MARCH 2004 $4.95 WWW.SCIAM.COM THE SILENT EARTHQUAKE MENACE • A SCIENCE OF FAIR VOTING The Time Bomb of Global Warming (and How to Defuse It) On twin rovers explore baffling landscapes On robotic vehicles race across the Mojave Desert MARS, MARS, EARTH, EARTH, Robots on Two Worlds Robots on Two Worlds How Addiction Reshapes Brains Flu Vaccines’ Biotech Future ( see page 16 ) How Addiction Reshapes Brains Flu Vaccines’ Biotech Future ( see page 16 ) COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
Transcript
MARCH 2004 $4.95WWW.SCIAM.COM
THE SILENT EARTHQUAKE MENACE • A SCIENCE OF FAIR VOTING
The Time Bombof Global Warming
(and How to Defuse It)
On twin rovers explore baffling landscapes
On roboticvehicles race acrossthe Mojave Desert
MARS,MARS,
EARTH,EARTH,
Robots on Two WorldsRobots on Two Worlds
How AddictionReshapes Brains
Flu Vaccines’Biotech Future(see page 16)
How AddictionReshapes Brains
Flu Vaccines’Biotech Future(see page 16)
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
P L A N E T A R Y S C I E N C E
52 The Spirit of ExplorationB Y G E O R G E M U S S E RNASA’s robot rover scouts unknown terrain on the Angry Red Planet.
I N F O R M A T I O N T E C H N O L O G Y
58 A New Race of RobotsB Y W . W A Y T G I B B SThis month a grueling off-road race through theMojave Desert may crown the most capablerobotic vehicles ever. But for the engineers behindthe machines, the race started long ago.
C L I M A T O L O G Y
68 Defusing the Global Warming Time BombB Y J A M E S H A N S E NTroubling geologic evidence verifies that human activities are shifting the climate. But practical actions to clean up theatmosphere could slow the process.
B I O T E C H N O L O G Y
78 The Addicted BrainB Y E R I C J . N E S T L E R A N D R O B E R T C . M A L E N K ABetter understanding of how drug abuse produces long-term changes in the brain’s reward circuitry opens up new possibilities for treating addictions.
E A R T H S C I E N C E
86 The Threat of Silent EarthquakesB Y P E T E R C E R V E L L INot all earthquakes cause a noticeable rumbling. Recognizing the quiet types could be a tip-off to imminent devastating tsunamis and ground-shaking shocks.
E L E C T O R A L S Y S T E M S
92 The Fairest Vote of AllB Y P A R T H A D A S G U P T A A N D E R I C M A S K I NSurprisingly, in elections best designed to read voters’ wishes, the winner should not always be the candidate who gets the most votes.
SCIENTIFIC AMERICAN Volume 290 Number 3
52 Mars yieldsgrudgingly torobot probes
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 3
contentsmarch 2004
features
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
4 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
departments6 SA Perspectives
The lack of leadership on climate policy.
8 How to Contact Us8 On the Web
10 Letters15 50, 100 & 150 Years Ago16 News Scan
■ Hatching flu vaccines without chicken eggs.■ The next linear collider?■ Ways to spot sniper fire.■ Verifying the pre-Columbian Vinland map.■ Early warnings for solar storms.■ It slices, it dices! It’s the nitrogen knife.■ Data Points: Mad cow disease spreads. ■ By the Numbers: Rise of black ghettos.
