what is the universe made of e
TRANSCRIPT
Every once in a while, cos-mologists are dragged,kicking and screaming,
into a universe much more unset-tling than they had any reason toexpect. In the 1500s and 1600s,Copernicus, Kepler, and Newtonshowed that Earth is just one ofmany planets orbiting one ofmany stars, destroying the com-fortable Medieval notion of aclosed and tiny cosmos. In the1920s, Edwin Hubble showedthat our universe is constantlyexpanding and evolving, a find-ing that eventually shattered theidea that the universe is unchang-ing and eternal. And in the pastfew decades, cosmologists havediscovered that the ordinary mat-ter that makes up stars and galax-ies and people is less than 5% ofeverything there is. Grapplingwith this new understanding ofthe cosmos, scientists face one overridingquestion: What is the universe made of?
This question arises from years of pro-gressively stranger observations. In the1960s, astronomers discovered that galaxiesspun around too fast for the collective pull ofthe stars’ gravity to keep them from flyingapart. Something unseen appears to bekeeping the stars from flinging themselvesaway from the center: unilluminated matterthat exerts extra gravitational force. This isdark matter.
Over the years, scientists have spottedsome of this dark matter in space; they haveseen ghostly clouds of gas with x-ray tele-scopes, watched the twinkle of distant stars asinvisible clumps of matter pass in front ofthem, and measured the distortion of space
and time caused by invisible mass in galaxies.And thanks to observations of the abun-dances of elements in primordial gas clouds,physicists have concluded that only 10% ofordinary matter is visible to telescopes.
But even multiplying all the visible “ordi-nary” matter by 10 doesn’t come close toaccounting for how the universe is structured.When astronomers look up in the heavenswith powerful telescopes, they see a lumpycosmos. Galaxies don’t dot the skies uni-formly; they cluster together in thin tendrilsand filaments that twine among vast voids.Just as there isn’t enough visible matter tokeep galaxies spinning at the right speed, thereisn’t enough ordinary matter to account forthis lumpiness. Cosmologists now concludethat the gravitational forces exerted by another
form of dark matter, made of an as-yet-undiscovered type of particle, must be
sculpting these vast cosmic structures.They estimate that this exotic dark matter
makes up about 25% of the stuff in the uni-verse—five times as much as ordinary matter.
But even this mysterious entity pales bycomparison to another mystery: dark energy.In the late 1990s, scientists examining distantsupernovae discovered that the universe isexpanding faster and faster, instead of slow-ing down as the laws of physics would imply.Is there some sort of antigravity force blow-ing the universe up?
All signs point to yes. Independent meas-urements of a variety of phenomena—cosmicbackground radiation, element abundances,galaxy clustering, gravitational lensing, gascloud properties—all converge on a consis-tent, but bizarre, picture of the cosmos. Ordi-nary matter and exotic, unknown particlestogether make up only about 30% of the stuffin the universe; the rest is this mysterious anti-gravity force known as dark energy.
This means that figuring out what the uni-verse is made of will require answers to threeincreasingly difficult sets of questions. Whatis ordinary dark matter made of, and wheredoes it reside? Astrophysical observations,such as those that measure the bending of lightby massive objects in space, are already yield-ing the answer. What is exotic dark matter?Scientists have some ideas, and with luck, adark-matter trap buried deep underground or ahigh-energy atom smasher will discover a newtype of particle within the next decade. Andfinally, what is dark energy? This question,which wouldn’t even have been asked adecade ago, seems to transcend knownphysics more than any other phenomenon yetobserved. Ever-better measurements of super-novae and cosmic background radiation aswell as planned observations of gravitationallensing will yield information about darkenergy’s “equation of state”—essentially ameasure of how squishy the substance is. Butat the moment, the nature of dark energy isarguably the murkiest question in physics—and the one that, when answered, may shedthe most light. –CHARLES SEIFE C
RED
IT:H
IGH
-Z S
UPER
NO
VA
SEA
RC
H T
EAM
/HST
/NA
SA
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org78
In the dark. Dark matter holds galaxies together; supernovaemeasurements point to a mysterious dark energy.
What Is the
Universe Made Of
Sp
ec
ial
Se
cti
on
From the nature of the cosmos to the nature of societies, the
following 100 questions span the sciences. Some are pieces
of questions discussed above; others are big questions in
their own right. Some will drive scientific inquiry for the next
century; others may soon be answered. Many will undoubtedly
spawn new questions.
So Much More to Know … >>
W H A T D O N ’ T W E K N O W ?
Is ours the only universe?
A number of quantum theorists
and cosmologists are trying to
figure out whether our universe
is part of a bigger “multiverse.”
But others suspect that this
hard-to-test idea may be a question
for philosophers.
What drove cosmic
inflation?
In the first moments
after the big bang, the
universe blew up at
an incredible rate. But
what did the blowing?
Measurements of the
cosmic microwave
background and other
astrophysical obser-
vations are narrowing
the possibilities.
Published by AAAS
For centuries, debating the nature ofconsciousness was the exclusivepurview of philosophers. But if therecent torrent of books on the topic
is any indication, a shift has taken place:Scientists are getting into the game.
Has the nature of consciousness finallyshifted from a philosophical question to ascientific one that can be solved by doingexperiments? The answer, as with any relatedto this topic, depends on whom you ask. Butscientific interest in this slippery, age-oldquestion seems to be gathering momentum.So far, however, although theories abound,hard data are sparse.
The discourse on consciousness has beenhugely influenced by René Descartes, theFrench philosopher who in the mid–17thcentury declared that body and mind aremade of differentstuff entirely. It mustbe so, Descartes con-cluded, because thebody exists in botht ime and space ,whereas the mind hasno spatial dimension.
Recent scientif-ically oriented acc-ounts of conscious-ness generally rejectDescartes’s solution;most prefer to treatbody and mind asdifferent aspects ofthe same thing. Inthis view, conscious-ness emerges fromthe properties andorganization of neu-rons in the brain. Buthow? And how can scientists, with theirdevotion to objective observation and meas-urement, gain access to the inherently privateand subjective realm of consciousness?
Some insights have come from examin-ing neurological patients whose injurieshave altered their consciousness. Damageto certain evolutionarily ancient structuresin the brainstem robs people of conscious-
ness entirely, leaving them in a coma or apersistent vegetative state. Although theseregions may be a master switch for con-sciousness, they are unlikely to be its solesource. Different aspects of consciousnessare probably generated in different brainregions. Damage to visual areas of the cere-bral cortex, for example, can producestrange deficits limited to visual awareness.One extensively studied patient, known asD.F., is unable to identify shapes or deter-mine the orientation of a thin slot in a verticaldisk. Yet when asked to pick up a card andslide it through the slot, she does so easily.At some level, D.F. must know the orienta-tion of the slot to be able to do this, but sheseems not to know she knows.
Cleverly designed experiments can pro-duce similar dissociations of unconscious
and conscious knowl-edge in people with-out neurological dam-age. And researchershope that scanningthe brains of subjectsengaged in such taskswil l reveal c luesabout the neuralactivity required forconscious awareness.Work with monkeysalso may elucidate
some aspects of consciousness, particularlyvisual awareness. One experimentalapproach is to present a monkey with an opti-cal illusion that creates a “bistable percept,”looking like one thing one moment andanother the next. (The orientation-flippingNecker cube is a well-known example.) Mon-keys can be trained to indicate which versionthey perceive. At the same time, researchers
hunt for neurons that track the monkey’s per-ception, in hopes that these neurons will leadthem to the neural systems involved in con-scious visual awareness and ultimately to anexplanation of how a particular pattern ofphotons hitting the retina produces the expe-rience of seeing, say, a rose.
Experiments under way at present gener-ally address only pieces of the consciousnesspuzzle, and very few directly address themost enigmatic aspect of the conscioushuman mind: the sense of self. Yet the exper-imental work has begun, and if the resultsdon’t provide a blinding insight into howconsciousness arises from tangles of neu-rons, they should at least refine the nextround of questions.
Ultimately, scientistswould like to understand
not just the biological basis ofconsciousness but also why it exists. Whatselection pressure led to its development,and how many of our fellow creatures shareit? Some researchers suspect that con-sciousness is not unique to humans, but ofcourse much depends on how the term isdefined. Biological markers for conscious-ness might help settle the matter and shedlight on how consciousness develops earlyin life. Such markers could also informmedical decisions about loved ones who arein an unresponsive state.
Until fairly recently, tackling the subjectof consciousness was a dubious career movefor any scientist without tenure (and perhapsa Nobel Prize already in the bag). Fortunately,more young researchers are now joining thefray. The unanswered questions should keepthem—and the printing presses—busy formany years to come.
–GREGMILLER
What Is the BiologicalBasis of Consciousness
79
CR
ED
IT:L
EST
ER
LEFK
OW
ITZ
/CO
RB
IS
Sp
ec
ial
Se
cti
on
continued >>
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005
W H A T D O N ’ T W E K N O W ?
When and how did the first stars
and galaxies form?
The broad brush strokes are visible, but the fine
details aren’t. Data from satellites and ground-
based telescopes may
soon help pinpoint,
among other particulars,
when the first genera-
tion of stars burned off
the hydrogen “fog” that
filled the universe.
Where do ultra-
high-energy cosmic
rays come from?
Above a certain
energy, cosmic
rays don’t travel
very far before being
destroyed. So why
are cosmic-ray
hunters spotting
such rays with no
obvious source within
our galaxy?
What powers
quasars?
The mightiest
energy fountains
in the universe
probably get their
power from matter
plunging into whirling
supermassive black
holes. But the details
of what drives
their jets remain
anybody’s guess.
What is the nature of
black holes?
Relativistic mass crammed
into a quantum-sized object?
It’s a recipe for disaster—and scientists are still
trying to figure out the ingredients.
JPL/NASA
E. J. SCHREIER, STSCI/NASA
Published by AAAS
CR
ED
IT:J
UPIT
ER
IM
AG
ES
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org80
Sp
ec
ial
Se
cti
on
When leading biologists wereunraveling the sequence of thehuman genome in the late 1990s,
they ran a pool on the number of genes con-tained in the 3 billion base pairs that makeup our DNA. Few bets came close. The con-ventional wisdom a decade or so ago wasthat we need about 100,000 genes to carryout the myriad cellular processes that keepus functioning. But it turns out that we haveonly about 25,000 genes—about the samenumber as a tiny flowering plant calledArabidopsis and barely more than the wormCaenorhabditis elegans.
That big surprise reinforced a growingrealization among geneticists: Our genomesand those of other mammals are far moreflexible and complicated than they onceseemed. The old notion of one gene/one pro-tein has gone by the board: It is now clear thatmany genes can make more than one protein.Regulatory proteins, RNA, noncoding bits ofDNA, even chemical and structural alter-ations of the genome itself control how,where, and when genes are expressed. Figur-ing out how all these elements work togetherto choreograph gene expression is one of thecentral challenges facing biologists.
In the past few years, it has become clearthat a phenomenon called alternative splicingis one reason human genomes can producesuch complexity with so few genes. Humangenes contain both coding DNA—exons—and noncoding DNA. In some genes, differentcombinations of exons can become active atdifferent times, and each combination yields adifferent protein. Alternative splicing waslong considered a rare hiccup during tran-scription, but researchers have concluded thatit may occur in half—some say close to all—of our genes. That finding goes a long waytoward explaining how so few genes canproduce hundreds of thousands of different
proteins. But how the transcription machin-ery decides which parts of a gene to read atany particular time is still largely a mystery.
The same could be said for the mechanismsthat determine which genes or suites of genesare turned on or off at particular times andplaces. Researchers are discovering that eachgene needs a supporting cast of hundreds to getits job done. They include proteins that shutdown or activate a gene, for example by addingacetyl or methyl groups to the DNA. Otherproteins, called transcription factors, interactwith the genes more directly: They bind tolanding sites situated near the gene under theircontrol. As with alternative splicing, activationof different combinations of landing sitesmakes possible exquisite controlof gene expression, butresearchers have yet tofigure out exactly howall these regulatoryelements really workor how they f it inwith alternativesplicing.
In the past decade or so, re-searchers have also come to appreci-
ate the key roles played by chromatinproteins and RNA in regulating geneexpression. Chromatin proteins areessentially the packaging for DNA,
holding chromosomes in well-def inedspirals. By slightly changing shape, chro-matin may expose different genes to thetranscription machinery.
Genes also dance to the tune of RNA.Small RNA molecules, many less than30 bases, now share the limelight with othergene regulators. Many researchers who oncefocused on messenger RNA and other rela-tively large RNA molecules have in the past5 years turned their attention to these smallercousins, including microRNA and smallnuclear RNA. Surprisingly, RNAs in thesevarious guises shut down and otherwisealter gene expression. They also are keyto cell differentiation in developing organ-
isms, but the mechanisms are notfully understood.
Researchers have madeenormous strides in pinpointing
these various mechanisms.By matching up genomesfrom organisms on differentbranches on the evolution-ary tree, genomicists arelocating regulatory regionsand gaining insights into
how mechanisms such asalternative splicing evolved.
These studies, in turn, shouldshed light on how these regions
work. Experiments in mice, such asthe addition or deletion of regulatoryregions and manipulating RNA,
and computer models shouldalso help. But the cen-tral question is likelyto remain unsolvedfor a long time: Howdo all these featuresmeld together to makeus whole?
–ELIZABETH PENNISI
0 10,000 20,000 30,000 40,000 50,000
D. melanogaster
C. elegans
Homo sapiens
Arabidopsis thaliana
Oryza sativa
Fugu rupides
Approximate number of genes
Why Do Humans Have So Few Genes
W H A T D O N ’ T W E K N O W ?
Why is there
more matter than
antimatter?
To a particle physicist,
matter and anti-
matter are almost
the same. Some
subtle difference
must explain why
matter is common
and antimatter rare.
Does the proton
decay?
In a theory of every-
thing, quarks (which
make up protons)
should somehow be
convertible to leptons
(such as electrons)—
so catching a proton
decaying into some-
thing else might
reveal new laws of
particle physics.
What is the
nature of gravity?
It clashes with
quantum theory.
It doesn’t fit in the
Standard Model.
Nobody has spotted
the particle that is responsible for it. Newton’s
apple contained a whole can of worms.
Why is time different
from other dimensions?
It took millennia for scien-
tists to realize that time is a
dimension, like the three
spatial dimensions, and that
time and space are inextrica-
bly linked. The equations
make sense, but they don’t
satisfy those who ask why
we perceive a “now” or why
time seems to flow the way
it does.
JUPITER IMAGES
JUPITER IMAGES
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 81
CR
EDIT
S (T
OP
TO B
OT
TOM
):JU
PIT
ER IM
AG
ES;A
FFY
MET
RIX
Sp
ec
ial
Se
cti
onF
orty years ago, doc-tors learned whysome patientswho received
the anesthetic succinyl-choline awoke normallybut remained tempo-rarily paralyzed andunable to breathe:They shared an inher-ited quirk that slowedtheir metabolism ofthe drug. Later, scien-tists traced sluggish succinyl-choline metabolism to a particular genevariant. Roughly 1 in 3500 people carry twodeleterious copies, putting them at high riskof this distressing side effect.
