I all a dream.

    t sun inguisarsDid wander . . .

    —Byron, “Darkness”

    IN 1815 on t mountainnamed tambora exploded spectacularly, killing a s blast andassociated tsunamis. It  volcanic explosion in ten timest St.  to sixty tom bombs.

    Ne travel terribly fast in times ran a small story—actually a letter from a merc—seven monter t. But by time tambora’seffects . ty-six cubic miles of smoky as, and grit mospo cool.

    Sunsets  blearily colorful, an effect memorably captured by tist J. M.

    . turner, ly ted under anoppressive, dusky pall. It  inspired the Byron lines above.

    Spring never came and summer never summer. Crops everyed typy-five teen o Deats continued until June andalmost no planted seed ock died or o be prematurelyslaug  certainly t for farmers inmodern times. Yet globally temperature fell by only about 1.5 degrees Fa. Eartural tat, as scientists e instrument.

    teentury le Ice Age, as it ted all kinds of ry events—frost fairs on ting races along Dutc are mostly impossible no y ury geologists for beingsloo realize t t balmy compared  muc fair.

    t t. tteredic reindeer in tranded in improbable places—and ten came up ive but not terriblyplausible explanations. One Frencuralist named de Luc, trying to explain eboulders o rest one flanks of tains, suggestedt per t of apopgun. term for a displaced boulder is an erratic, but in teentury to apply more often to to the rocks.

    t Britis Arted t if James ton, ted Szerland,  once triations, telltale strand lines  point to passing ice ss. Unfortunately, ton  a traveler.

    But even ter at s, ton rejected out of  up mountainsides byfloods—all ter in t make a boulder float, ed out—and becameone of t to argue for ion. Unfortunately ice, andfor anotury most naturalists continued to insist t ttributed to passing carts or even ts.

    Local  peasants,  uncontaminated  by  scientific ortter, uralist Jean de Cier told tory of rylane ter o talking about tter matter-of-factly told  te some distance a tones ion,  ation: ‘transported t glacier extended in t as far as town of Bern.’ ”

    Cier  scientific gat  friends uralist, Louis Agassiz, ial skepticism came to embrace,and eventually all but appropriate, theory.

    Agassiz udied under Cuvier in Paris and nouralory at tel in Szerland. Anotanistnamed Karl Scually t to coin term ice age (in German Eiszeit ), in1837, and to propose t to s ice  just t over muc ion.  Agassiz es—to regret it as Agassizincreasingly got t for imacy, was heory.

    Cier liketer enemy of , yetanot least partly in mind ages in scientific discovery: first, people deny t it is true; t it isimportant; finally t the wrong person.

    At all events, Agassiz made t to understand tion,  everyo ts of t Alpine peaks, often apparently una eam  to climbtered an unyielding reluctance to accept heories.

    urged o return to ise, fossil fis Agassiz was a man possessed by an idea.

    Agassiz’s t in Britain, uralists en couldn’t grasp t ice in bulk exerts. “Could scratc be due to ice ?” asked Roderick Murcone at one meeting,evidently imagining t and glassy rime. to  incredulity at ts  for so mucy, endorsed t tion t ice could transportboulders presented “sucies” as to make it uny’s attention.

    Undaunted, Agassiz traveled tirelessly to promote o ameeting of tision for t of Science in Glasgo  Cy ofEdinburgion conceding t t be some general merit in t t certainly none of it applied to Scotland.

    Lyell did eventually come round.  of epip amoraine, or line of rocks, near ate in Scotland, ood if one accepted t a glacier ed, Lyell t  of t rating time for Agassiz. ly accusing  of ier  speak to est living geologist offered support of only t tepid and vacillating kind.

    In 1846, Agassiz traveled to America to give a series of lectures and t last found teem  -rate museum, tive Zoology. Doubtless it  tled in Neain sympaterminable periods ofcold. It also  six years after  scientific expedition to Greenlandreported t nearly t semicontinent  like t one imagined in Agassiz’s t long last, o find a realfolloral defect of Agassiz’s t assistance  to come from an unlikely quarter.

    In tions in Britain began to receive papers onatics, electricity, and otific subjects from a James Croll of Anderson’sUniversity in Glasgoions in Eart migated ice ages,  standard. So t a touc, urned out t Croll  an academic at ty, but a janitor.

    Born in 1821, Croll greed only to teen.  a variety of jobs—as a carpenter, insurance salesman, keeper of atemperance el—before taking a position as a janitor at Anderson’s (noy ofStrato do muco pass many quiet evenings in ty library teacronomy, atics, and to produce a string of papers, icular empions ofEart on climate.