41 Staking ClaimsA university mimics corporations in greedily gaming the patent system.
44 InnovationsNanotechnology brings chips one step closer to assembling themselves.
48 InsightsCan veteran pathogen fighter David L. Heymannrepeat his SARS-control success with polio?
98 Working KnowledgeThe crystalline workings of watches.
100 Voyages Free services help volunteers make their mark on archaeological and forestry research.
103 ReviewsThree new books by brain researchers tackle the hard problem of explaining consciousness.
42 Skeptic B Y M I C H A E L S H E R M E RThe gorilla in our midst: How beliefs shape what we see—and don’t see.
108 Puzzling Adventures B Y D E N N I S E . S H A S H ATraffic on the grid.
110 Anti Gravity B Y S T E V E M I R S K Y“Regulatory intrusion” may be why you’re not dead.
111 Ask the Experts Why doesn’t the body reject blood transfusions? How can deleted computer files be retrieved?
112 Fuzzy Logic B Y R O Z C H A S T
Cover image by Daniel Maas, Maas Digital LLC, NASA/JPL/Cornell University;preceding page: NASA /JPL/Malin Space Science Systems.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
If you still doubt that global warming is real and thathumans contribute to it, read the article beginning onpage 68. Its author, James Hansen of the NASA God-dard Institute for Space Studies, is no doomsayer. In-stead of relying on just computer climate models,which skeptics don’t trust, Hansen builds a powerfulcase for global warming based on the geologic recordand simple thermodynamics. He sees undeniable signs
of danger, especially from ris-ing ocean levels, but he alsobelieves that we can slow orhalt global warming afford-ably—if we start right away.
Politically, that’s the rub.As time slips by, our leverageover the problem melts away.Even small reductions in gasand aerosol emissions todayforestall considerable warm-ing and damage in the longrun. In our view, the interna-
tional community needs a leader, but the obvious nationfor the job still has its head in the sand.
President George W. Bush’s administration impliesthat it will get more serious about global warming af-ter further years of study determine the scope of theproblem (tick . . . tick . . . tick . . . ). The Kyoto Protocolis the most internationally acceptable approach to a so-lution yet devised. Largely at the insistence of Ameri-can negotiators, it adopts a market-based strategy.Nevertheless, the White House in 2001, like the U.S.Senate in 1997, rejected the treaty as economically ru-inous and environmentally inadequate. The adminis-tration has yet to propose a workable alternative.
Two years ago the president committed the coun-try to reducing its greenhouse gas “intensity”—theemissions per unit of economic output—by 18 percent
in 10 years. But he has not enunciated a clear and cred-ible strategy for doing even that. The White Houseboasts of the $4.3 billion budgeted for climate change–related programs in 2004 as well as its backing for hy-drogen-based energy. But those initiatives don’t set anygoals by which they can be judged. All they do is throwmoney at new technologies in the hope that business-es might eventually adopt them. In other areas of en-vironmental policy, the administration insists on cost-benefit analyses—but not for climate change policy.
A real action plan is feasible. Current technologycan stop the increase of soot emissions from dieselcombustion at a reasonable cost. Reductions in air-borne soot would boost the reflection of sunlight fromsnow back into space. Minimizing soot also directlybenefits human health and agricultural productivity.
Suitably controlling greenhouse gases is a greaterchallenge, but it can be done. Kyoto establishes a cap-and-trade program for carbon dioxide and other emis-sions. The administration has favored programs totrade credits for industrial pollutants such as mercury.Carbon dioxide is an even more appropriate subjectfor such an effort: creating environmental mercury“hot spots” raises local health risks, but concentratingcarbon dioxide production is harmless.
The expense of reducing carbon dioxide could bekept low by letting the marketplace identify cost-ef-fective ways to meet targets. Domestic emissions trad-ing for sulfur dioxide under the first Bush administra-tion was highly successful. Output levels were cutahead of schedule and at half the expected cost.
The only significant U.S. activity in carbon dioxidetrading now is at the state level. Ten northeastern stateshave established a regional initiative to explore such amarket. Meanwhile the administration sits on the side-lines. That’s not good enough: it needs to show spe-cific, decisive, meaningful leadership today.
6 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
P E R M I S S I O N SFor permission to copy or reuse
material from SA:www.sciam.com/permissions
or 212-451-8546 for procedures
or Permissions Department
Scientific American415 Madison Ave.
New York, NY 10017-1111 Please allow three to six weeks
for processing.
A D V E R T I S I N Gwww.sciam.com has electronic contactinformation for sales representatives of Scientific American in all regions ofthe U.S. and in other countries.