The solution to the succinylcholine mys-tery was among the first links drawn betweengenetic variation and an individual’s responseto drugs. Since then, a small but growingnumber of differences in drug metabolismhave been linked to genetics, helping explainwhy some patients benefit from a particulardrug, some gain nothing, and others suffertoxic side effects.
The same sort of variation, it is now clear,plays a key role in individual risks of comingdown with a variety of diseases. Gene vari-ants have been linked to elevated risks for dis-orders from Alzheimer’s disease to breastcancer, and they may help explain why, forexample, some smokers develop lung cancerwhereas many others don’t.
These developments have led to hopes—and some hype—that we are on the verge ofan era of personalized medicine, one in whichgenetic tests will determine disease risks andguide prevention strategies and therapies. Butdigging up the DNA responsible—if in factDNA is responsible—and converting thatknowledge into gene tests that doctors canuse remains a formidable challenge.
Many conditions, including various can-cers, heart attacks, lupus, and depression,likely arise when a particular mix of genescollides with something in the environment,such as nicotine or a
fatty diet. These multigeneinteractions are subtler and knot-
tier than the single gene drivers ofdiseases such as hemophilia and cystic
fibrosis; spotting them callsfor statistical inspiration
and rigorous experimentsrepeated again and againto guard against intro-ducing unproven gene
tests into the clinic. Anddetermining treatmentstrategies will be no lesscomplex: Last summer,for example, a team of sci-entists linked 124 differentgenes to resistance to fourleukemia drugs.
But identifying gene networks like these isonly the beginning. One of the toughest tasksis replicating these studies—an especiallydifficult proposition in diseases that are notoverwhelmingly heritable, such as asthma, orones that affect fairly small patient cohorts,such as certain childhood cancers. Many clin-ical trials do not routinely collect DNA fromvolunteers, making it sometimes difficult forscientists to correlate disease or drug responsewith genes. Gene microarrays, which measureexpression of dozens of genes at once, can befickle and supply inconsistent results. Genestudies can also be prohibitively costly.
Nonetheless, genetic dissection of somediseases—such as cancer, asthma, and heartdisease—is galloping ahead. Progress in otherareas, such as psychiatric disorders, is slower.Severely depressed or schizophrenic patientscould benefit enormously from tests that
reveal which drug and dose will help them themost, but unlike asthma, drug response can bedifficult to quantify biologically, making gene-drug relations tougher to pin down.
As DNA sequence becomes more avail-able and technologies improve, the geneticpatterns that govern health will likely comeinto sharper relief. Genetic tools still underconstruction, such as a haplotype map thatwill be used to discern genetic variationbehind common diseases, could furtheraccelerate the search for disease genes.
The next step will be designing DNA teststo guide clinical decision-making—and usingthem. If history is any guide, integrating suchtests into standard practice will take time. Inemergencies—a heart attack, an acute cancer,or an asthma attack—such tests will
be valuable only if they rapidly deliver results. Ultimately, comprehensive personal-
ized medicine will come only if pharma-ceutical companies want it to—and it willtake enormous investments in research anddevelopment. Many companies worry thattesting for genetic differences will narrowtheir market and squelch their profits.
Still, researchers continue to identify newopportunities. In May, the Icelandic companydeCODE Genetics reported that an experi-mental asthma drug that pharmaceutical giantBayer had abandoned appeared to decrease therisk of heart attack in more than 170 patientswho carried particular gene variants. The drugtargets the protein produced by one of thosegenes. The finding is likely to be just a fore-taste of the many surprises in store, as thebraids binding DNA, drugs, and disease areslowly unwound. –JENNIFER COUZIN
continued >>
To What Extent Are
Genetic Variation and
Personal Health Linked
W H A T D O N ’ T W E K N O W ?
Are there smaller
building blocks
than quarks?
Atoms were
“uncuttable.” Then
scientists discovered
protons, neutrons, and
other subatomic parti-
cles—which were, in
turn, shown to be made
up of quarks and glu-
ons. Is there something
more fundamental still?
Are neutrinos their
own antiparticles?
Nobody knows this
basic fact about
neutrinos, although a
number of under-
ground experiments
are under way.
Answering this ques-
tion may be a crucial
step to understanding
the origin of matter in
the universe.
Is there a unified theory
explaining all correlated
electron systems?
High-temperature superconductors
and materials with giant and
colossal magnetoresistance are all
governed by the collective rather
than individual behavior of electrons.
There is currently no common
framework for understanding them.
What is the most powerful laser
researchers can build?
Theorists say an intense enough
laser field would rip pho-
tons into electron-
positron pairs,
dousing the
beam. But no one
knows whether
it’s possible to
reach that point.
CERN
SANDIA NATIONAL LABORATORY
Published by AAAS
CR
EDIT
S (T
OP
TO B
OT
TOM
):JU
PIT
ER IM
AG
ES;J
UPI
TER
IMA
GES
;MEH
AU
KU
LYK
/PH
OTO
RES
EAR
CH
ERS
INC
.;C
ERN
/PH
OTO
RES
EAR
CH
ERS
INC
.
Sp
ec
ial
Se
cti
on
At its best, physics eliminates complex-ity by revealing underlying simplicity.Maxwell’s equations, for example,
describe all the confusing anddiverse phenomena of classicalelectricity and magnetism bymeans of four simple rules. Theseequations are beautiful; they havean eerie symmetry, mirroring oneanother in an intricate dance ofsymbols. The four together feel aselegant, as whole, and as completeto a physicist as a Shakespeareansonnet does to a poet.
The Standard Model of particlephysics is an unfinished poem.Most of the pieces are there, andeven unfinished, it is arguably themost brilliant opus in the literatureof physics. With great precision, itdescribes all known matter—allthe subatomic particles such asquarks and leptons—as well as theforces by which those particlesinteract with one another. Theseforces are electromagnetism,which describes how chargedobjects feel each other’s influence:the weak force, which explainshow particles can change theiridentities, and the strong force,which describes how quarks sticktogether to form protons and othercomposite particles. But as lovelyas the Standard Model’s descrip-tion is, it is in pieces, and some ofthose pieces—those that describegravity—are missing. It is a fewshards of beauty that hint at some-thing greater, like a few lines ofSappho on a fragment of papyrus.
The beauty of the StandardModel is in its symmetry; mathematiciansdescribe its symmetries with objects known
as Lie groups. And a mereglimpse at the Standard Model’sLie group betrays its fragmentednature: SU(3) × SU(2) × U(1).Each of those pieces representsone type of symmetry, but thesymmetry of the whole is bro-ken. Each of the forces behavesin a slightly different way, soeach is described with a slightlydifferent symmetry.
But those differences mightbe superficial. Electromagnet-ism and the weak force appearvery dissimilar, but in the 1960sphysicists showed that at hightemperatures, the two forces“unify.” It becomes apparent thatelectromagnetism and the weakforce are really the same thing,just as it becomes obvious thatice and liquid water are the samesubstance if you warm them uptogether. This connection ledphysicists to hope that the strongforce could also be unified withthe other two forces, yieldingone large theory described by asingle symmetry such as SU(5).
A unified theory should haveobservable consequences. Forexample, if the strong force trulyis the same as the electroweakforce, then protons might not be
truly stable; once in a long while, theyshould decay spontaneously. Despitemany searches, nobody has spotted a
proton decay, nor has anyone sighted anyparticles predicted by some symmetry-enhancing modifications to the StandardModel, such as supersymmetry. Worse yet,even such a unified theory can’t be com-plete—as long as it ignores gravity.
Gravity is a troublesome force. The theorythat describes it, general relativity, assumesthat space and time are smooth and continu-ous, whereas the underlying quantum physicsthat governs subatomic particles and forces isinherently discontinuous and jumpy. Gravityclashes with quantum theory so badly thatnobody has come up with a convincing wayto build a single theory that includes all theparticles, the strong and electroweak forces,and gravity all in one big bundle. But physi-cists do have some leads. Perhaps the mostpromising is superstring theory.
Superstring theory has a large followingbecause it provides a way to unify every-thing into one large theory with a singlesymmetry—SO(32) for one branch ofsuperstring theory, for example—but itrequires a universe with 10 or 11 dimen-sions, scads of undetected particles, and a lotof intellectual baggage that might never beverifiable. It may be that there are dozensof unified theories, only one of which is cor-rect, but scientists may never have the meansto determine which. Or it may be that thestruggle to unify all the forces and particlesis a fool’s quest.
In the meantime, physicists will continueto look for proton decays, as well as searchfor supersymmetric particles in undergroundtraps and in the Large Hadron Collider(LHC) in Geneva, Switzerland, when itcomes online in 2007. Scientists believe thatLHC will also reveal the existence of theHiggs boson, a particle intimately related tofundamental symmetries in the model ofparticle physics. And physicists hope that oneday, they will be able to finish the unfinishedpoem and frame its fearful symmetry.
–CHARLES SEIFE
Can the Laws ofPhysics Be Unified
Fundamental forces. A theory thatties all four forces together is stilllacking.
W H A T D O N ’ T W E K N O W ?
Is it possible to
create magnetic
semiconductors
that work at room
temperature?
Such devices have
been demonstrated
at low temperatures
but not yet in a
range warm enough
for spintronics
applications.
What is the pairing mechanism
behind high-temperature
superconductivity?
Electrons in superconductors surf
together in pairs. After 2 decades of
intense study, no one knows what
holds them together in the complex,
high-temperature materials.
Can we develop a
general theory of
the dynamics of
turbulent flows and
the motion of gran-
ular materials?
So far, such “nonequi-
librium systems” defy
the tool kit of statisti-
cal mechanics, and
the failure leaves a
gaping hole in
physics.
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org82
Can researchers make a perfect
optical lens?
They’ve done it with microwaves
but never with visible light.
JUPITER IMAGES
JUPITER IMAGES
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 83
CR
ED
IT:G
ET
TY
IM
AG
ES
Sp
ec
ial
Se
cti
on
When Jeanne Calment died in anursing home in southernFrance in 1997, she was 122years old, the longest-living
human ever documented. But Calment’suncommon status will fade in subsequentdecades if the predictions of some biolo-gists and demographers come true.Life-span extension in speciesfrom yeast to mice and extrapola-tion from life expectancy trends inhumans have convinced a swath ofscientists that humans will routinelycoast beyond 100 or 110 years of age.(Today, 1 in 10,000 people in industrial-ized countries hold centenarian status.)Others say human life span may be farmore limited. The elasticity found inother species might not apply to us. Further-more, testing life-extension treatments inhumans may be nearly impossible for practicaland ethical reasons.
Just 2 or 3 decades ago, research on agingwas a backwater. But when molecular biolo-gists began hunting for ways to prolong life,they found that life span was remarkablypliable. Reducing the activity of an insulinlikereceptor more than doubles the life span ofworms to a startling—for them—6 weeks. Putcertain strains of mice on near-starvation butnutrient-rich diets, and they live 50% longerthan normal.
Some of these effects may not occur in otherspecies. A worm’s ability to enter a “dauer”state, which resembles hibernation, may be
critical, for example. And shorter-lived speciessuch as worms and fruit flies, whose aging hasbeen delayed the most, may be more susceptibleto life-span manipulation. But successfulapproaches are converging on a few key areas:calorie restriction; reducing levels of insulinlikegrowth factor 1 (IGF-1), a protein; and prevent-
ing oxidative damage to the body’s tissues.
All three might be interconnected, but so farthat hasn’t been confirmed (although calorie-restricted animals have low levels of IGF-1).
Can these strategies help humans livelonger? And how do we determine whetherthey will? Unlike drugs for cancer or heartdisease, the benefits of antiaging treatmentsare fuzzier, making studies difficult to set upand to interpret. Safety is uncertain; calorierestriction reduces fertility in animals, and labflies bred to live long can’t compete with theirwild counterparts. Furthermore, garneringresults—particularly from younger volun-teers, who may be likeliest to benefit becausethey’ve aged the least—will take so long thatby the time results are in, those who began thestudy will be dead.
That hasn’t stopped scientists, some ofwhom have founded companies, fromsearching for treatments to slow aging. Oneintriguing question is whether calorierestriction works in humans. It’s beingtested in primates, and the National Instituteon Aging in Bethesda, Maryland, is fundingshort-term studies in people. Volunteers inthose trials have been on a stringent dietfor up to 1 year while researchers monitortheir metabolism and other factors thatcould hint at how they’re aging.
Insights could also come from geneticstudies of centenarians, who may haveinherited long life from their parents. Manyscientists believe that average human life spanhas an inherent upper limit, although theydon’t agree on whether it’s 85 or 100 or 150.
One abiding question in the antiagingworld is what the goal of all this work ought tobe. Overwhelmingly, scientists favor treat-ments that will slow aging and stave off age-
related diseases rather than simply extendinglife at its most decrepit. But even so, slowingaging could have profound social effects,upsetting actuarial tables and retirement plans.
Then there’s the issue of fairness: If anti-aging therapies become available, who willreceive them? How much will they cost?Individuals may find they can stretch theirlife spans. But that may be tougher toachieve for whole populations, althoughmany demographers believe that the averagelife span will continue to climb as it has con-sistently for decades. If that happens, muchof the increase may come from less dramaticstrategies, such as heart disease and cancerprevention, that could also make the end of along life more bearable. –JENNIFER COUZIN
How Much Can HumanLife Span Be Extended
continued >>
W H A T D O N ’ T W E K N O W ?
Is superfluidity
possible in a solid?
If so, how?
Despite hints in solid
helium, nobody is
sure whether a crys-
talline material can
flow without resist-
ance. If new types of
experiments show
that such outlandish
behavior is possible,
theorists would have
to explain how.
Are there stable
high-atomic-number
elements?
A superheavy element
with 184 neutrons
and 114 protons
should be relatively
stable, if physicists
can create it.
What is the
structure of
water?
Researchers continue
to tussle over how many
bonds each H2O molecule
makes with its nearest neighbors.
What is the nature
of the glassy state?
Molecules in a glass
are arranged much
like those in liquids
but are more tightly
packed. Where and
why does liquid end
and glass begin?
Are there limits to rational
chemical synthesis?
The larger synthetic mole-
cules get, the harder it is to
control their shapes and make
enough copies of them to be
useful. Chemists will need
new tools to keep their cre-
ations growing.
JUPITER IMAGES
Published by AAAS
CR
EDIT
:J.P
.BR
OC
KES
AN
D A
.KU
MA
R,N
AT
UR
E R
EV
IEW
S M
OL
EC
UL
AR
CE
LL
BIO
LO
GY3
,56
6 (
20
02
)
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org84
Sp
ec
ial
Se
cti
on
Unlike automobiles, humans get alongpretty well for most of their lives withtheir original parts. But organs do
sometimes fail, and we can’t go to themechanic for an engine rebuild or a new waterpump—at least not yet. Medicine has battledback many of the acute threats, such as infec-tion, that curtailed human life in past centuries.Now, chronic illnesses and deterioratingorgans pose the biggest drain on human healthin industrialized nations, andthey will only increase inimportance as the populationages. Regenerative medi-cine—rebuilding organs andtissues—could conceivably bethe 21st century equivalent ofantibiotics in the 20th. Beforethat can happen, researchersmust understand the signalsthat control regeneration.