    Croll  to suggest t cyclical c, fromelliptical (o elliptical again, mig and retreat of ice ages. No one  before to consider an astronomicalexplanation for variations in Eart entirely to Croll’s persuasivetain began to become more responsive to tion t at some formertime parts of ty and aptitude  tland and widely honored:

    y in London and of ty of St. Andrews, among much else.

    Unfortunately, just as Agassiz’s t last beginning to find converts in Europe, aking it into ever more exotic territory in America. o find evidence forglaciers practically everyually  ice inguised. None of ted supported suced country ature greil ly beloy.  it necessary to appoint to take hisplace.

    Yet, as sometimes ly out of faser e t ts may noed ation.”

    Part of t Croll’s computations suggested t t recent ice ageoccurred eigedt Eart of dramatic perturbation mucly t.

    it a plausible explanation for  migime except t in tin Milankovitcial motionsat all—raining—developed an unexpected interest in tter. Milankovitc t t it  but t it oo simple.

    As Eart is subject not just to variations in ts orbit, but also to rs in its angle of orientation to ts tilt and pitcing tensity of sunligcicular it is subject to tion, knos obliquity,precession, and eccentricity, over long periods of time. Milankovitcbe a relationsy  t lengtely 20,000,40,000, and 100,000 years, but varying in eaco a fe t determining ts of intersection over long spans of time involved a nearlyendless amount of devoted computation. Essentially Milankovitco  tion of incoming solar radiation at every latitude on Earted for three ever-changing variables.

    of repetitive toil t suited Milankovitcemperament. For t ty years, even ables of  noed ina day or ter. tions all o be made in ime, but in1914 Milankovitc a great deal of t o ion as a reservist in t most of tfour years under loose  in Budapest, required only to report to t of ime   prisoner of war in ory.

    tual  outcome  of   scribblings icalClimatology and tronomical tic Cc ttionsary  people  it  led to t eorologist, ladimir K?ppen—fatectonic friend Alfred egener—le, and rat.

    to be found in cool summers, not brutal ers.

    If summers are too cool to melt all t falls on a given area, more incoming sunligive surface, exacerbating t and encouraging yetmore snoo fall. tend to be self-perpetuating. As snoedinto an ice s, ting more ice to accumulate. As t G is not necessarily t of sno causes icess but t t snole, lasts.” It is t t an ice age could startfrom a single unseasonal summer. tover snos  and exacerbates t. “toppable, and once tmoves,” says McPhee. You have advancing glaciers and an ice age.

    In t dating tecists o correlateMilankovitc cycles es of ice ages as tcions increasingly fell out of favor. o prove t . By time, e Joo find a geologist or meteorologist y.” Not until t of a potassium-argon meting ancient seafloor sediments were ed.

    tc enougo explain cycles of ice ages. Many otors are involved—not least tion of tinents, in particular t tly understood. It ed,  if you  t and inescapable ice ages. e are very lucky, itappears, to get any good  all. Even less ood are tive balminess erglacials. It is mildly unnerving toreflect t tory—t of farming, tionof toics and ing and science and all t—aken placeypical patcerglacials ed as little as eigs ten th anniversary.

    t is, ill very muc’s just a some of t period of glaciation, aroundty t 30 percent of tenpercent still is—and a furt is in a state of permafrost. ters of all ter on Eart botuation t may be unique in Eartory. t ters t glaciers even in temperate places suce natural, but in fact it is a most unusual situation for t.

    For most of its ory until fairly recent times ttern for Earto be  ice anyed about fortymillion years ago, and o not bad at all. Ice ages tend to evidence of earlier ice ages, so tcuregro it appears t  seventeen severe glacial episodes in t 2.5million years or so—t coincides us in Africafolloed culprits for t epocion of t disrupting air flos. India, once an island, ers into t forty-five million years, raising not only t alsot tibetan plateau be tonly cooler, but diverted  made to more susceptible to long-term c five millionyears ago, Panama rose from ting ts betlantic, and cterns of precipitation across at least  ofAfrica,  of trees and go looking for a nehe emerging savannas.

    At all events, inents arranged as t appears t iceerm part of our future. According to Jo fifty more glacialepisodes can be expected, eacing a haw.

    Before fifty million years ago, Eart o be colossal. A massive freezing occurred about 2.2 billion years ago, follo—so large t some scientists are noo t occurred as tion is more popularly knoh.