New YorkScientific American415 Madison Ave.New York, NY 10017-1111212-451-8893fax: 212-754-1138Los Angeles310-234-2699fax: 310-234-2670San Francisco415-403-9030fax: 415-403-9033DetroitKaren Teegarden & Associates248-642-1773fax: 248-642-6138MidwestDerr Media Group847-615-1921fax: 847-735-1457 SouthwestPublicitas North America, Inc.972-386-6186fax: 972-233-9819Direct ResponseA & A Media Options, LLC203-267-4251fax: 203-267-1552BelgiumPublicitas Media S.A.+32-(0)2-639-8420fax: +32-(0)2-639-8430CanadaDerr Media Group847-615-1921fax: 847-735-1457 France and SwitzerlandPEM-PEMA+33-1-46-37-2117fax: +33-1-47-38-6329GermanyPublicitas Germany GmbH+49-211-862-092-0fax: +49-211-862-092-21Hong KongHutton Media Limited+852-2528-9135fax: +852-2528-9281IndiaYogesh Rao Convergence Media +91-22-2414-4808 fax: +91-22-2414-5594JapanPacific Business, Inc.+813-3661-6138fax: +813-3661-6139KoreaBiscom, Inc.+822-739-7840fax: +822-732-3662Middle East Peter Smith Media & Marketing+44-140-484-1321fax: +44-140-484-1320SwedenPublicitas Nordic AB+46-8-442-7050fax: +46-8-442-7059U.K.The Powers Turner Group+44-207-592-8331fax: +44-207-630-9922
On the WebWWW.SCIENTIFICAMERICAN.COM
FEATURED THIS MONTHVisit www.sciam.com/onthewebto find these recent additions to the site:
Stardust SpaceProbe Flies by CometThe first mission to collect asample from a body beyond themoon and return it to Earth,the Stardust space probe hassuccessfully made its closeapproach to Comet Wild-2.During the encounter, Stardustdeployed a dust collectorroughly the size and shape of a
large-head tennis racket. The gathered dust, ranging in sizefrom a few to a few hundred microns, is thought to be apiece of the swirling cloud from which the planets emerged.
Unmaking Memories: Interview with James McGaughIn the recent sci-fi movie Paycheck, a crack reverseengineer helps companies to steal and improve on thetechnology of their rivals and then has his memory of thetime he spent working for them erased. The plot, based onPhilip K. Dick’s short story of the same name, is set in thenear future, but such selective memory erasure is still highlyspeculative at best. ScientificAmerican.com asked neuro-biologist James McGaugh of the University of California atIrvine, who studies learning and memory, to talk aboutwhat kinds of memory erasure are currently possible.
Ask the ExpertsWhy do people snore?Lynn A. D’Andrea, a sleep specialist at the University of Michigan Medical School, explains.
SIGN UP NOW AND GET INSTANT ONLINE ACCESS TO SCIENTIFIC AMERICAN DIGITALwww.sciamdigital.comMORE THAN JUST A DIGITAL MAGAZINE!
Three great reasons to subscribe:Get it first—all new monthly issues before they reach the newsstand.10+ years of Scientific American—more than 140 issues,from 1993 to the present.Enhanced search—easy, fast, convenient.
SUBSCRIBE TODAY AND SAVE! CO
UR
TESY
OF
NAS
A/JP
L-C
ALTE
CH
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
GENOME REVIEWIn “The Unseen Genome,” W. Wayt Gibbsdeplores the dogmatism that led biolo-gists to write off large parts of the genomeas junk and prevented them from recog-nizing several processes that may play animportant role in heredity. I want to sug-gest a different perspective: This narrowfocus by the research community led todetailed discoveries that have, in turn,challenged the guiding dogma and doneso in a relatively short time on the scale ofhuman history.
Closely constrained communal re-search may be a more effective long-termmeans of pursuing knowledge than re-search in which resources are continual-ly diverted to following up any apparentlead. The idea that tightly organized re-search leads (despite itself) to the recog-nition of anomalies that generate new ap-proaches was one of the themes of ThomasS. Kuhn’s The Structure of Scientific Rev-olutions. This theme was largely forgot-ten by those who read Kuhn as attackingscience, whether their aim was to defendscience or join in the supposed attack.