Researchers have puzzledfor centuries over how bodyparts replenish themselves. Inthe mid-1700s, for instance,Swiss researcher AbrahamTrembley noted that whenchopped into pieces, hydra—tubelike crea-tures with tentacles that live in fresh water—could grow back into complete, new organ-isms. Other scientists of the era examined thesalamander’s ability to replace a severed tail.And a century later, Thomas Hunt Morganscrutinized planaria, flatworms that canregenerate even when whittled into 279 bits.But he decided that regeneration was anintractable problem and forsook planaria infavor of fruit flies.
Mainstream biology has followed in Mor-gan’s wake, focusing on animals suitable forstudying genetic and embryonic development.But some researchers have pressed on with
studies of regeneration superstars, and they’vedevised innovative strategies to tackle thegenetics of these organisms. These efforts andinvestigations of new regeneration models—such as zebrafish and special mouse lines—are beginning to reveal the forces that guideregeneration and those that prevent it.
Animals exploit three principal strategiesto regenerate organs. First, working organcells that normally don’t divide can multiply
and grow to replenish lost tissue, as occurs ininjured salamander hearts. Second, special-ized cells can undo their training—a processknown as dedifferentiation—and assume amore pliable form that can replicate and laterrespecialize to reconstruct a missing part.Salamanders and newts take this approach toheal and rebuild a severed limb, as dozebrafish to mend clipped fins. Finally, poolsof stem cells can step in to perform requiredrenovations. Planaria tap into this resourcewhen reconstructing themselves.
Humans already plug into these mecha-nisms to some degree. For instance, after surgi-cal removal of part of a liver, healing signals
tell remaining liver cells to resumegrowth and division to expand the organ
back to its original size. Researchers havefound that when properly enticed, some typesof specialized human cells can revert to a morenascent state (see p. 85). And stem cells helpreplenish our blood, skin, and bones. So whydo our hearts fill with scar tissue, our lensescloud, and our brain cells perish?
Animals such as salamanders and planariaregenerate tissues by rekindling genetic mech-anisms that guide the patterning of body struc-tures during embryonic development. Weemploy similar pathways to shape our parts asembryos, but over the course of evolution,humans may have lost the ability to tap into itas adults, perhaps because the cell divisionrequired for regeneration elevated the likeli-hood of cancer. And we may have evolved thecapacity to heal wounds rapidly to repel infec-tion, even though speeding the pace meansmore scarring. Regeneration pros such assalamanders heal wounds methodically andproduce pristine tissue. Avoiding fibrotic tissuecould mean the difference between regenerat-ing and not: Mouse nerves grow vigorously ifexperimentally severed in a way that preventsscarring, but if a scar forms, nerves wither.
Unraveling the mysteries of regenerationwill depend on understanding what separatesour wound-healing process from that of ani-mals that are able to regenerate. The differencemight be subtle: Researchers have identifiedone strain of mice that seals up ear holes inweeks, whereas typical strains never do. A rel-atively modest number of genetic differencesseems to underlie the effect. Perhaps altering ahandful of genes would be enough to turn usinto superhealers, too. But if scientists succeedin initiating the process in humans, new ques-tions will emerge. What keeps regeneratingcells from running amok? And what ensuresthat regenerated parts are the right size andshape, and in the right place and orientation? Ifresearchers can solve these riddles—and it’s abig “if ”—people might be able to order upreplacement parts for themselves, not just their’67 Mustangs. –R. JOHNDAVENPORT
R. John Davenport is an editor of Science’s SAGE KE.
What Controls Organ Regeneration
Self-repair.Newts reprogram their cells to reconstruct a severed limb.
W H A T D O N ’ T W E K N O W ?
What is the ulti-
mate efficiency of
photovoltaic cells?
Conventional solar
cells top out at con-
verting 32% of the
energy in sunlight to
electricity. Can
researchers break
through the barrier?
Will fusion always
be the energy
source of the
future?
It’s been 35 years away
for about 50 years, and
unless the international
community gets its
act together, it’ll be
35 years away for
many decades to come.
What drives the
solar magnetic
cycle?
Scientists believe
differing rates of
rotation from place
to place on the sun
underlie its 22-year
sunspot cycle. They
just can’t make it
work in their simula-
tions. Either a detail
is askew, or it’s back
to the drawing board.
How do planets form?
How bits of dust and ice
and gobs of gas came
together to form the
planets without the
sun devouring
them all is still
unclear. Planetary
systems around
other stars should
provide clues.
SOLA
R E
LEC
TR
IC P
OW
ER A
SSO
CIA
TIO
N
JPL/NASA
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 85
Sp
ec
ial
Se
cti
onL
ike Medieval alchemists whosearched for an elixir that couldturn base metals into gold, biol-ogy’s modern alchemists have
learned how to use oocytes to turn normalskin cells into valuable stem cells, and evenwhole animals. Scientists, with practice,have now been able to make nuclear transfernearly routine to produce cattle, cats, mice,sheep, goats, pigs, and—as a Korean teamannounced in May—even human embryonicstem (ES) cells. They hope to go still furtherand turn the stem cells into treatments forpreviously untreatable diseases. But like themedieval alchemists, today’s cloning andstem cell biologists are working largely withprocesses they don’t fully understand: Whatactually happens inside the oocyte to repro-gram the nucleus is still a mystery, and scien-tists have a lot to learn before they can directa cell’s differentiation as smoothly as nature’sprogram of development does every time afertilized egg gives rise to the multiple celltypes that make up a live baby.
Scientists have been investigating thereprogramming powers of the oocyte for halfa century. In 1957, developmental biologistsf irst discovered that theycould insert the nucleus ofadult frog cells into frog eggsand create dozens of geneti-cally identical tadpoles. But in50 years, the oocyte has yet togive up its secrets.
The answers lie deep incell biology. Somehow, scien-tists know, the genes that con-trol development—generallyturned off in adult cells—getturned back on again by theoocyte, enabling the cell totake on the youthful potentialof a newly fertilized egg. Sci-entists understand relativelylittle about these on-and-offswitches in normal cells, how-ever, let alone the unusualreversal that takes place dur-ing nuclear transfer.
As cells differentiate, their DNA becomesmore tightly packed, and genes that are nolonger needed—or those which should not beexpressed—are blocked. The DNA wrapstightly around proteins called histones, andgenes are then tagged with methyl groups thatprevent the proteinmaking machinery in thecell from reaching them. Several studies haveshown that enzymes that remove thosemethyl groups are crucial for nuclear transferto work. But they are far from the only thingsthat are needed.
If scientists could uncover the oocyte’ssecrets, it might be possible to replicate itstricks without using oocytes themselves, aresource that is fairly difficult to obtain and theuse of which raises numerous ethical ques-tions. If scientists could come up with a cell-free bath that turned the clock back on already-
differentiated cells,the implications couldbe enormous. Labscould rejuvenate cellsfrom patients and per-haps then grow theminto new tissue thatcould repair partsworn out by old ageor disease.
But scientists arefar from sure if suchcell-free alchemy ispossible. The egg’svery structure, itsscaffolding of pro-teins that guide thechromosomes duringcell division, mayalso play a key role inturning on the neces-
sary genes. If so, developing an elixir of pro-teins that can turn back a cell’s clock mayremain elusive.
To really make use of the oocyte’s power,scientists still need to learn how to direct thedevelopment of the rejuvenated stem cellsand guide them into forming specific tis-sues. Stem cells, especially those fromembryos, spontaneously form dozens ofcell types, but controlling that developmentto produce a single type of cell has provedmore difficult. Although some teams havemanaged to produce nearly pure colonies ofcertain kinds of neural cells from ES cells,no one has managed to concoct a recipe thatwill direct the cells to become, say, a purepopulation of dopamine-producingneurons that could replace thosemissing in Parkinson’s disease.
Scientists are just beginning to under-stand how cues interact to guide a celltoward its final destiny. Decades of workin developmental biology have provideda star t : Biologists have used mutantfrogs, flies, mice, chicks, and f ish toidentify some of the main genes that con-trol a developing cel l ’s decis ion tobecome a bone cell or a muscle cell. Butobserving what goes wrong when a geneis missing is easier than learning toorchestrate differentiation in a culturedish. Understanding how the roughly25,000 human genes work together toform tissues—and tweaking the rightones to guide an immature cell’s develop-ment—will keep researchers occupiedfor decades. If they succeed, however, theresult will be worth far more than itsweight in gold.
–GRETCHENVOGEL
How Can a Skin Cell
Become a Nerve Cell
continued >>
Cellular alchemist. A human oocyte.CR
EDIT
:DO
N W
.FA
WC
ETT
/VIS
UA
LS U
NLI
MIT
EDW H A T D O N ’ T W E K N O W ?
What causes reversals in
Earth’s magnetic field?
Computer models and laboratory
experiments are generating new
data on how Earth’s magnetic
poles might flip-flop. The trick will
be matching simulations to enough
aspects of the magnetic field
beyond the inaccessible core to
build a convincing case.
What causes
ice ages?
Something about the
way the planet tilts,
wobbles, and careens
around the sun pre-
sumably brings on ice
ages every 100,000
years or so, but reams
of climate records
haven’t explained
exactly how.
Are there earthquake precursors that
can lead to useful predictions?
Prospects for finding
signs of an imminent
quake have been waning
since the 1970s. Under-
standing faults will
progress, but routine
prediction would require
an as-yet-unimagined
breakthrough.
Is there—or was there—life
elsewhere in the solar system?
The search for life—past or pres-
ent—on other planetary bodies
now drives NASA’s planetary explo-
ration program, which focuses on
Mars, where water abounded when
life might have first arisen.
USGS
Published by AAAS
CR
EDIT
S:U
SH
RL/
AR
S/U
SD
A
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org86
Sp
ec
ial
Se
cti
on
It takes a certain amount of flexibility for aplant to survive and reproduce. It canstretch its roots toward water and its
leaves toward sunlight, but it has few optionsfor escaping predators or findingmates. To compensate, manyplants have evolved repair mecha-nisms and reproductive strategiesthat allow them to produce off-spring even without the meetingof sperm and egg. Some canreproduce from outgrowths ofstems, roots, and bulbs, but othersare even more radical, able to cre-ate new embryos from singlesomatic cells. Most citrus trees,for example, can form embryosfrom the tissues surrounding theunfertilized gametes—a feat noanimal can manage. The house-plant Bryophyllum can sproutembryos from the edges of itsleaves, a bit like Athena springingfrom Zeus’s head.
Nearly 50 years ago, scientists learnedthat they could coax carrot cells to undergosuch embryogenesis in the lab. Since then,people have used so-called somatic embryo-genesis to propagate dozens of species,including coffee, magnolias, mangos, androses. A Canadian company has plantedentire forests of fir trees that started life intissue culture. But like researchers who cloneanimals (see p. 85), plant scientists under-stand little about what actually controls theprocess. The search for answers might shedlight on how cells’ fates become fixed duringdevelopment, and how plants manage toretain such flexibility.
Scientists aren’t even sure which cells arecapable of embryogenesis. Although earlierwork assumed that all plant cells were equallylabile, recent evidence suggests that only a sub-
set of cells can transform into embryos. Butwhat those cells look like before their transfor-mation is a mystery. Researchers have video-taped cultures in which embryos develop butfound no visual pattern that hints at which cellsare about to sprout, and staining for certain pat-terns of gene expression has been inconclusive.
Researchers do have a few clues about themolecules that might be involved. In the lab,the herbicide 2,4-dichlorophenoxyacetic acid(sold as weed killer and called 2,4-D) canprompt cells in culture to elongate, build anew cell wall, and start dividing to formembryos. The herbicide is a synthetic analogof the plant hormones called auxins, which
control everything from theplant’s response to light and grav-ity to the ripening of fruit. Auxins
might also be important in naturalsomatic embryogenesis: Embryos that
sprout on top of veins near the leaf edgeare exposed to relatively high levels ofauxins. Recent work has also shown thatover- or underexpression of certain genes
in Arabidopsis plants can prompt embryo-genesis in otherwise normal-looking leaf cells.
Sorting out sex-free embryogenesis mighthelp scientists understand the cellularswitches that plants use to stay flexible while
still keeping growth under control. Develop-mental biologists are keen to learn how thosemechanisms compare in plants and animals.Indeed, some of the processes that controlsomatic embryogenesis may be similar tothose that occur during animal cloning orlimb regeneration (see p. 84).
On a practical level, scientists would liketo be able to use lab-propagation techniqueson crop plants such as maize that stillrequire normal pollination. That wouldspeed up both breeding of new varieties andthe production of hybrid seedlings—a flex-ibility that farmers and consumers couldboth appreciate. –GRETCHENVOGEL
How Does a SingleSomatic Cell Become A Whole Plant
W H A T D O N ’ T W E K N O W ?
What is the origin
of homochirality in
nature?
Most biomolecules
can be synthesized in
mirror-image shapes.
Yet in organisms,
amino acids are
always left-handed,
and sugars are
always right-handed.
The origins of this
preference remain a
mystery.
Can we predict how
proteins will fold?
Out of a near infini-
tude of possible ways
to fold, a protein
picks one in just tens
of microseconds.
The same task takes
30 years of computer
time.
How many proteins
are there in
humans?
It has been hard
enough counting
genes. Proteins can
be spliced in different
ways and decorated
with numerous func-
tional groups, all of
which makes count-
ing their numbers
impossible for now.
How do proteins find their
partners?
Protein-protein interactions are at
the heart of life. To understand how
partners come together in precise
orientations in seconds,
researchers need to know more
about the cell’s biochemistry and
structural organization.
PROTEIN DATA BANK
Power of one. Orange tree embryos can sprout from a single somatic cell.
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 87
CR
EDIT
:RO
GER
RES
SM
EYER
/CO
RB
IS
Sp
ec
ial
Se
cti
onT
he plate tectonics revolution wentonly so deep. True, it made wonderfulsense of most of the planet’s geology.But that’s something like understand-
ing the face of Big Ben; there must be a lotmore inside to understand about how and whyit all works. In the case of Earth, there’s another6300 kilometers of rock and iron beneath thetectonic plates whose churnings constitute theinner workings of a planetary heat engine.Tectonic plates jostling about the surface arelike the hands sweeping across the clock face:informative in many ways but largely mute asto what drives them.
Earth scientists inherited a rather simplepicture of Earth’s interior from theirpre–plate tectonics colleagues.Earth was like an onion. Seismicwaves passing through the deepEarth suggested that beneath thebroken skin of plates lies a 2800-kilometer layer of rocky mantle over-lying 3470 kilometers of molten and—at the center—solid iron. The mantle wasfurther subdivided at a depth of 670 kilo-meters into upper and lower layers, witha hint of a layer a couple of hundred kilo-meters thick at the bottom of the lower mantle.
In the postrevolution era, the onion modelcontinued to loom large. The dominant pictureof Earth’s inner workings divided the planetat the 670-kilometer depth, forming with thecore a three-layer machine. Above the 670, themantle churned slowly like a very shallow potof boiling water, delivering heat and rock atmid-ocean ridges to make new crust and coolthe interior and accepting cold sinking slabs ofold plate at deep-sea trenches. A plume of hotrock might rise from just above the 670 to forma volcanic hot spot like Hawaii. But no hot rockrose up through the 670 barrier, and no coldrock sank down through it. Alternatively,argued a smaller contingent, the mantlechurned from bottom to top like a deep stock-pot, with plumes rising all the way from thecore-mantle boundary.