    “Snoions. tbecause of a fall in solar radiation of about 6 percent and a dropoff in tion (orretention) of greenially lost its ability to o its . Itbecame a kind of all-over Antarctica. temperatures plunged by as muc. tire surface of t may o  itudes and tens of yards tropics.

    t tes iceevery asfirmly t t er someosynt t, but as youo peer t, ice quickly becomes opaque and after onlya feies ed. One is t alittle ocean er did remain exposed (pera  spot); t maybe t it remained translucent—a condition t does sometimes ure.

    If Eart question of  ever got  s so muc t it ay frozen forever. It appearst rescue may en interior. Once again, ed totectonics for alloo be  s of  and gases t melted tmosperestingly, tburst—time event of life’s ory. In fact, it may not ranquil as all t. As Eart probably   o raise o ts of skyscrapersand rainfalls of indescribable intensity.

    t all tubeo deep oceanvents undoubtedly  on as if not all ot ever o c entirely. It ime ago and at tage  don’t know.

    Compared burst, t times seem pretty smallscale, but of course tandards of anyto be found onEartoday. t,  a rate of about four  a year.

    a t must o be ts could benearly anding at t high.

    Be more ice, allest mountain summits poking tinents sagged under t of so mucertill rising back into place. ts didn’t just dribble outboulders and long lines of gravelly moraines, but dumped entire landmasses—Long Islandand Cape Cod and Nantucket, among ot along. It’s little  geologists before Agassiz rouble grasping tal capacity to reworklandscapes.

    If ice ss advanced again,  t Prince illiam Sound in Alaska, one of t glacial fields in Nort by trongest eartinent. It measured 9.2 on ter scale. Along t line, ty feet. t, in fact, t it made er slos of pools in texas. And  did tburst  all. tsoaked it up and kept on moving.

    For a long time it  t o and out of ice ages gradually, over  t  been toice cores from Greenland e for somet is found t comforting. It s for most of its recentory Eartable and tranquil place t civilization  ratly betal chill.

    to big glaciation, some toe rapidly, but tly plunged back into bitter cold for a t knoo science as ticplant t to recolonize land after an ice s  it  so s t average temperatures leapt again, by as mucy years,erribly dramatic but is equivalent to exce ofScandinavia for t of terranean in just tures teen degrees in ten years, drastically altering rainfall patterns and groions. t tling enouged planet. today tty well unimaginable.

    is most alarming is t ural ply rattle Earter. As Elizabet, ing in ternal force, or even any t emperature back and fortly, and as often, as to be to be, s and terrible feedback loop,”

    probably involving tions of tterns of ocean circulation, butall tood.

    One t ter to t tiness (and ty) of nortream to so trying to avoid a collision. Deprived of tream’s itudes returned to cions. But t begin toexplain veer as before. Instead, ranquility knoime in which we live now.

    to suppose t tretcic stability s much longer.

    In fact, some auties believe t   before. It isnatural to suppose t global  as a useful counterendency to plunge back into glacial conditions.  ed out, uating and unpredictable climate “t t todo is conduct a vast unsupervised experiment on it.” It ed, y t first seem evident, t an ice age migually be induced by arise in temperatures. t a sligion rates andincrease cloud cover, leading in titudes to more persistent accumulations of snow.

    In fact, global o pohern Europe.

    Climate is t of so many variables—rising and falling carbon dioxide levels, ts of continents, solar activity, tately c it is asdifficult to compres of t as it is to predict ture. Mucake Antarctica. For at least ty million years after it settled over tarctica remained covered in plants and free of ice. t simply s havebeen possible.

    No less intriguing are te dinosaurs. tisStepes t forests itude of togreat beasts, including tyrannosaurus rex. “t is bizarre,” es, “for sucitude is continually dark for t titudes suffered severe ers. Oxygen isotope studies suggest t te around Fairbanks, Alaska,  te Cretaceous period as it isnoyrannosaurus doing t migrated seasonally over enormousdistances or it spent mucs in tralia—time s orientation—a retreat to  possible. o survive in sucions can only be guessed.

    One t to bear in mind is t if ts did start to form again for  more er for to draime. t Lakes, less lakes of Canada—t to fuel t ice age. tedby it.

    On t pory could see us melting a lot of ice rat. If all ts melted, sea levels of a ty-story building—and every coastal city in ted. Morelikely, at least in t term, is t Antarctic ice s. In t fiftyyears ters around it igrade, and collapses ically. Because of tty quickly—by beteen and ty feet on average.

    traordinary fact is t  knoeamy . Only one tain: we live on a knife edge.

    In tally, ice ages are by no means bad ne. tuous ric freser lakest provide abundant nutritive possibilities for  as aspur to migration and keep t dynamic. As tim Flannery ion you need ask of a continent in order to determine te of its people: ‘Did you in mind, it’s time to look at a species of ape t trulydid.