Harold I. BrownDepartment of Philosophy
Northern Illinois University
After reading “The Unseen Genome,” wewere surprised and disappointed that theauthor gave all credit for the discovery ofriboswitches to Ronald R. Breaker’s lab.We made this finding independently ofBreaker; our paper in Cell describing tworiboswitch families at once was publishedat the same time as the Breaker group’s
(“Sensing Small Molecules by NascentRNA,” by Mironov et al. in Cell, Vol.111, No. 5, pages 747–756; November27, 2002). Moreover, Gibbs refers toBreaker’s August 2003 paper reportingthat one family of riboswitches regulatesthe expression of no fewer than 26 genes.Our paper describing that same family ofriboswitches ran several months earlier(“The Riboswitch-Mediated Control ofSulfur Metabolism in Bacteria,” by Ep-shtein et al. in PNAS USA, Vol. 100, No.9, pages 5052–5056; April 29, 2003).
Evgeny NudlerDepartment of Biochemistry
New York University School of Medicine
SOLAR SOLUTIONS“The Asteroid Tugboat,” by Russell L.Schweickart, Edward T. Lu, Piet Hut andClark R. Chapman, discussed using larg-er launch vehicles and possibly nuclearpush mechanisms to deflect threateningasteroids into unthreatening orbits. Theseideas unnerved my sense of simplicity. Af-ter reading Philip Yam’s story about so-lar sails [“Light Sails to Orbit,” NewsScan], I wonder if painting the asteroid sil-ver would turn the whole spinning nuggetinto a “solar sail” opposed to the sun andif this method would alter the orbit.Would the solar wind be enough to pushsuch a painted asteroid away?
David T. Hanawaltvia e-mail
SCHWEICKART AND CHAPMAN REPLY: A sim-ilar proposal was raised by J. N. Spitale in theApril 5, 2002, issue of Science (Vol. 295, page
10 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
SCIENCE IS A PROJECT in a constant state of revision. The-ories are tweaked, probabilities adjusted, limits pushed, ele-ments added, maps redrawn. And every once in a while, awhole chapter gets a rewrite. In the November 2003 issue ofScientific American, “The Unseen Genome,” by W. Wayt Gibbs,reviewed one such change currently under way in genetics asnew research challenges the long-respected central dogma.In the field of space technology, “The Asteroid Tugboat,” byRussell L. Schweickart, Edward T. Lu, Piet Hut and Clark R.Chapman, posited a new way to divert unpredictable Earth-bound asteroids. Reader reactions to these and other innova-tive ideas from the issue follow.
LettersE D I T O R S @ S C I A M . C O M
E D I T O R I N C H I E F : John Rennie E X E C U T I V E E D I T O R : Mariette DiChristina M A N A G I N G E D I T O R : Ricki L. Rusting N E W S E D I T O R : Philip M. Yam S P E C I A L P R O J E C T S E D I T O R : Gary Stix S E N I O R E D I T O R : Michelle Press S E N I O R W R I T E R : W. Wayt Gibbs E D I T O R S : Mark Alpert, Steven Ashley, Graham P. Collins, Steve Mirsky, George Musser, Christine Soares C O N T R I B U T I N G E D I T O R S : Mark Fischetti, Marguerite Holloway, Philip E. Ross, Michael Shermer, Sarah Simpson, Carol Ezzell Webb
E D I T O R I A L D I R E C T O R , O N L I N E : Kate Wong A S S O C I A T E E D I T O R , O N L I N E : Sarah Graham
A R T D I R E C T O R : Edward Bell S E N I O R A S S O C I A T E A R T D I R E C T O R : Jana Brenning A S S O C I A T E A R T D I R E C T O R : Mark Clemens A S S I S T A N T A R T D I R E C T O R : Johnny JohnsonP H O T O G R A P H Y E D I T O R : Bridget Gerety P R O D U C T I O N E D I T O R : Richard Hunt
C O P Y D I R E C T O R : Maria-Christina Keller C O P Y C H I E F : Molly K. Frances C O P Y A N D R E S E A R C H : Daniel C. Schlenoff, Rina Bander, Emily Harrison, Michael Battaglia
E D I T O R I A L A D M I N I S T R A T O R : Jacob Lasky S E N I O R S E C R E T A R Y : Maya Harty