Forty years of probing inner Earth with evermore sophisticated seismic imaging hasboosted the view of the engine’s complexity
without much calming the debate about how itworks. Imaging now clearly shows that the670 is no absolute barrier. Slabs penetrate theboundary, although with difficulty. Layered-earth advocates have duly dropped their impen-etrable boundary to 1000 kilometers or deeper.Or maybe there’s a flexible, semipermeableboundary somewhere that limits mixing to onlythe most insistent slabs or plumes.
Now seismic imaging is also outliningtwo great globs of mantle rock standingbeneath Africa and the Pacific like pistons.Researchers disagree whether they are hotterthan average and rising under their ownbuoyancy, denser and sinking, or merely pas-sively being carried upward by adjacent cur-rents. Thin lenses of partially melted rock dotthe mantle bottom, perhaps marking the bot-tom of plumes, or perhaps not. Geochemistsreading the entrails of elements and isotopesin mantle-derived rocks find signs of fivelong-lived “reservoirs” that must have resis-ted mixing in the mantle for billions of years.But they haven’t a clue where in the depths of
the mantle those reservoirs might be hiding.How can we disassemble the increasingly
complex planetary machine and find whatmakes it tick? With more of the same, plus alarge dose of patience. After all, plate tectonicswas more than a half-century in the making,and those revolutionaries had to look littledeeper than the sea floor.
Seismic imaging will continue to improveas better seismometers are spread more evenlyabout the globe. Seismic data are alreadydistinguishing between temperature andcompositional effects, painting an even morecomplex picture of mantle structure. Mineralphysicists working in the lab will tease outmore properties of rock under deep mantleconditions to inform interpretation of theseismic data, although still handicapped bythe uncertain details of mantle composition.And modelers will more faithfully simulate thewhole machine, drawing on seismics, mineralphysics, and subtle geophysical observationssuch as gravity variations. Another 40 yearsshould do it. –RICHARDA.KERR
How Does Earth’s
Interior Work
continued >>
A long way to go. Grasping theworkings of plate tectonics willrequire deeper probing.
How many forms of cell
death are there?
In the 1970s, apoptosis
was finally recognized as
distinct from necrosis.
Some biologists now
argue that the cell death
story is even more com-
plicated. Identifying new
ways cells die could lead
to better treatments for
cancer and degenerative
diseases.
What keeps intracellular traffic
running smoothly?
Membranes inside cells transport
key nutrients around, and through,
various cell compartments without
sticking to each other or
losing their way. Insights
into how membranes
stay on track could help
conquer diseases, such
as cystic fibrosis.
What enables
cellular components
to copy themselves
independent of
DNA?
Centrosomes, which
help pull apart paired
chromosomes, and
other organelles repli-
cate on their own
time, without DNA’s
guidance. This inde-
pendence still defies
explanation.
What roles do different
forms of RNA play in
genome function?
RNA is turning out to play a
dizzying assortment of roles,
from potentially passing genetic
information to offspring to muting
gene expression. Scientists are scram-
bling to decipher this versatile molecule.
C. SLAYDEN/SCIENCE
KATHARINE SUTLIFF/SCIENCE
Published by AAAS
CR
EDIT
:SET
I
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org88
Sp
ec
ial
Se
cti
on
Alone, in all that space? Not likely.Just do the numbers: Several hun-dred billion stars in our galaxy, hun-
dreds of billions of galaxies in the observ-able universe, and 150 planetsspied already in the immediateneighborhood of the sun. Thatshould make for plenty of warm,scummy little ponds where lifecould come together to beginbillions of years of evolutiontoward technology-wieldingcreatures like ourselves. No, thereally big question is when, ifever, we’ll have the technologi-cal wherewithal to reach out andtouch such intelligence. With abit of luck, it could be in thenext 25 years.
Workers in the search forextrater restrial intelligence(SETI) would have neededmore than a little luck in the first45 years of the modern hunt forlike-minded colleagues outthere. Radio astronomer Frank Drake’slandmark Project Ozma was certainly atriumph of hope over daunting odds. In1960, Drake pointed a 26-meter radio tele-scope dish in Green Bank, West Virginia, attwo stars for a few days each. Given thevacuum-tube technology of the time, hecould scan across 0.4 megahertz of themicrowave spectrum one channel at a time.
Almost 45 years later, the SETI Institutein Mountain View, California, completedits 10-year-long Project Phoenix. Oftenusing the 350-meter antenna at Arecibo,Puerto Rico, Phoenix researchers searched710 star systems at 28 million channelssimultaneously across an 1800-megahertz
range. All in all, the Phoenix searchwas 100 trillion times more effectivethan Ozma was.
Besides stunning advances in search
power, the first 45 years of modern SETIhave also seen a diversification of searchstrategies. The Search for ExtraterrestrialRadio Emissions from Nearby DevelopedIntelligent Populations (SERENDIP) hasscanned billions of radio sources in theMilky Way by piggybacking receivers onantennas in use by observational astro-nomers, including Arecibo. And othergroups are turning modest-sized opticaltelescopes to searching for nanosecondflashes from alien lasers.
Still, nothing has been heard. But then,Phoenix, for example, scanned just one ortwo nearby sunlike stars out of each 100 mil-lion stars out there. For such sparse sampling
to work, advanced, broadcasting civi-lizations would have to be abundant, or
searchers would have to get very lucky.To f ind the needle in a galaxy-size
haystack, SETI workers are counting on theconsistently exponential growth of computingpower to continue for another couple ofdecades. In northern California, the SETIInstitute has already begun constructing anarray composed of individual 6-meter anten-nas. Ever-cheaper computer power will even-tually tie 350 such antennas into “virtual
telescopes,” allowing scientists tosearch many targets at once. IfMoore’s law—that the cost of com-putation halves every 18 months—holds for another 15 years or so,SETI workers plan to use thisantenna array approach to checkout not a few thousand but perhapsa few million or even tens of mil-lions of stars for alien signals. Ifthere were just 10,000 advancedcivilizations in the galaxy, theycould well strike pay dirt beforeScience turns 150.
The technology may well beavailable in coming decades, butSETI will also need money. That’sno easy task in a field with as higha “giggle factor” as SETI has. TheU.S. Congress forced NASA towash its hands of SETI in 1993
after some congressmen mocked the wholeidea of spending federal money to look for“little green men with misshapen heads,” asone of them put it. Searching for anothertippy-top branch of the evolutionary tree stillisn’t part of the NASA vision. For more thana decade, private funding alone has drivenSETI. But the SETI Institute’s planned$35 million array is only a prototype of theSquare Kilometer Array that would putthose tens of millions of stars within reach ofSETI workers. For that, mainstream radioastronomers will have to be onboard—orwe’ll be feeling alone in the universe a long
time indeed.–RICHARDA. KERR
Are We Alone
In the Universe
Listening for E.T. The SETI Institute is deploying an array of antennas andtying them into a giant “virtual telescope.”
W H A T D O N ’ T W E K N O W ?
What role do
telomeres and cen-
tromeres play in
genome function?
These chromosome
features will remain
mysteries until new
technologies can
sequence them.
Why are some genomes really
big and others quite compact?
The puffer fish genome is
400 million bases; one
lungfish’s is 133 billion
bases long. Repeti-
tive and duplicated
DNA don’t explain
why this and other
size differences
exist.
What is all that
“junk” doing in our
genomes?
DNA between genes is
proving important for
genome function and
the evolution of new
species. Comparative
sequencing, microarray
studies, and lab work
are helping genomicists
find a multitude of genetic
gems amid the junk.
How much will new
technologies lower
the cost of
sequencing?
New tools and concep-
tual breakthroughs are
driving the cost of
DNA sequencing down
by orders of magni-
tude. The reductions
are enabling research
from personalized
medicine to evolution-
ary biology to thrive.
NIST
GETTY IMAGES
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 89
Sp
ec
ial
Se
cti
onF
or the past 50 years, scientists haveattacked the question of how lifebegan in a pincer movement. Someapproach it from the present, moving
backward in time from life today to its sim-pler ancestors. Others march forward fromthe formation of Earth 4.55 billion years ago,exploring how lifeless chemicals might havebecome organized into living matter.
Working backward, paleontologists havefound fossils of microbes dating back at least3.4 billion years. Chemical analysis of evenolder rocks suggests that photosyntheticorganisms were already well established onEarth by 3.7 billion years ago. Researcherssuspect that the organisms that left these tracesshared the same basic traits found in all lifetoday. All free-living organisms encodegenetic information in DNA and catalyzechemical reactions using proteins. BecauseDNA and proteins depend so intimately oneach other for their survival, it’s hard to imag-ine one of them having evolved first. But it’sjust as implausible for them to have emergedsimultaneously out of a prebiotic soup.
Experiments now suggestthat earlier forms of lifecould have been based ona third kind of moleculefound in today’s organ-isms: RNA. Onceconsidered nothingmore than a cellularcourier, RNA turnsout to be astonish-ingly versatile, notonly encoding geneticinformation but alsoacting like a protein.Some RNA moleculesswitch genes on and off, forexample, whereas others bind toproteins and other molecules. Laboratoryexperiments suggest that RNA could havereplicated itself and carried out the other func-tions required to keep a primitive cell alive.
Only after life passed through this “RNAworld,” many scientists now agree, did it takeon a more familiar cast. Proteins are thou-
sands of times more efficient as a catalystthan RNA is, and so once they emerged theywould have been favored by natural selec-tion. Likewise, genetic information can bereplicated from DNA with far fewer errorsthan it can from RNA.
Other scientists have focused their effortson figuring out how the lifeless chemistry of aprebiotic Earth could have given rise to anRNA world. In 1953, working at the Univer-sity of Chicago, Stanley Miller and HaroldUrey demonstrated that experiments couldshed light on this question. They ran an elec-tric current through a mix of ammonia,methane, and other gases believed at the timeto have been present on early Earth. They
found that they couldproduce amino acidsand other importantbuilding blocks of life.
Today, many sci-entists argue that theearly atmosphere wasdominated by other
gases, such as carbondioxide. But experi-
ments in recent years haveshown that under these con-
ditions, many building blocksof life can be formed. In addition,
comets and meteorites may have deliveredorganic compounds from space.
Just where on Earth these building blockscame together as primitive life forms is a sub-ject of debate. Starting in the 1980s, many sci-entists argued that life got its start in the scald-ing, mineral-rich waters streaming out of deep-
sea hydrothermal vents. Evidence for a hotstart included studies on the tree of life, whichsuggested that the most primitive species ofmicrobes alive today thrive in hot water. But thehot-start hypothesis has cooled off a bit. Recentstudies suggest that heat-loving microbes arenot living fossils. Instead, they may havedescended from less hardy species and evolvednew defenses against heat. Some skeptics alsowonder how delicate RNA molecules couldhave survived in boiling water. No singlestrong hypothesis has taken the hot start’splace, however, although suggestions includetidal pools or oceans covered by glaciers.
Research projects now under waymay shed more light on how life
began. Scientists are running experi-ments in which RNA-based cells may be ableto reproduce and evolve. NASA and theEuropean Space Agency have launchedprobes that will visit comets, narrowingdown the possible ingredients that mighthave been showered on early Earth.
Most exciting of all is the possibility offinding signs of life on Mars. Recent missionsto Mars have provided strong evidence thatshallow seas of liquid water once existed on theRed Planet—suggesting that Mars might oncehave been hospitable to life. Future Mars mis-sions will look for signs of life hiding in under-ground refuges, or fossils of extinct creatures.If life does turn up, the discovery could meanthat life arose independently on both planets—suggesting that it is common in the universe—or that it arose on one planet and spread to theother. Perhaps martian microbes were carriedto Earth on a meteorite 4 billion years ago,infecting our sterile planet. –CARL ZIMMER
Carl Zimmer is the author of Soul Made Flesh: TheDiscovery of the Brain—and How it Changed the World.
How and Where Did
Life on Earth Arise
continued >>
Cauldron of life? Deep-sea vents areone proposed site for life’s start.
CR
EDIT
:DA
VID
A.H
AR
DY
/PH
OTO
RES
EAR
CH
ERS IN
C.
W H A T D O N ’ T W E K N O W ?
How can genome
changes other
than mutations
be inherited?
Researchers are find-
ing ever more exam-
ples of this process,
called epigenetics,
but they can’t explain
what causes and pre-
serves the changes.
How do organs and
whole organisms
know when to stop
growing?
A person’s right and
left legs almost
always end up the
same length, and the
hearts of mice and
elephants each fit the
proper rib cage. How
genes set limits on
cell size and number
continues to mystify.
How is asymme-
try determined
in the embryo?
Whirling cilia
help an embryo
tell its left from its
right, but scientists
are still looking for the
first factors that give a rel-
atively uniform ball of cells a head,
tail, front, and back.
How do limbs, fins, and
faces develop and evolve?
The genes that determine the
length of a nose or the breadth of a
wing are subject to natural and sexual
selection. Understanding how selection
works could lead to new ideas about the
mechanics of evolution with respect to
development.JUPITER IMAGES
CENTER FOR FUNCTIONAL GENOMICS,SUNY AT ALBANY
Published by AAAS
CR
EDIT
S (T
OP
TO B
OT
TOM
):ST
EPH
EN F
RIN
K/C
OR
BIS
IMA
GES
;KEN
LU
CA
S/V
ISU
ALS
UN
LIM
ITED
;HA
NS
PFLE
TSC
HIN
GER
/PET
ER A
RN
OLD
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org90
Sp
ec
ial
Se
cti
on
Countless species of plants, animals,and microbes f ill every crack andcrevice on land and in the sea. They
make the world go ’round, converting sun-light to energy that fuels the rest of life,cycling carbon and nitrogen between inor-ganic and organic forms, and modifying thelandscape.
In some places and some groups, hun-dreds of species exist, whereas in others,very few have evolved; the tropics, forexample, are a complex paradise comparedto higher latitudes. Biologists are striving tounderstand why. The interplay betweenenvironment and living organisms andbetween the organisms themselves play keyroles in encouraging or discouraging diver-sity, as do human disturbances, predator-prey relationships, and other food web con-nections. But exactly how these and otherforces work together to shape diversity islargely a mystery.
The challenge is daunting. Baseline dataare poor, for example: We don’t yet knowhow many plant and animal species thereare on Earth, and researchers can’t evenbegin to predict the numbers and kinds oforganisms that make up the microbialworld. Researchers probing the evolutionof, and limits to, diversity also lack a stan-dardized time scale because evolution takesplace over periods lasting from days to mil-lions of years. Moreover, there can bealmost as much variation within a speciesas between two closely related ones. Nor isit clear what genetic changes will result in anew species and what their true influenceon speciation is.
Understanding what shapes diversitywill require a major interdisciplinaryeffort, involving paleontological interpre-
tation, f ield studies, laboratory experi-mentation, genomic comparisons, andeffective statistical analyses. A fewexhaustive inventories, such as the UnitedNations’ Millennium Project and anaround-the-world assessment of genes
from marine microbes, shouldimprove baseline data, but they will
barely scratch the surface. Modelsthat predict when one species will split
into two will help. And an emerging disci-pline called evo-devo is probing howgenes involved in development con-
tribute to evolution. Together, theseefforts will go a long way toward clarify-ing the history of life.
