigger mechanism, most likely from outside the Earth or its atmosphere, affecting the climate seems to be necessary. Solar energy intensities have not yet been found to vary sufficiently to have single-handedly produced an ice age, but solar activity as expressed in sunspots and radio blackouts on an 11-year cycle has been shown to relate to short-term fluctuations in the Earth's climate. Longer cycles are currently being sought by statistical means. At present the \TMilankovitch theory\t, first proposed in 1924, is gaining widespread acceptance. According to this theory, the combination of a trio of orbital variations--eccentricity (a 97,000-year cycle), precession (22,000 years), and axial tilt (41,000 years)--produces effects that are large enough to cause cyclical expansion and contraction of the ice sheets. By taking the effects on climate and solar insolation of these variations and applying them to computer models of the behavior of ice sheets, scientists have succeeded in demonstrating a correlation between the cycles and the cyclic growth and decay of Pleistocene ice sheets over the past 600,000 years. Computer modeling provides a critical line of evidence by showing the combinations of orbital cycles result in lower summer insolation at 55 deg north latitude. Such cooler summers in high latitudes tend to preserve each winter's snowfall, and the accumulation of snow from a succession of winters induces growth of the northern ice sheets and the onset of a glacial stage. Richard P. Goldthwait Bibliography: Bahn, Paul G., Images of the Ice Age (1989); Calvin, William H., The Ascent of Mind: Ice Age Climates and the Evolution of Intelligence (1991); Chorlton, Windsor, Ice Ages (1983); Covey, Curt, "The Earth's Orbit and the Ice Ages," Scientific American, February 1984; Embleton, C., and King, C.A., Glacial and Periglacial Morphology, 2 vols., 2d ed. (1975); Flint, Richard F., Glacial and Quaternary Geology (1971); Fodor, Ronald, Frozen Earth: Explaining the Ice Ages (1981); Frenzel, Burkhard, Climatic Fluctuations of the Ice Age, trans. by A. E. Nairn (1973); Gray, Murray, The Quaternary Ice Age (1986); Imbrie, John and Katherine, The Ice Ages: Solving the Mystery (1979; repr. 1986); Kerr, R.A., "Milankovitch Climate Cycles through the Ages," Science, Feb. 27, 1987; Matthews, William H., The Story of Glaciers and the Ice Age (1974); Monastersky, Richard, "Ice Age Insights," Science News, Sept. 17, 1988; Sutcliffe, A.J., On the Track of Ice Age Mammals (1988); Williams, George, "Cosmic Signals Laid Down in Stone," New Scientist, June 25, 1987; Sutcliffe, A.J., On the Track of Ice Age Mammals (1988); Wright, A. E., and Moseley, F., Ice Ages: Ancient and Modern, Geological Journal Special Issue No. 6 (1980).