A S S O C I A T E P U B L I S H E R , P R O D U C T I O N : William Sherman M A N U F A C T U R I N G M A N A G E R : Janet Cermak A D V E R T I S I N G P R O D U C T I O N M A N A G E R : Carl Cherebin P R E P R E S S A N D Q U A L I T Y M A N A G E R : Silvia Di Placido P R I N T P R O D U C T I O N M A N A G E R : Georgina Franco P R O D U C T I O N M A N A G E R : Christina Hippeli C U S T O M P U B L I S H I N G M A N A G E R : Madelyn Keyes-Milch
A S S O C I A T E P U B L I S H E R / V I C E P R E S I D E N T , C I R C U L A T I O N :Lorraine Leib Terlecki C I R C U L A T I O N D I R E C T O R : Katherine Corvino F U L F I L L M E N T A N D D I S T R I B U T I O N M A N A G E R : Rosa Davis
V I C E P R E S I D E N T A N D P U B L I S H E R : Bruce Brandfon A S S O C I A T E P U B L I S H E R : Gail Delott W E S T E R N S A L E S M A N A G E R : Debra Silver S A L E S D E V E L O P M E N T M A N A G E R : David Tirpack WESTERN SALES DEVELOPMENT MANAGER: Valerie Bantner SALES REPRESENTATIVES: Stephen Dudley, Hunter Millington, Stan Schmidt
ASSOCIATE PUBLISHER, STRATEGIC PLANNING: Laura Salant P R O M O T I O N M A N A G E R : Diane Schube R E S E A R C H M A N A G E R : Aida Dadurian P R O M O T I O N D E S I G N M A N A G E R : Nancy Mongelli G E N E R A L M A N A G E R : Michael Florek B U S I N E S S M A N A G E R : Marie Maher M A N A G E R , A D V E R T I S I N G A C C O U N T I N G A N D C O O R D I N A T I O N : Constance Holmes
D I R E C T O R , S P E C I A L P R O J E C T S : Barth David Schwartz
M A N A G I N G D I R E C T O R , O N L I N E : Mina C. Lux S A L E S R E P R E S E N T A T I V E , O N L I N E : Gary BronsonW E B D E S I G N M A N A G E R : Ryan Reid
D I R E C T O R , A N C I L L A R Y P R O D U C T S : Diane McGarvey P E R M I S S I O N S M A N A G E R : Linda Hertz M A N A G E R O F C U S T O M P U B L I S H I N G : Jeremy A. Abbate
C H A I R M A N E M E R I T U S : John J. Hanley C H A I R M A N : Rolf Grisebach P R E S I D E N T A N D C H I E F E X E C U T I V E O F F I C E R :Gretchen G. Teichgraeber V I C E P R E S I D E N T A N D M A N A G I N G D I R E C T O R ,I N T E R N A T I O N A L : Dean SandersonV I C E P R E S I D E N T : Frances Newburg
Established 1845
®
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
77). Spitale’s proposal calls on the potential-ly more powerful Yarkovsky effect, in whichemission of thermal photons changes an as-teroid’s momentum, rather than pressurefrom the solar wind (light pressure), but it isroughly the same idea. Recent and relevantinformation about the Yarkovsky effect is online at http://neo.jpl.nasa.gov/news/news141.html. There are practical problems withpainting a whole asteroid, and no design hasbeen looked at seriously yet. Attaching an ac-tual, separate and necessarily large solar sailto an asteroid has also been proposed but like-wise presents serious engineering challenges.
ASTRO LOTTOWhen reflecting on the odds estimate pre-sented in “Penny-Wise, Planet-Foolish”[SA Perspectives]—“every year Earth hasa one-in-600,000 chance of gettingwhacked by an asteroid wider than onekilometer”—I found the lottery ticket inmy hand to be quite disconcerting. To har-vest the $160-million bounty on my tick-et, I would have to beat the winning oddsof 1:120,526,770, yet I’m willing to in-vest. While looking over the odds assignedto the remaining prizes, I find I have a sim-ilar chance of winning the $5,000 as per-ishing in the wake of an asteroid this year.Thanks for making me aware, I think.
Nicholas KulkeMadison, Wis.
CALL FOR BETTER BAFFLERS“Baffling the Bots,” by Lee Bruno [Inno-vations], left one important questionunanswered: How do Web visitors withvisual impairments use a service that isguarded with such visual trickery? Websites that use CAPTCHAs (for “complete-ly automated public Turing test to tellcomputers and humans apart”) and simi-lar barriers to bots need to provide alter-native access paths for users who are noless human for being visually impaired!