Paleontologists have already madeheadway in tracking the expansion andcontraction of the ranges of various organ-isms over the millennia. They are findingthat geographic distribution plays a keyrole in speciation. Future studies shouldcontinue to reveal large-scale patterns ofdistribution and perhaps shed more light onthe origins of mass extinctions and theeffects of these catastrophes on the evolu-tion of new species.
From field studies of plants and animals,researchers have learned that habitat caninfluence morphology and behavior—particularly sexual selection—in ways thathasten or slow down speciation. Evolutionarybiologists have also discovered that specia-tion can stall out, for example, as separatedpopulations become reconnected, homoge-nizing genomes that would otherwisediverge. Molecular forces, such as low muta-tion rates or meiotic drive—in which certainalleles have an increased likelihood of beingpassed from one generation to the next—influence the rate of speciation.
And in some cases, differences in diver-sity can vary within an ecosystem: Edges ofecosystems sometimes support fewer speciesthan the interior.
Evolutionary biologists are just begin-ning to sort out how all these factors areintertwined in different ways for differentgroups of organisms. The task is urgent:Figuring out what shapes diversity couldbe important for understanding the natureof the wave of extinctions the world isexperiencing and for determining strate-
gies to mitigate it. –ELIZABETH PENNISI
What DeterminesSpecies Diversity
W H A T D O N ’ T W E K N O W ?
What triggers
puberty?
Nutrition—including
that received in
utero—seems to help
set this mysterious
biological clock, but
no one knows exactly
what forces child-
hood to end.
Are stem cells at the
heart of all cancers?
The most aggressive
cancer cells look a
lot like stem cells.
If cancers are
caused by stem
cells gone awry,
studies of a cell’s
“stemness” may lead
to tools that could
catch tumors sooner
and destroy them
more effectively.
Is cancer susceptible
to immune control?
Although our immune
responses can suppress
tumor growth, tumor
cells can combat those
responses with counter-
measures. This defense
can stymie researchers hoping to develop
immune therapies against cancer.
Can cancers be controlled
rather than cured?
Drugs that cut off a tumor’s
fuel supplies—say, by stop-
ping blood-vessel growth—
can safely check or even
reverse tumor growth. But
how long the drugs remain
effective is still unknown.
M. CLARKE ET AL., PNAS 100 (7), 3983COPYRIGHT (2003) NAS, U.S.A.
L.H
EUSE
R A
ND
R.A
CK
LAN
D/U
NIV
ERSI
TY
OF
LOU
ISV
ILLE
SC
HO
OL
OF
MED
ICIN
E
Published by AAAS
91
CR
EDIT
S (L
EFT
TO
RIG
HT
):FR
ITZ
PO
LKIN
G/V
ISU
ALS
UN
LIM
ITED
;TER
RY H
USE
BY
E/G
ETT
Y IM
AG
ES
Sp
ec
ial
Se
cti
onE
very generation of anthropologistssets out to explore what it is thatmakes us human. Famed paleo-anthropologist Louis Leakey thought
tools made the man, and so when he uncov-ered hominid bones near stone tools in Tan-zania in the 1960s, he labeled the putativetoolmaker Homo habilis, the earliest memberof the human genus. But then primatologistJane Goodall demonstrated that chimps alsouse tools of a sort, and today researchersdebate whether H. habilis truly belongs inHomo. Later studies have honed in on traitssuch as bipedality, culture, language, humor,and, of course, a big brain as the uniquebirthright of our species. Yet many of thesetraits can also be found, at least to somedegree, in other creatures: Chimps have rudi-
mentary culture, parrots speak, and some ratsseem to giggle when tickled.
What is beyond doubt is that humans,like every other species, have a uniquegenome shaped by our evolutionary history.Now, for the first time, scientists can addressanthropology’s fundamental question at anew level: What are the genetic changes thatmake us human?
With the human genome in hand and pri-mate genome data beginning to pour in, weare entering an era in which it may becomepossible to pinpoint the genetic changes thathelp separate us from our closest relatives. Arough draft of the chimp sequence has alreadybeen released, and a more detailed version isexpected soon. The genome of the macaque isnearly complete, the orangutan is under way,and the marmoset was recently approved. All
these will help reveal the ancestral genotype atkey places on the primate tree.
The genetic differences revealed betweenhumans and chimps are likely to be profound,despite the oft-repeated statistic that onlyabout 1.2% of our DNA differs from that ofchimps. A change in every 100th base couldaffect thousands of genes, and the percentagedifference becomes much larger if you countinsertions and deletions. Even if we documentall of the perhaps 40 million sequence differ-ences between humans and chimps, what dothey mean? Many are probably simply theconsequence of 6 million years of geneticdrift, with little effect on body or behavior,whereas other small changes—perhaps inregulatory, noncoding sequences—may havedramatic consequences.
Half of the differences might define achimp rather than a human. How can wesort them all out?
One way is to zero in on the genes thathave been favored by natural selection inhumans. Studies seeking subtle signs ofselection in the DNA of humans andother primates have identified dozens ofgenes, in particular those involved inhost-pathogen interactions, reproduction,sensory systems such as olfaction and taste,and more.
But not all of these genes helped set usapart from our ape cousins originally. Ourgenomes reveal that we have evolved inresponse to malaria, but malaria defense didn’tmake us human. So some researchers havestarted with clinical mutations that impair keytraits, then traced the genes’ evolution, an
approach that has identified a handful of tanta-lizing genes. For example, MCPH1 and ASPM
cause microcephaly when mutated, FOXP2
causes speech defects, and all three show signsof selection pressure during human, but notchimp, evolution. Thus they may have playedroles in the evolution of humans’ large brainsand speech.
But even with genes like these, it is oftendifficult to be completely sure of what they do.Knockout experiments, the classic way toreveal function, can’t be done in humans andapes for ethical reasons. Much of the workwill therefore demand comparative analysesof the genomes and phenotypes of largenumbers of humans and apes. Already, someresearchers are pushing for a “great ape‘phenome’ project” to match the incomingtide of genomic data with more phenotypicinformation on apes. Other researchers arguethat clues to functioncan best be gleanedby mining naturalhuman variability,matching mutationsin living people to
subtle differences inbiology and behavior.Both strategies facelogistical and ethicalproblems, but some
progress seems likely.A complete understanding of uniquely
human traits will, however, include more thanDNA. Scientists may eventually circle backto those long-debated traits of sophisticatedlanguage, culture, and technology, in whichnurture as well as nature plays a leading role.We’re in the age of the genome, but we canstill recognize that it takes much more thangenes to make the human.
–ELIZABETH CULOTTA
What Genetic Changes Made UsUniquely Human
continued >>
W H A T D O N ’ T W E K N O W ?
Is inflammation a
major factor in all
chronic diseases?
It’s a driver of arthri-
tis, but cancer and
heart disease? More
and more, the answer
seems to be yes, and
the question remains
why and how.
How do prion
diseases work?
Even if one accepts
that prions are just
misfolded proteins,
many mysteries
remain. How can they
go from the gut to the
brain, and how do
they kill cells once
there, for example.
How much do vertebrates
depend on the innate immune
system to fight infection?
This system predates the verte-
brate adaptive immune response.
Its relative importance is
unclear, but immunol-
ogists are working
to find out.
Does immunologic memory
require chronic exposure to
antigens?
Yes, say a few prominent
thinkers, but experiments with
mice now challenge the theory.
Putting the debate to rest
would require proving that
something is not there, so the
question likely will not go away.
ART DAVIS/USDA
JUPITER IMAGES
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005Published by AAAS
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org92
Sp
ec
ial
Se
cti
on
Why doesn’t a pregnant woman
reject her fetus?
Recent evidence suggests that the mother’s
immune system doesn’t “realize”
that the fetus is foreign even
though it gets half its
genes from the father.
Yet just as Nobelist
Peter Medawar said
when he first raised
this question in 1952,
“the verdict has yet to
be returned.”
What synchronizes
an organism’s cir-
cadian clocks?
Circadian clock genes
have popped up in all
types of creatures and
in many parts of the
body. Now the chal-
lenge is figuring out
how all the gears fit
together and what
keeps the clocks set
to the same time.
How do migrating
organisms find their way?
Birds, butterflies, and
whales make annual
journeys of thou-
sands of kilometers.
They rely on cues
such as stars and magnetic
fields, but the details remain unclear.
Why do we sleep?
A sound slumber may
refresh muscles and organs
or keep animals safe from
dangers lurking in the dark.
But the real secret of sleep
probably resides in the brain,
which is anything but still
while we’re snoring away.
Packed into the kilogram or so of neuralwetware between the ears is every-thing we know: a compendium of use-
ful and trivial facts about the world, the his-tory of our lives, plus every skill we’ve everlearned, from riding a bike to persuading aloved one to take out the trash. Memoriesmake each of us unique, and they give conti-nuity to our lives. Understanding how mem-ories are stored in the brain is an essentialstep toward understanding ourselves.
Neuroscientists have already made greatstrides, identifying key brain regions andpotential molecular mechanisms. Still, manyimportant questions remainunanswered, and a chasm gapesbetween the molecular andwhole-brain research.
The birth of the modern era ofmemory research is often peggedto the publication, in 1957, of anaccount of the neurologicalpatient H.M. At age 27, H.M. hadlarge chunks of the temporallobes of his brain surgicallyremoved in a last-ditch effort torelieve chronic epilepsy. The sur-gery worked, but it left H.M.unable to remember anythingthat happened—or anyone hemet—after his surgery. The caseshowed that the medial temporallobes (MTL), which include thehippocampus, are crucial formaking new memories. H.M.’scase also revealed, on closerexamination, that memory is nota monolith: Given a tricky mirrordrawing task, H.M.’s perform-ance improved steadily over 3 dayseven though he had no memory of his previ-ous practice. Remembering how is not the
same as remembering what, as far as thebrain is concerned.
Thanks to experiments on animals andthe advent of human brain imaging, scien-tists now have a working knowledge of thevarious kinds of memory as well as whichparts of the brain are involved in each. Butpersistent gaps remain. Although the MTLhas indeed proved critical for declarativememory—the recollection of facts andevents—the region remains something of ablack box. How its various components inter-act during memory encoding and retrieval isunresolved. Moreover, the MTL is not the
f inal repository ofdeclarative memo-ries. Such memoriesare apparently filedto the cerebral cortexfor long-term stor-age, but how this hap-pens, and how mem-ories are representedin the cortex, remainsunclear.
More than a cen-tury ago, the greatS p a n i s h n e u r o -anatomist SantiagoRamón y Cajal pro-posed that makingm e m o r i e s m u s trequire neurons tostrengthen their con-nections with oneanother. Dogma atthe time held that nonew neurons a rebor n in the adul tbrain, so Ramón y
Cajal made the reasonable assumption thatthe key changes must occur between exist-
ing neurons. Until recently,scientists had few clues about
how this might happen. Since the 1970s, however, work
on isolated chunks of nervous-systemtissue has identified a host of molecu-lar players in memory formation.
Many of the same molecules have beenimplicated in both declarative and nonde-clarative memory and in species as varied assea slugs, fruit flies, and rodents, suggestingthat the molecular machinery for memoryhas been widely conserved. A key insightfrom this work has been that short-termmemory (lasting minutes) involves chemicalmodifications that strengthen existing con-nections, called synapses, between neurons,whereas long-term memory (lasting days orweeks) requires protein synthesis and proba-bly the construction of new synapses.
Tying this work to the whole-brainresearch is a major challenge. A potentialbridge is a process called long-term potentia-tion (LTP), a type of synaptic strengtheningthat has been scrutinized in slices of rodenthippocampus and is widely considered alikely physiological basis for memory. Aconclusive demonstration that LTP reallydoes underlie memory formation in vivowould be a big breakthrough.
Meanwhile, more questions keep pop-ping up. Recent studies have found that pat-terns of neural activity seen when an animalis learning a new task are replayed later dur-ing sleep. Could this play a role in solidifyingmemories? Other work shows that our mem-ories are not as trustworthy as we generallyassume. Why is memory so labile? A hintmay come from recent studies that revive thecontroversial notion that memories arebriefly vulnerable to manipulation each timethey’re recalled. Finally, the no-new-neuronsdogma went down in flames in the 1990s,with the demonstration that the hippocam-pus, of all places, is a virtual neuron nurserythroughout life. The extent to which thesenewborn cells support learning and memoryremains to be seen.
–GREGMILLER
How Are Memories
Stored and Retrieved
Memorable diagram. SantiagoRamón y Cajal’s drawing of thehippocampus. He proposed thatmemories involve strengthenedneural connections.
W H A T D O N ’ T W E K N O W ?
JUPITER IMAGES
JUPITER IMAGES
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 93
CR
EDIT
S (T
OP
TO B
OT
TOM
):TO
M B
RA
KEF
IELD
/CO
RB
IS;M
.TH
OM
SEN
/CO
RB
IS
Sp
ec
ial
Se
cti
onW
hen Charles Darwin wasworking out his grand theoryon the origin of species, hewas perplexed by the fact that
animals from ants to people form socialgroups in which most individuals work forthe common good. This seemed to runcounter to his proposal that individual fitnesswas key to surviving over the long term.
By the time hewrote The Descent ofMan, however, hehad come up with afew explanations. Hesuggested that natu-ral selection couldencourage altruisticbehavior among kinso as to improve thereproductive poten-tial of the “family.”He also introducedthe idea of reciproc-ity: that unrelated butfamiliar individualswould help eachother out if both werealtruistic. A centuryof work with dozensof social species has borne out his ideas tosome degree, but the details of how and whycooperation evolved remain to be workedout. The answers could help explain humanbehaviors that seem to make little sense froma strict evolutionary perspective, such asrisking one’s life to save a drowning stranger.
Animals help each other out in manyways. In social species from honeybees tonaked mole rats, kinship fosters cooperation:Females forgo reproduction and instead helpthe dominant female with her young. Andcommon agendas help unrelated individualswork together. Male chimpanzees, for exam-ple, gang up against predators, protectingeach other at a potential cost to themselves.
Generosity is pervasive among humans.Indeed, some anthropologists argue that theevolution of the tendency to trust one’s rela-tives and neighbors
helped humans become Earth’s dominantvertebrate: The ability to work together pro-vided our early ancestors more food, betterprotection, and better childcare, which inturn improved reproductive success.
However, the degree of cooperation varies.“Cheaters” can gain a leg up on the rest ofhumankind, at least in the short term. Butcooperation prevails among many species,
suggesting that thisbehavior is a bettersurvival strategy, overthe long run, despiteall the strife amongethnic, political, reli-gious, even familygroups now rampantwithin our species.
Evolutionary biologists and animalbehavior researchers are searching out thegenetic basis and molecular drivers of coop-erative behaviors, as well as the physiologi-cal, environmental, and behavioral impetusfor sociality. Neuroscientists studying mam-mals from voles to hyenas are discovering keycorrelations between brain chemicals andsocial strategies.
Others with a more mathematical bentare applying evolutionary game theory, amodeling approach developed for econom-ics, to quantify cooperation and predictbehavioral outcomes under different cir-cumstances. Game theory has helpedreveal a seemingly innate desire for fair-ness: Game players will spend time andenergy to punish unfair actions, eventhough there’s nothing to be gained by
these actions for themselves. Sim-
ilar studies have shown that even when twopeople meet just once, they tend to be fairto each other. Those actions are hard toexplain, as they don’t seem to follow thebasic tenet that cooperation is really basedon self-interest.
The models developed through thesegames are still imperfect. They do not ade-quately consider, for example, the effect ofemotions on cooperation. Nonetheless, withgame theory’s increasing sophistication,researchers hope to gain a clearer sense ofthe rules that govern complex societies.
Together, these efforts are helping socialscientists and others build on Darwin’sobservations about cooperation. As Darwinpredicted, reciprocity is a powerful fitnesstactic. But it is not a pervasive one.
Modern researchers have discovered thata good memory is a prerequisite: It seemsreciprocity is practiced only by organismsthat can keep track of those who are helpfuland those who are not. Humans have a greatmemory for faces and thus can maintain life-long good—or hard—feelings toward peoplethey don’t see for years. Most other speciesexhibit reciprocity only over very short timescales, if at all.
Limited to his personal observations,Darwin was able to come up with onlygeneral rationales for cooperative behavior.Now, with new insights from game the-ory and other promising experimentalapproaches, biologists are refining Darwin’sideas and, bit by bit, hope that one day theywill understand just what it takes to bring outour cooperative spirit.
–ELIZABETH PENNISI
continued >>
How Did Cooperative Behavior Evolve
W H A T D O N ’ T W E K N O W ?
Why do we dream?
Freud thought dream-
ing provides an outlet
for our unconscious
desires. Now, neuro-
scientists suspect
that brain activity
during REM sleep—
when dreams occur—
is crucial for learning.
Is the experience of
dreaming just a side
effect?
Why are there critical periods
for language learning?
Monitoring brain activity in young
children—including infants—may
shed light on why children pick up
languages with ease while adults
often struggle to learn train station
basics in a foreign tongue.
Do pheromones influence
human behavior?
Many animals use airborne chemi-
cals to communicate, particularly
when mating. Controversial studies
have hinted that humans too use
pheromones. Identifying them will
be key to assessing their sway on
our social lives.
How do general
anesthetics work?
Scientists are chip-
ping away at the
drugs’ effects on indi-
vidual neurons, but
understanding how
they render us uncon-
scious will be a
tougher nut to crack.
CORBIS
JUPITER IMAGES
Published by AAAS
W H A T D O N ’ T W E K N O W ?
CR
EDIT
:MIC
HA
EL T
.SH
IPLE
Y
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org94
Sp
ec
ial
Se
cti
on
Biology is rich in descriptive data—and getting richer all the time.Large-scale methods of probing
samples, such as DNA sequencing,microarrays, and automated gene-functionstudies, are f illing new databases to thebrim. Many subfields from biomechanics toecology have gone digital, and as a result,observations are more precise and moreplentiful. A central question now con-fronting virtually all f ields of biology iswhether scientists can deduce from this tor-rent of molecular data how systems andwhole organisms work. All this informationneeds to be sifted,organized, compiled,and—most impor-tantly—connected ina way that enablesresearchers to makepredictions based ongeneral principles.
Enter sys temsbiology. Looselydef ined and s t i l lstruggling to f indits way, this newlyemerging approacha i m s t o c o n n e c tthe dots that haveemerged from dec-ades of molecular,cellular, organismal, and even environmentalobservations. Its proponents seek to makebiology more quantitative by relying onmathematics, engineering, and computerscience to build a more rigid framework forlinking disparate findings. They argue that itis the only way the field can move forward.And they suggest that biomedicine, particu-larly deciphering risk factors for disease,will benefit greatly.
The field got a big boost from the comple-tion of the human genome sequence. Theproduct of a massive, trip-to-the-moon logis-tical effort, the sequence is now a hard andfast fact. The biochemistry of human inheri-tance has been defined and measured. Andthat has inspired researchers to try to makeother aspects of life equally knowable.
Molecular geneticists dream of having asimilarly comprehensive view of networksthat control genes: For example, they wouldlike to identify rules explaining how a singleDNA sequence can express different proteins,or varying amounts of protein, in different cir-
cumstances (see p. 80). Cell biologists wouldlike to reduce the complex communicationpatterns traced by molecules that regulate thehealth of the cell to a set of signaling rules.Developmental biologists would like a com-prehensive picture of how the embryo man-ages to direct a handful of cells into a myriadof specialized functions in bone, blood, andskin tissue. These hard puzzles can only besolved by systems biology, proponents say.
The same can be said for neuro-scientists trying to work out theemergent properties—higher thought, forexample—hidden in complex brain circuits.To understand ecosystem changes, includingglobal warming, ecologists need ways toincorporate physical as well as biological datainto their thinking.
Today, systems biologists have onlybegun to tackle relatively simple networks.They have worked out the metabolic path-way in yeast for breaking down galactose,a carbohydrate. Others have tracked thefirst few hours of the embryonic develop-
ment of sea urchins and otherorganisms with the goal of see-ing how various transcriptionfactors alter gene expressionover time. Researchers are alsodeveloping rudimentary modelsof signaling networks in cellsand simple brain circuits.
Progress is limited by the dif-ficulty of translating biologicalpatterns into computer models.Network computer programsthemselves are relatively simple,and the methods of portraying theresults in ways that researcherscan understand and interpret needimproving. New institutionsaround the world are gathering
interdisciplinary teams of biologists, mathe-maticians, and computer specialists to helppromote systems biology approaches. But itis still in its early days.
No one yet knows whether intensiveinterdisciplinary work and improved com-putational power will enable researchersto create a comprehensive, highly struc-tured picture of how life works.
–ELIZABETH PENNISI
How Will Big Pictures Emerge From a Sea of Biological Data
Systems approach. Circuit diagrams help clarify nerve cell functions.
What causes schiz-
ophrenia?
Researchers are try-
ing to track down
genes involved in this
disorder. Clues may
also come from
research on traits
schizophrenics share
with normal people.
What causes
autism?
Many genes probably
contribute to this baf-
fling disorder, as well
as unknown environ-
mental factors. A bio-
marker for early diag-
nosis would help
improve existing ther-
apy, but a cure is a
distant hope.
To what extent can we stave
off Alzheimer’s?
A 5- to 10-year delay in this
late-onset disease would
improve old age for millions.
Researchers are determining
whether treatments with
hormones or antioxidants, or
mental and physical exercise,
will help.
What is the biological
basis of addiction?
Addiction involves the dis-
ruption of the brain’s reward
circuitry. But personality
traits such as impulsivity
and sensation-seeking also
play a part in this complex
behavior.
ROYALTY-FREE/CORBIS
Published by AAAS
W H A T D O N ’ T W E K N O W ?
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 95
CR
ED
IT:E
RIC
V.G
RA
VE/P
HO
TO R
ESEA
RR
CH
ER
S IN
C.
Sp
ec
ial
Se
cti
onM
ost physical scientists nowadaysfocus on uncovering nature’smysteries; chemists build things.There is no synthetic astronomy
or synthetic physics, at least for now. Butchemists thrive on finding creative new waysto assemble molecules. For the last 100 years,they have done that mostly by makingand breaking the strong covalent bondsthat form when atomsshare electrons. Usingthat trick, they havelearned to combine asmany as 1000 atomsinto essentially anymolecular configura-tion they please.
Impressive as it is,this level of complex-ity pales in compari-son to what natureflaunts all around us.Everything from cellsto cedar trees is knittogether using a myr-iad of weaker links between small molecules.These weak interactions, such as hydrogenbonds, van der Waals forces, and π–π inter-actions, govern the assembly of everythingfrom DNA in its famous double helix to thebonding of H2O molecules in liquid water.More than just riding herd on molecules,such subtle forces make it possible for struc-tures to assemble themselves into an evermore complex hierarchy. Lipids coalesce toform cell membranes. Cells organize to formtissues. Tissues combine to create organisms.Today, chemists can’t approach the complex-ity of what nature makes look routine. Willthey ever learn to make complex structuresthat self-assemble?
Well, they’ve made a start. Over thepast 3 decades, chemists have made keystrides in learning the fundamental rules ofnoncovalent bonding. Among these rules:Like prefers like. We see this in hydro-phobic and hydrophilic interactions thatpropel lipid molecules in water to corraltogether to form the
two-layer membranes that serve as thecoatings surrounding cells. They bunchtheir oily tails together to avoid any inter-action with water and leave their morepolar head groups facing out into the liq-uid. Another rule: Self-assembly is gov-erned by energetically favorable reactions.Leave the right component moleculesalone, and they will assemble themselves
into complex ordered structures.Chemists have learned to take advantage
of these and other rules to design self-assembling systems with a modest degreeof complexity. Drug-carrying liposomes,made with lipid bilayers resembling those incells, are used commercially to ferry drugsto cancerous tissues in patients. And self-assembled molecules called rotaxanes,which can act as molecular switches thatoscillate back and forth between two stablestates, hold promise as switches in futuremolecular-based computers.
But the need for increased complexity isgrowing, driven by the miniaturization ofcomputer circuitry and the rise of nanotech-nology. As features on computer chips con-tinue to shrink, the cost of manufacturingthese ever-smaller components is skyrocket-ing. Right now, companies make them bywhittling materials down to the desired size.At some point, however, it will becomecheaper to design and build them chemicallyfrom the bottom up.
Self-assembly is also the only practicalapproach for building a wide variety ofnanostructures. Making sure the compo-nents assemble themselves correctly, how-ever, is not an easy task. Because the forcesat work are so small, self-assembling mole-cules can get trapped in undesirable confor-mations, making defects all but impossibleto avoid. Any new system that relies onself-assembly must be able either to toler-ate those defects or repair them. Again,biology offers an example in DNA. Whenenzymes copy DNA strands dur-
ing cell division, they invariably make mis-takes—occasionally inserting an A whenthey should have inserted a T, for example.Some of those mistakes get by, but most arecaught by DNA-repair enzymes that scanthe newly synthesized strands and correctcopying errors.
Strategies like that won’t be easy forchemists to emulate. But if they want tomake complex, ordered structures from theground up, they’ll have to get used to think-ing a bit more like nature.
–ROBERT F. SERVICE
How Far Can We Push Chemical Self-Assembly
continued >>
Is morality hard-
wired into the
brain?
That question has
long puzzled philoso-
phers; now some
neuroscientists think
brain imaging will
reveal circuits
involved in reasoning.
What are the limits
of learning by
machines?
Computers can
already beat the
world’s best chess
players, and they
have a wealth of infor-
mation on the Web to
draw on. But abstract
reasoning is still
beyond any machine.
How much of personality
is genetic?
Aspects of personality are
influenced by genes;
environment modi-
fies the genetic
effects. The rela-
tive contribu-
tions remain
under debate.
What is the biological root of
sexual orientation?
Much of the “environmental”
contribution to homosexuality
may occur before birth in the
form of prenatal hormones, so
answering this question will
require more than just the hunt
for “gay genes.”
LOUIE PSIHOYOS/CORBIS
JUPITER IMAGES
Published by AAAS
CR
EDIT
:JU
PIT
ER IM
AG
ES
96
Sp
ec
ial
Se
cti
on
At first glance, the ultimate limit ofcomputation seems to be an engi-neering issue. How much energy
can you put in a chip without melting it?How fast can you flip a bit in your siliconmemory? How big can youmake your computer andstill f it it in a room? Thesequestions don’t seem terri-bly profound.
In fact, computation ismore abstract and funda-mental than figuring out thebest way to build a com-puter. This realization camein the mid-1930s, whenPrinceton mathematiciansAlonzo Church and AlanTuring showed—roughlyspeaking—that any calcula-tion involving bits and bytescan be done on an idealizedcomputer known as a Turingmachine. By showing thatall classical computers areessentially alike, this dis-covery enabled scientists and mathemati-cians to ask fundamental questions aboutcomputation without getting bogged downin the minutiae of computer architecture.
For example, theorists can now classifycomputational problems into broad cate-gories. P problems are those, broadlyspeaking, that can be solved quickly, such asalphabetizing a list of names. NP problemsare much tougher to solve but relativelyeasy to check once you’ve reached ananswer. An example is the traveling sales-man problem, finding the shortest possibleroute through a series of locations. Allknown algorithms for getting an answertake lots of computing power, and even rel-
atively small versions might be out of reachof any classical computer.
Mathematicians have shown that if youcould come up with a quick and easy short-cut to solving any one of the hardest type
of NP problems, you’d be able to crackthem all. In effect, the NP problems wouldturn into P problems. But it’s uncertainwhether such a shortcut exists—whetherP = NP. Scientists think not, but provingthis is one of the great unanswered ques-tions in mathematics.
In the 1940s, Bell Labs scientist ClaudeShannon showed that bits are not just forcomputers; they are the fundamental unitsof describing the information that flowsfrom one object to another. There are phys-ical laws that govern how fast a bit canmove from place to place, how much infor-mation can be transferred back and forthover a given communications channel, and
how much energy it takes to erase a bitfrom memory. All classical information-processing machines are subject to theselaws—and because information seems tobe rattling back and forth in our brains, do
the laws of informationmean that our thoughts arereducible to bits and bytes?Are we merely computers?It’s an unsettling thought.
But there is a realmbeyond the classical com-puter: the quantum. Theprobabilistic nature of quan-tum theory allows atomsand other quantum objectsto store information that’snot restricted to only thebinary 0 or 1 of informationtheory, but can also be 0 and1 at the same time. Physicistsaround the world are buildingrudimentary quantum com-puters that exploit this andother quantum effects to dothings that are provably
impossible for ordinary computers, such asfinding a target record in a database with toofew queries. But scientists are still trying tofigure out what quantum-mechanical prop-erties make quantum computers so powerfuland to engineer quantum computers bigenough to do something useful.
By learning the strange logic of thequantum world and using it to do comput-ing, scientists are delving deep into thelaws of the subatomic world. Perhapssomething as seemingly mundane as thequest for computing power might lead to a newfound understanding of the
quantum realm.–CHARLES SEIFE
What Are the Limits of Conventional Computing
W H A T D O N ’ T W E K N O W ?
What is a species?
A “simple” concept
that’s been muddied by
evolutionary data; a
clear definition may be
a long time in coming.
Will there ever be a tree of life
that systematists can agree on?
Despite better morphological,
molecular, and statistical methods,
researchers’ trees
don’t agree. Expect
greater, but
not complete,
consensus.
Who was LUCA (the
last universal com-
mon ancestor)?
Ideas about the origin
of the 1.5-billion-
year-old “mother” of
all complex organ-
isms abound. The
continued discovery
of primitive microbes,
along with compara-
tive genomics, should
help resolve life’s
deep past.
How many species are
there on Earth?
Count all the stars in the sky?
Impossible. Count all the species
on Earth? Ditto. But the biodiver-
sity crisis demands that we try.
Why does lateral
transfer occur in so
many species and
how?
Once considered rare,
gene swapping, par-
ticularly among
microbes, is proving
quite common. But
why and how genes
are so mobile—and
the effect on fit-
ness—remains to be
determined.