Carl ZetieWaterford, Va.
SOLAR-SAIL SUPPORT“Light Sails to Orbit,” by Philip Yam[News Scan], correctly described the
emerging interest in solar-sail technologyin the aerospace community but incor-rectly leaves the impression that NASA isunwilling to support solar-sail develop-ment efforts in the private sector. Further,the article’s claim that the Cosmos 1 mis-sion is the “lone player” in the private de-velopment of solar sails for spaceflight isalso incorrect.
Since 1999 Team Encounter has beendeveloping a series of privately financedsolar-sail missions. Our sailcraft technol-ogy, developed with our partner L’Garde,represents a significantly different ap-proach from that of Cosmos 1 and hasbeen well received and supported byNASA as well as the National Oceanic andAtmospheric Administration.
Charles M. ChaferPresident, Team Encounter
Houston
YAM RESPONDS: Certainly many groupsaround the world are committed to solar sail-ing besides the Cosmos 1 team. The Germanspace agency, for instance, is close to a testlaunch. And, as I noted in the story, NASAspends millions every year researching suchadvanced propulsion systems. I also wrotethat NASA chose to be a bystander in the Cos-mos 1 flight, not in solar-sail technology as awhole. Indeed, I described the kinds of goalsNASA seeks in a test flight. Such goals are notpart of the Cosmos 1 flight, which is meant todemonstrate feasibility and helps to explain
w w w . s c i a m . c o m
Letters
w w w . s c i a m . c o m
CO
UR
TESY
OF
THE
PLA
NE
TAR
Y SO
CIE
TY A
ND
CO
SMO
S ST
UD
IOS;
IM
AGE
PR
EP
ARE
D B
Y B
ABAK
IN S
PAC
E C
EN
TER
COSMOS 1 is one of the many team efforts to harness the solar winds.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
why NASA is not participating. Future testflights of more complex sail designs by TeamEncounter and other groups would do much topush solar-sail technology forward.
Thomas Gold’s assertion, noted in themarginalia of “Light Sails to Orbit,” thatthe solar sail cannot work because “per-fect mirrors do not create temperature dif-ferences, which are necessary to convertheat into kinetic energy,” is false, becausethe force results from radiation pressure,not heat. Radiation pressure, given by thepower flux divided by the speed of light,follows from 19th-century physics, specif-ically electrodynamics. The existence ofthis force was verified at least as early as1901 using a torsional balance and hasbeen used recently to manipulate smallobjects. The solar-sail concept is on firmtheoretical and experimental ground.
Thomas G. MoranNASA Goddard Space Flight Center
TWO TAKES ON TELLERAs a longtime reader of your magazine, Iwas appalled at the bad taste of GaryStix’s obituary of Edward Teller [NewsScan]. Contrary to Isidor Rabi’s ill-tem-pered political opinion, Teller’s contribu-tions were significant in keeping the Sovi-et threat in check and preserving the free-doms of the West.
Georgette P. ZoltaniLutherville, Md.
I find it hard to believe that Stix defendedTeller, stating that Isidor Rabi’s commentthat the world would have been a betterplace without Teller was “unquestionablyharsh.” I might also add that most of theimportant breakthroughs regarding thehydrogen bomb were the result of Stanis-law Ulam’s work and brains, not Teller’s.
Joseph Michael CierniakGlen Burnie, Md.
ERRATUM In “The Unseen Genome,” by W.Wayt Gibbs, the statement that riboswitcheshave been extracted from species “in all threekingdoms of life” should have read “in all threedomains of life.”