M. DWORKIN,MICROBIOL. REVS. 60,70 (1996)
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org
W. F. DOOLITTLE, SCIENCE284, 2124 (1999)
Published by AAAS
W H A T D O N ’ T W E K N O W ?
97
Sp
ec
ial
Se
cti
onI
n the past few decades, organ transplan-tation has gone from experimental toroutine. In the United States alone, morethan 20,000 heart, liver, and kidney
transplants are performed every year. But fortransplant recipients, one prospect hasremained unchanged: a lifetime of taking pow-erful drugs to suppress the immune system, atreatment that can have serious side effects.Researchers have long sought ways to inducethe immune system to tolerate a transplantwithout blunting the body’s entire defenses,but so far, they have had limited success.
They face formidable challenges. Althoughimmune tolerance can occur—in rare cases,transplant recipients who stop taking immuno-suppressants have not rejected their foreignorgans—researchers don’t have a clear pictureof what is happening at the molecular and cel-lular levels to allow this to happen. Tinkeringwith the immune system is also a bit like tin-kering with a mechanical watch: Fiddle withone part, and you may disrupt the whole mech-anism. And there is a big roadblock to testingdrugs designed to induce tolerance: It is hard toknow if they work unless immunosuppressantdrugs are withdrawn, and that would risk rejec-tion of the transplant. But if researchers canfigure out how to train the immune system totolerate transplants, the knowledge could haveimplications for the treatment of autoimmunediseases, which also result from unwantedimmune attack—in these cases on some of thebody’s own tissues.
A report in Science 60 years ago fired thestarting gun in the race to induce transplant tol-erance—a race that has turned into a marathon.Ray Owen of the University of Wisconsin,Madison, reported that fraternal twin cattlesometimes share a placenta and are born witheach other’s red blood cells, a state referred toas mixed chimerism. The cattle tolerated theforeign cells with no apparent problems.
A few years later, Peter Medawar and histeam at the University of Birmingham, U.K.,showed that fraternal twin cattle with mixedchimerism readily accept skin grafts fromeach other. Medawar did not immediatelyappreciate the link to Owen’s work, but when
he saw the connection, he decided to injectfetal mice in utero with tissue from mice of adifferent strain. In a publication in Nature in1953, the researchers showed that, after birth,some of these mice tolerated skin grafts fromdifferent strains. This influential experimentled many to devote their careers to transplan-tation and also raised hopes that the workwould lead to cures for autoimmune diseases.
Immunologists, many of them workingwith mice, have since spelled out severaldetailed mechanisms behind tolerance. Theimmune system can, for example, dispatch“regulatory” cells that suppress immuneattacks against self. Or the system can forceharmful immune cells to commit suicide or to
go into a dysfunctional stupor called anergy.Researchers indeed now know fine detailsabout the genes, receptors, and cell-to-cellcommunications that drive these processes.
Yet it’s one matter to unravel how theimmune system works and another to figure
out safe ways to manipulate it. Transplantresearchers are pursuing three main strategiesto induce tolerance. One builds on Medawar’sexperiments by trying to exploit chimerism.Researchers infuse the patient with the organdonor’s bone marrow in hopes that the donor’simmune cells will teach the host to tolerate thetransplant; donor immune cells that comealong with the transplanted organ also, somecontend, can teach tolerance. A second strategyuses drugs to train T cells to become anergic orcommit suicide when they see the foreign anti-gens on the transplanted tissue. Thethird approach turns up produc-tion of T regulatory cells, whichprevent specific immune cells
from copying themselves and can also sup-press rejection by secreting biochemicalscalled cytokines that direct the immuneorchestra to change its tune.
All these strategies face a common prob-lem: It is maddeningly diff icult to judgewhether the approach has failed or succeededbecause there are no reliable “biomarkers” thatindicate whether a person has become tolerantto a transplant. So the only way to assess toler-ance is to stop drug treatment, which puts thepatient at risk of rejecting the organ. Similarly,ethical concerns often require researchers totest drugs aimed at inducing tolerance in con-cert with immunosuppressive therapy. This, inturn, can undermine the drugs’ effectivenessbecause they need a fully functioning immunesystem to do their job.
If researchers can complete their 50-yearquest to induce immune tolerance safely andselectively, the prospects for hundreds of thou-sands of transplant recipients would be greatlyimproved, and so, too, might the prospects forcontrolling autoimmune diseases.
–JON COHEN
Can We Selectively ShutOff Immune Responses
continued >>
CR
ED
IT:M
AR
K H
AR
MEL/
GET
TY
IM
AG
ES
How did flowers evolve?
Darwin called this question an “abominable mys-
tery.” Flowers arose in the cycads and conifers,
but the details of their evolution remain obscure.
How is plant growth
controlled?
Redwoods grow to be
hundreds of meters tall,
Arctic willows barely
10 centimeters.
Understanding the
difference could lead to
higher-yielding crops.
How do plants make
cell walls?
Cellulose and pectin walls
surround cells, keeping water
in and supporting tall trees.
The biochemistry holds the
secrets to turning its bio-
mass into fuel.
Why aren’t all plants immune
to all diseases?
Plants can mount a general
immune response, but they also
maintain molecular snipers that
take out specific pathogens. Plant
pathologists are asking why differ-
ent species, even closely related
ones, have different sets of defend-
ers. The answer could result in
hardier crops.
What is the basis of
variation in stress
tolerance in plants?
We need crops that better
withstand drought, cold, and
other stresses. But there are
so many genes involved, in
complex interactions, that
no one has yet figured out
which ones work how.
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005
J. MOGLIA/SCIENCE ARS
Published by AAAS
CR
EDIT
:RIM
MA
GER
LOV
INA
AN
D V
ALE
RIY
GER
LOV
IN “
TW
O E
GG
S”
© 2
00
2,C
-PR
INT.C
OU
RT
ESY
RO
BER
T B
RO
WN
GA
LLER
Y,W
ASH
ING
TON
,D.C
.
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org98
Sp
ec
ial
Se
cti
on
Quantum mechanics is very impres-sive,” Albert Einstein wrote in 1926.“But an inner voice tells me that it isnot yet the real thing.” As quantum
theory matured over the years, that voicehas gotten quieter—but it has not beensilenced. There is a relentless murmur ofconfusion underneath the chorus of praise forquantum theory.
Quantum theory was born at the very endof the 19th century and soon became one ofthe pillars of modern physics. It describes,with incredible precision, the bizarre andcounterintuitive behavior of the very small:atoms and electrons and other wee beastiesof the submicroscopic world. But that suc-cess came with the price of discomfort. Theequations of quantum mechanics work verywell; they just don’t seem to make sense.
No matter how you look at the equations ofquantum theory, they allow a tiny object tobehave in ways that defy intuition. For exam-ple, such an object can be in “superposition”:It can have two mutually exclusive propertiesat the same time. The mathematics of quan-tum theory says that an atom, for example, canbe on the left side of a box and the right sideof the box at the very same instant, as long asthe atom is undisturbed and unobserved. Butas soon as an observer opens the box and triesto spot where the atom is, the superpositioncollapses and the atom instantly “chooses”whether to be on the right or the left.
This idea is almost as unsettling today as itwas 80 years ago, when Erwin Schrödingerridiculed superposition by describing a half-living, half-dead cat. That is because quantumtheory changes what the meaning of “is” is. Inthe classical world, an object has a solid reality:Even a cloud of gas is well described by hardlittle billiard ball–like pieces, each of whichhas a well-defined position and velocity.Quantum theory seems to undermine that solid
reality. Indeed, the famous Uncertainty Princi-ple, which arises directly from the mathemat-ics of quantum theory, says that objects’ posi-tions and momenta are smeary and ill defined,and gaining knowledge about one implies los-ing knowledge about the other.
The early quantum physicists dealt withthis unreality by saying that the “is”—the fun-damental objects handled by the equations ofquantum theory—were not actually particlesthat had an extrinsic reality but “probability
waves” that merely had the capability ofbecoming “real” when an observer makes ameasurement. This so-called CopenhagenInterpretation makes sense, if you’re willing toaccept that reality is probability waves and notsolid objects. Even so, it still doesn’t suffi-ciently explain another weirdness of quantumtheory: nonlocality.
In 1935, Einstein came up with a scenariothat still defies common sense. In his thought
experiment, two particles fly away from eachother and wind up at opposite ends of thegalaxy. But the two particles happen to be“entangled”—linked in a quantum-mechanicalsense—so that one particle instantly “feels”what happens to its twin. Measure one, and theother is instantly transformed by that measure-ment as well; it’s as if the twins mysticallycommunicate, instantly, over vast regions ofspace. This “nonlocality” is a mathematicalconsequence of quantum theory and has beenmeasured in the lab. The spooky action appar-ently ignores distance and the flow of time; intheory, particles can be entangled after theirentanglement has already been measured.
On one level, the weirdness of quantumtheory isn’t a problem at all. The mathematicalframework is sound and describes all thesebizarre phenomena well. If we humans can’timagine a physical reality that corresponds toour equations, so what? That attitude has beencalled the “shut up and calculate” interpretationof quantum mechanics. But to others, our diffi-culties in wrapping our heads around quantumtheory hint at greater truths yet to be understood.
Some physicists in the second group arebusy trying to design experiments that can getto the heart of the weirdness of quantum theory.They are slowly testing what causes quantumsuperpositions to “collapse”—research thatmay gain insight into the role of measurementin quantum theory as well as into why bigobjects behave so differently from small ones.Others are looking for ways to test variousexplanations for the weirdnesses of quantumtheory, such as the “many worlds” interpreta-tion, which explains superposition, entangle-ment, and other quantum phenomena bypositing the existence of parallel universes.Through such efforts, scientists might hope toget beyond the discomfort that led Einstein todeclare that “[God] does not play dice.”
–CHARLESSEIFE
Do Deeper Principles Underlie Quantum Uncertainty and Nonlocality
W H A T D O N ’ T W E K N O W ?
What caused mass
extinctions?
A huge impact did in
the dinosaurs, but
the search for other
catastrophic triggers
of extinction has had
no luck so far. If more
subtle or stealthy cul-
prits are to blame,
they will take consid-
erably longer to find.
Can we prevent
extinction?
Finding cost-effective
and politically
feasible ways
to save many
endangered
species
requires cre-
ative thinking.
How will ecosystems
respond to global
warming?
To anticipate the effects
of the intensifying green-
house, climate modelers
will have to focus on
regional changes and
ecologists on the right combination
of environmental changes.
Why were some dinosaurs so large?
Dinosaurs reached almost unimaginable sizes,
some in less than
20 years. But how
did the long-necked
sauropods, for instance,
eat enough to pack on
up to 100 tons without
denuding their world?
FLO
RID
A K
EYS N
AT
ION
AL
MA
RIN
E SA
NC
TU
AR
Y/N
OA
A
LOUIE PSIHOYOS/CORBIS
‘‘
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 99
Sp
ec
ial
Se
cti
onI
n the 2 decades since researchers identi-fied HIV as the cause of AIDS, moremoney has been spent on the search for avaccine against the virus than on any vac-
cine effort in history. The U.S. National Insti-tutes of Health alone invests nearly $500 mil-lion each year, and more than 50 differentpreparations have entered clinical trials. Yetan effective AIDS vaccine, which potentiallycould thwart millions of new HIV infectionseach year, remains a distant dream.
Although AIDS researchers have turnedthe virus inside-out and carefully detailedhow it destroys the immune system, they haveyet to unravel which immune responses canfend off an infection. That means, as oneAIDS vaccine researcher famously put itmore than a decade ago, the field is “flyingwithout a compass.”
Some skeptics contend that no vaccine willever stop HIV. They argue that the virus repli-cates so quickly and makes so many mistakesduring the process that vaccines can’t possiblyfend off all the types of HIV that exist. HIValso has developed sophisticated mechanismsto dodge immune attack, shrouding its surfaceprotein in sugars to hide vulnerable sites fromantibodies and producing proteins that thwartproduction of other immune warriors. And theskeptics point out that vaccine developers havehad little success against pathogens like HIVthat routinely outwit the immune system—themalaria parasite, hepatitis C virus, and thetuberculosis bacillus are prime examples.
Yet AIDS vaccine researchers have solidreasons to believe they can succeed. Mon-key experiments have shown that vaccinescan protect animals from SIV, a simian rela-tive of HIV. Several studies have identifiedpeople who repeatedly expose themselves toHIV but remain uninfected, suggesting thatsomething is stopping the virus. A small per-centage of people who do become infectednever seem to suffer any harm, and othershold the virus at bay for a decade or morebefore showing damage to their immunesystems. Scientists also have found thatsome rare antibodies do work powerfullyagainst the virus in test tube experiments.
At the start, researchers pinned their hopeson vaccines designed to trigger production ofantibodies against HIV’s surface protein. Theapproach seemed promising because HIVuses the surface protein to latch onto white
blood cells and establish an infection. Butvaccines that only contained HIV’s surfaceprotein looked lackluster in animal and testtube studies, and then proved worthless inlarge-scale clinical trials.
Now, researchers are intensely investigat-ing other approaches. When HIV manages tothwart antibodies and establish an infection, asecond line of defense, cellular immunity,specif ically targets and eliminates HIV-infected cells. Several vaccines which are nowbeing tested aim to stimulate production ofkiller cells, the storm troopers of the cellularimmune system. But cellular immunityinvolves other players—such as macrophages,the network of chemical messengers called
cytokines, and so-called natural killer cells—that have received scant attention.
The hunt for an antibody-based vaccinealso is going through something of a renais-sance, although it’s requiring researchers tothink backward. Vaccine researchers typi-cally start with antigens—in this case,pieces of HIV—and then evaluate the anti-bodies they elicit. But now researchers haveisolated more than a dozen antibodies frominfected people that have blocked HIVinfection in test tube experiments. The trickwill be to figure out which specific antigenstriggered their production.
It could well be that a successful AIDSvaccine will need to stimulate both the pro-duction of antibodies and cellular immunity,a strategy many are attempting to exploit.Perhaps the key will be stimulating immu-nity at mucosal surfaces, where HIV typi-cally enters. It’s even possible thatresearchers will discover an immuneresponse that no one knowsabout today. Or perhaps the
answer lies in the interplay betweenthe immune system and human genetic vari-ability: Studies have highlighted genes thatstrongly influence who is most susceptible—and who is most resistant—to HIV infectionand disease.
Wherever the answer lies, the insightscould help in the development of vaccinesagainst other diseases that, like HIV, don’teasily succumb to immune attack and that killmillions of people. Vaccine developers forthese diseases will probably also have to lookin unusual places for answers. The maps cre-ated by AIDS vaccine researchers currentlyexploring uncharted immunologic terraincould prove invaluable. –JON COHEN
Is an Effective HIV
Vaccine Feasible
continued >>
CR
ED
IT:R
EU
TER
S/C
OR
BIS
W H A T D O N ’ T W E K N O W ?
How many kinds of humans
coexisted in the recent past,
and how did they relate?
The new dwarf human species fossil
from Indonesia suggests that at
least four kinds of
humans thrived in the
past 100,000
years. Better dates
and additional
material will help
confirm or revise
this picture.
What gave rise to modern
human behavior?
Did Homo sapiens acquire abstract
thought, language, and art grad-
ually or in a cultural “big bang,”
which in Europe occurred
about 40,000 years ago?