14 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
Letters
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
MARCH 1954CRUNCH, BANG—“A theory which sug-gests that our Universe started from anextremely compressed concentration ofmatter and radiation naturally raises thequestion: How did it get into that state?Relativistic formulae tell us that variousparts of the Universe are flyingapart with an energy exceedingthe forces of Newtonian attrac-tion between them. Extrapolat-ing these formulae to the periodbefore the Universe reached thestage of maximum contraction,we find that the Universe mustthen have been collapsing, withjust as great speed as it is now ex-panding! Thus, we conclude thatour Universe has existed for aneternity of time; that until aboutfive billion years ago it was col-lapsing uniformly from a state ofinfinite rarefaction; and that theUniverse is currently on the re-bound, dispersing irreversibly toward a state of infinite rarefac-tion. —George Gamow”
MARCH 1904DARWIN’S ATOLL—“Darwin hadearnestly desired a fuller exami-nation of coral reefs, in situ, andin fact went so far as to expresshis conviction (in a letter to Agas-siz in 1881) that nothing reallysatisfactory could be brought for-ward as contributory evidenceon their origin until a boring wasmade in one of the Pacific or Indianatolls, and a core obtained down to adepth of at least 500 feet. That hoped-forconsummation has, however, been over-achieved, since the boring of Funafutiwas carried down to a limit of 1,114 feet,during the third expedition to this ring-shaped spot of land in the South Pacific.The evidence derived goes to show thatthe material appears to be entirely of or-ganic character, traceable to the calcare-
ous skeletons of marine invertebrate an-imals and calcareous algae.”
ABRUZZI IN THE ARCTIC—“Great interestattaches to the polar expeditions of HisRoyal Highness Luigi Amedeo of Savoy,Duke of the Abruzzi. The ‘Polar Star’ was
to sail as far to the north as possiblealong some coast line, and then a partywas to travel on sledges toward the pole.The pole was not reached, but a latitudewas reached which no man had previ-ously attained, and it was proved thatwith determination and sturdy men anda number of well-selected dogs, the frozenArctic Ocean can actually be crossed tothe highest latitude. However, at the Em-peror Franz Josef archipelago, the ice
field trapped and threatened to sink theboat. Therefore, the crew were obliged toland with the utmost haste the stores forwinter [see illustration], and to secure thenecessary materials for building a dwell-ing. A retreat was carried out in the fol-lowing spring.”
MARCH 1854A FARADAY LECTURE—“The open-ing lecture of the Royal Institu-tion of London was delivered byMichael Faraday to a very crowd-ed audience. The subject was thedevelopment of electrical princi-ples produced by the working ofthe electric telegraph. To illus-trate the subject, there was an ex-tensive apparatus of voltaic bat-teries, consisting of 450 pairs ofplates, and eight miles of wirecovered with gutta-percha, fourmiles of which were immersed intubs of water. The principal pointwhich Professor Faraday wasanxious to illustrate was the con-firmation—which experiments onthe large scale of the electric tele-graph have afforded—of the iden-tity of dynamic or voltaic elec-tricity with static or frictionalelectricity.”
DINO DINER—“Professor RichardOwen was recently entertainedat dinner in the garden of theCrystal Palace at Sydenham, inthe model of an Iguanadon. The
animal in whose mould the dinner wasgiven was one of the former inhabitantsof Sussex, several of his bones havingbeen found near Horsham. His dimen-sions have been kept strictly within thelimits of anatomical knowledge. Thelength from the snout to the end of thetail was 35 feet. Twenty-one gentlemendined comfortably within the interior ofthe creature, and Professor Owen sat inhis head as substitute for brains.”
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 15
Gamow ■ Darwin ■ Faraday
50, 100 & 150 Years AgoFROM SCIENTIFIC AMERICAN
POLAR STAR trapped in the ice, Arctic Ocean, 1904
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
16 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
BE
TTM
ANN
/CO
RB
IS
I f you want to make an omelet, you have tobreak some eggs. And if you want to sup-ply the U.S. with flu vaccine, you have to
break about 100 million.That may change someday, as leading vac-
cine manufacturers explore the possibility oftrading their chicken eggs forstainless-steel culture vats andgrowing their flu virus in celllines derived from humans,monkeys or dogs. The tech-nology could allow compa-nies to produce their vaccinesin a more timely and less la-borious manner and to re-spond more quickly in anemergency.
Today’s flu vaccines areprepared in fertilized chickeneggs, a method developedmore than 50 years ago. Theeggshell is cracked, and theinfluenza virus is injected intothe fluid surrounding the em-bryo. The egg is resealed, theembryo becomes infected,and the resulting virus isthen harvested, purified andused to produce the vaccine.Even with robotic assistance,“working with eggs is te-dious,” says Samuel L. Katz
of the Duke University School of Medicine,a member of the vaccine advisory committeefor the U.S. Food and Drug Administration.“Opening a culture flask is a heck of a lotsimpler.”