Data from Africa, where our
species arose, may hold the
key to the answer.
What are the roots of human culture?
No animal comes close to having humans’ ability
to build on previous
discoveries and pass the
improvements on.
What determines those
differences could help
us understand how
human culture evolved.
What are the evolu-
tionary roots of lan-
guage and music?
Neuroscientists
exploring how we
speak and make
music are just begin-
ning to find clues as
to how these prized
abilities arose.JUPITER IMAGESNSF
Published by AAAS
CR
EDIT
:PA
SC
AL
LE S
EGR
ETA
IN/G
ETT
Y IM
AG
ES
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org100
Sp
ec
ial
Se
cti
on
Scientists know that theworld has warmed lately,and they believe humankind
is behind most of that warming.But how far might we push theplanet in coming decades and cen-turies? That depends on just howsensitively the climate system—air, oceans, ice, land, and life—responds to the greenhouse gaseswe’re pumping into the atmos-phere. For a quarter-century,expert opinion was vague aboutclimate sensitivity. Expertsallowed that climate might bequite touchy, warming sharplywhen shoved by one climatedriver or another, such as the car-bon dioxide from fossil fuel burn-ing, volcanic debris, or dimming of the sun.On the other hand, the same experts concededthat climate might be relatively unresponsive,warming only modestly despite a hard pushtoward the warm side.
The problem with climate sensitivity isthat you can’t just go out and directly meas-ure it. Sooner or later a climate model mustenter the picture. Every model has its ownsensitivity, but each is subject to all theuncertainties inherent in building a hugelysimplified facsimile of the real-world cli-mate system. As a result, climate scientistshave long quoted the same vague range forsensitivity: A doubling of the greenhousegas carbon dioxide, which is expected tooccur this century, would eventually warmthe world between a modest 1.5°C and awhopping 4.5°C. This range—based on just
two early climate models—first appeared in1979 and has been quoted by every majorclimate assessment since.
Researchers are f inally beginning totighten up the range of possible sensitivities,at least at one end. For one, the sensitivities ofthe available models (5% to 95% confidencerange) are now falling within the canonicalrange of 1.5°C to 4.5°C; some had gone con-siderably beyond the high end. And the firsttry at a new approach—running a singlemodel while varying a number of modelparameters such as cloud behavior—has pro-duced a sensitivity range of 2.4°C to 5.4°Cwith a most probable value of 3.2°C.
Models are only models, however. Howmuch better if nature ran the experiment?Enter paleoclimatologists, who sort out howclimate drivers such as greenhouse gases have
varied naturally in the distant past and how theclimate system of the time responded. Nature,of course, has never run the perfect analog forthe coming greenhouse warming. And esti-mating how much carbon dioxide concentra-tions fell during the depths of the last ice age orhow much sunlight debris from the eruption ofMount Pinatubo in the Philippines blockedwill always have lingering uncertainties. Butpaleoclimate estimates of climate sensitivitygenerally fall in the canonical range, with abest estimate in the region of 3°C.
The lower end at least of likely climatesensitivity does seem to be firming up; it’snot likely below 1.5°C, say researchers. Thatwould rule out the negligible warmings pro-posed by some greenhouse contrarians. Butclimate sensitivity calculations still put afuzzy boundary on the high end. Studiesdrawing on the past century’s observed cli-mate change plus estimates of natural andanthropogenic climate drivers yield up to30% probabilities of sensitivities above4.5°C, ranging as high as 9°C. The lateststudy that varies model parameters allowssensitivities up to 11°C, with the authorscontending that they can’t yet say what thechances of such extremes are. Others arepointing to times of extreme warmth in thegeologic past that climate models fail toreplicate, suggesting that there’s a dangerouselement to the climate system that the mod-els do not yet contain.
Climate researchers have their work cut outfor them. They must inject a better understand-ing of clouds and aerosols—the biggestsources of uncertainty—into their modeling.Ten or 15 years ago, scientists said that wouldtake 10 or 15 years; there’s no sign of it hap-pening anytime soon. They must increase thefidelity of models, a realistic goal given thecontinued acceleration of affordable comput-ing power. And they must retrieve more andbetter records of past climate changes andtheir drivers. Meanwhile, unless a rapid shiftaway from fossil fuel use occurs worldwide, adoubling of carbon dioxide—and more—will
be inevitable.–RICHARDA.KERR
How Hot Will
The Greenhouse
World Be
A harbinger? Coffins being lined up during the record-breaking2003 heat wave in Europe.
W H A T D O N ’ T W E K N O W ?
What are human
races, and how did
they develop?
Anthropologists have
long argued that race
lacks biological real-
ity. But our genetic
makeup does vary
with geographic
origin and as such
raises political and
ethical as well as
scientific questions.
Why do some
countries grow and
others stagnate?
From Norway to Nige-
ria, living standards
across countries vary
enormously, and
they’re not becoming
more equal.
What impact do
large government
deficits have on a
country’s interest
rates and economic
growth rate?
The United States
could provide a
test case.
Why has poverty
increased and life
expectancy
declined in sub-
Saharan Africa?
Almost all efforts to
reduce poverty in
sub-Saharan Africa
have failed. Figuring
out what will work is
crucial to alleviating
massive human
suffering.
Are political and economic
freedom closely tied?
China may provide one answer.
LYLE CONRAD/CDCJUPITER IMAGES
Published by AAAS
www.sciencemag.org SCIENCE VOL 309 1 JULY 2005 101
CR
ED
IT:J
UPIT
ER
IM
AG
ES
Sp
ec
ial
Se
cti
onT
he road from old to new energysources can be bumpy, but the transi-tions have gone pretty smoothly inthe past. After millennia of depend-
ence on wood, society added coal and gravity-driven water to the energy mix. Industrializa-tion took off. Oil arrived, and transportation byland and air soared, with hardly a worry aboutwhere the next log or lump of coal was comingfrom, or what the explosive growth in energyproduction might be doing to the world.
Times have changed. The price of oil hasbeen climbing, and ice is melting around bothpoles as the mercury in the global thermometerrises. Whether the next big energy transitionwill be as smooth as past ones will depend inlarge part on three sets of questions: When willworld oil production peak? How sensitive isEarth’s climate to the carbon dioxide we arepouring into the atmosphere by burning fossilfuels? And will alternative energy sources beavailable at reasonable costs? The answers reston science and technology, but how societyresponds will be firmly in the realm of politics.
There is little disagreement that the worldwill soon be running short of oil. The debate isover how soon. Global demand for oil hasbeen rising at 1% or 2% each year, and we arenow sucking almost 1000 barrels of oil fromthe ground every second. Pessimists—mostlyformer oil company geologists—expect oilproduction to peak very soon. They point toAmerican geologist M. King Hubbert’s suc-cessful 1956 prediction of the 1970 peak inU.S. production. Using the same methodinvolving records of past production and dis-coveries, they predict a world oil peak by theend of the decade. Optimists—mostlyresource economists—argue that oil produc-tion depends more on economics and politicsthan on how much happens to be in theground. Technological innovation will inter-vene, and production will continue to rise,they say. Even so, midcentury is about as far asanyone is willing to push the peak. That’s still“soon” considering that the United States, forone, will need to begin replacing oil’s 40%contribution to its energy consumption bythen. And as concerns about climate change
intensify, the transition to nonfossil fuelscould become even more urgent (see p. 100).
If oil supplies do peak soon or climate con-cerns prompt a major shift away from fossilfuels, plenty of alternative energy supplies arewaiting in the wings. The sun bathes Earth’ssurface with 86,000 trillion watts, or terawatts,of energy at all times, about 6600 times theamount used by all humans on the planet eachyear. Wind, biomass, and nuclear power are alsoplentiful. And there is no shortage of opportu-nities for using energy more efficiently.
Of course, alternative energy sourceshave their issues. Nuclear fission supportershave never found a noncontroversial solutionfor disposing of long-lived radioactivewastes, and concerns over liability and capi-
tal costs are scaring utility companies off.Renewable energy sources are diffuse, mak-ing it diff icult and expensive to corralenough power from them at cheap prices. Sofar, wind is leading the way with a globalinstalled capacity of more than 40 billionwatts, or gigawatts, providing electricity forabout 4.5 cents per kilowatt hour.
That sounds good, but the scale of renew-able energy is still very small when comparedto fossil fuel use. In the United States, renew-ables account for just 6% of overall energy pro-duction. And, with global energy demandexpected to grow from approximately 13 tera-watts a year now to somewhere between30 and 60 terawatts by the middle ofthis century, use of renewables
will have to expand enormously to dis-place current sources and have a significantimpact on the world’s future energy needs.
What needs to happen for that to take place?Using energy more efficiently is likely to be thesine qua non of energy planning—not least tobuy time for efficiency improvements in alter-native energy. The cost of solar electric powermodules has already dropped two orders ofmagnitude over the last 30 years. And mostexperts figure the price needs to drop 100-foldagain before solar energy systems will bewidely adopted. Advances in nanotechnologymay help by providing novel semiconductorsystems to boost the efficiency of solar energycollectors and perhaps produce chemical fuelsdirectly from sunlight, CO2, and water.
But whether these will come in time toavoid an energy crunch depends in part on howhigh a priority we give energy research anddevelopment. And it will require a global polit-ical consensus on what the science is telling us.
–RICHARDA. KERR AND ROBERT F. SERVICE
continued >>
What Can ReplaceCheap Oil—and When
W H A T D O N ’ T W E K N O W ?
Is there a simple test for deter-
mining whether an elliptic
curve has an infinite number of
rational solutions?
Equations of the form y2 = x3 + ax
+ b are powerful mathematical
tools. The Birch and Swinnerton-
Dyer conjecture tells how to
determine how many solutions
they have in the realm of rational
numbers—information that could
solve a host of problems, if the
conjecture is true.
Can a Hodge cycle be written as
a sum of algebraic cycles?
Two useful mathematical struc-
tures arose independently in geom-
etry and in abstract algebra. The
Hodge conjecture posits a surpris-
ing link between them, but the
bridge remains to be built.
The following six mathematics questions are drawn
from a list of seven outstanding problems selected
by the Clay Mathematics Institute. (The seventh
problem is discussed on p. 96.) For more details, go to
www.claymath.org/millennium.
Published by AAAS
CR
EDIT
:VIV
IAN
E M
OO
S/C
OR
BIS
1 JULY 2005 VOL 309 SCIENCE www.sciencemag.org102
Sp
ec
ial
Se
cti
on
In 1798, a 32-year-old curate at a smallparish church in Albury, England, pub-lished a sobering pamphlet entitled An
Essay on the Principle of Population. As agrim rebuttal of the utopian philosophers ofhis day, Thomas Malthus argued that humanpopulations will always tend to grow and,eventually, they will always be checked—either by foresight, such as birth control, or asa result of famine,war, or disease. Thosespeculations haveinspired many a direwarning from envi-ronmentalists.
Since Malthus’stime, world popula-tion has risen sixfoldto more than 6 billion.Yet happily, apoca-lyptic collapses havemostly been pre-vented by the adventof cheap energy, therise of science andtechnology, and thegreen revolution.Most demographerspredict that by 2100,global population will level off at about10 billion.
The urgent question is whether currentstandards of living can be sustained whileimproving the plight of those in need.Consumption of resources—not just foodbut also water, fossil fuels, timber, andother essentials—has grown enormouslyin the developed world. In addition,humans have compounded the directthreats to those resources in many ways,including by changing climate (see p. 100),polluting land and water, and spreadinginvasive species.
How can humans live sustainably on theplanet and do so in a way that manages topreserve some biodiversity? Tackling thatquestion involves a broad range of research fornatural and social scientists. It’s abundantlyclear, for example, that humans are degradingmany ecosystems and hindering their abilityto provide clean water and other “goods andservices” (Science, 1 April, p. 41). But exactly
how bad is the situation? Researchers needbetter information on the status and trends ofwetlands, forests, and other areas. To set prior-ities, they’d also like a better understanding ofwhat makes ecosystems more resistant orvulnerable and whether stressed ecosystems,such as marine fisheries, have a threshold atwhich they won’t recover.
Agronomists face the task of feeding4 billion more mouths. Yields may be max-ing out in the developed world, but muchcan still be done in the developing world,particularly sub-Saharan Africa, which des-perately needs more nitrogen. Although
agricultural biotechnology clearlyhas potential to boost yields and
lessen the environmental impact offarming, it has its own risks, and winningover skeptics has proven difficult.
There’s no shortage of work for socialscientists either. Perverse subsidies that
encourage overuse of resources—taxloopholes for luxury Hummers and otherinefficient vehicles, for example—remain achronic problem. A new area of activity is theattempt to place values on ecosystems’ serv-ices, so that the price of clear-cut lumber, forinstance, covers the loss of a forest’s ability toprovide clean water. Incorporating those“externalities” into pricing is a daunting chal-lenge that demands much more knowledge ofecosystems. In addition, economic decisionsoften consider only net present value and dis-count the future value of resources—soil ero-sion, slash-and-burn agriculture, and the min-ing of groundwater for cities and farming areprime examples. All this complicates theprocess of transforming industries so thatthey provide jobs, goods, and services whiledamaging the environment less.
Researchers must also grapple with thechanging demographics of housing and howit will impact human well-being: In the next35 to 50 years, the number of people livingin cities will double. Much of the growthwill likely happen in the developing world incities that currently have 30,000 to 3 millionresidents. Coping with that huge urbaninflux will require everything from energy-efficient ways to make concrete to simpleways to purify drinking water.
And in an age of global television andrelentless advertising, what will happen topatterns of consumption? The world clearlycan’t support 10 billion people living likeAmericans do today. Whether science—both the natural and social sciences—andtechnology can crank up eff iciency andsolve the problems we’ve created is per-haps the most critical question the worldfaces. Mustering the political will to makehard choices is, however, likely to be aneven bigger challenge. –ERIK STOKSTAD
Will Malthus Continue to Be Wrong
Out of balance. Sustaining a growing world population is threatened byinefficient consumption of resources—and by poverty.
W H A T D O N ’ T W E K N O W ?
Will mathematicians unleash
the power of the Navier-Stokes
equations?
First written down in the 1840s,
the equations hold the keys to
understanding both smooth and
turbulent flow. To harness them,
though, theorists must find out
exactly when they work and under
what conditions they break down.
Does Poincaré’s test identify
spheres in four-dimensional
space?
You can tie a string around a
doughnut, but it will slide right
off a sphere. The mathematical
principle behind that observation
can reliably spot every spherelike
object in 3D space. Henri Poincaré
conjectured that it should also
work in the next dimension up,
but no one has proved it yet.
Do mathematically interesting
zero-value solutions of the Rie-
mann zeta function all have the
form a + bi?
Don’t sweat the details. Since the
mid-19th century, the “Riemann
hypothesis” has been the monster
catfish in mathematicians’ pond.
If true, it will give them a wealth of
information about the distribution
of prime numbers and other long-
standing mysteries.
Does the Standard Model of
particle physics rest on solid
mathematical foundations?
For almost 50 years, the model
has rested on “quantum Yang-Mills
theory,” which links the behavior
of particles to structures found
in geometry. The theory is breath-
takingly elegant and useful—but
no one has proved that it’s sound.
Published by AAAS