Better yet, using cells could shave weeksoff the production process, notes Dinko Va-lerio, president and CEO of Crucell, a Dutchbiotechnology company developing one of thehuman cell lines. Now when a new strain offlu is discovered, researchers often need totinker with the virus to get it to reproduce inchicken eggs. Makers using cultured cellscould save time by skipping that step, per-haps even starting directly from the circulat-ing virus isolated from humans. As an addedbonus, the virus harvested from cells ratherthan eggs might even look more like the virusencountered by humans, making it betterfodder for a vaccine, adds Michel DeWilde,executive vice president of R&D at Aventis,the world’s largest producer of flu vaccinesand a partner with Crucell in developing flushots made from human cells.
Whether vaccines churned out by barrelsof cells will be any better than those producedin eggs “remains to be seen,” says the FDA’sRoland A. Levandowski. And for a persongetting jabbed in the arm during a regular fluseason, observes Richard Webby, a virologistat St. Jude Children’s Research Hospital inMemphis, Tenn., “it’s not going to matter
VAC
CIN
ES
Egg BeatersFLU VACCINE MAKERS LOOK BEYOND THE CHICKEN EGG BY K AREN HOPKIN
SCANnews
OVER EASY? Researchers hope to replace the decades-oldway of making flu vaccines, which involves injectingviruses into fertilized eggs pierced with a drill.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
18 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
GR
EG
CAM
PB
ELL
AP
Ph
oto
newsSCAN
H igh-energy physicists have a newmachine in mind: an unprecedented ac-celerator 30 kilometers long that would
offer a precise tool to explore some of themost important unanswered questions inphysics. But the specter of the defunct Super-conducting Supercollider—and the moneythe project ended up wasting—looms large.Advocates of the machine, however, think
they can overcome national doubts by goingglobal.
Since they first began discussing a linearcollider in earnest at a 2001 conference atSnowmass, Colo., the world’s physicists haveconsistently and vigorously planned an inter-national effort. Their hopes recently rose whenU.S. Secretary of Energy Spencer Abrahamnamed it the highest “midterm” priority in a
where the vaccine came from.”Where the cell-based vac-
cine will become invaluable,Webby states, is in the case of aglobal pandemic. Should a newstrain of flu crop up outside thenormal season—one that is dif-ferent enough from previousstrains that people will have noimmunity—cell-based systemswill allow health officials to re-spond more rapidly. “Cell cul-tures are a lot easier to scale upfaster,” he explains. Techni-cians would simply remove cellsfrom a freezer and grow them inlarge volumes—something thatis not possible with chickeneggs. Although flocks of chick-ens kept in clean environmentsare available almost year-round, companiesgenerally place their egg orders six months be-fore they start vaccine production. And pre-venting a pandemic could require 10 times asmuch vaccine as a normal flu season. “Ifhalfway into manufacturing, you need a bil-lion more eggs, you’re not going to get them,”remarks Wayne Morges, a vice president atBaxter in Deerfield, Ill.
Preparing vaccines in cell cultures is notnew. Aventis, for example, currently pro-duces polio vaccines in the same monkeykidney cells that Baxter is gearing up to useto produce flu injections. And Baxter usedthe monkey cell line to replenish the U.S.supply of smallpox vaccine. So converting tocell-based systems, Katz says, would be“moving flu vaccine production into the
20th century at the beginning of the 21st.”Why has it has taken manufacturers so
long to come around to considering cell-based systems? Perhaps because current egg-based systems work so well, Webby surmis-es. Up-front costs for preparing productionplants to function with cells rather than eggsmight also be an impediment.
Clinical trials of cell-based flu vaccineswon’t begin in the U.S. until this fall, and ifapproved, the new vaccines will at first prob-ably just supplement those produced inchicken eggs. Having several different for-mulations of flu vaccine can’t hurt. Exceptmaybe for that muscle soreness that lingersfor a day or two after you roll up your sleeve.
Karen Hopkin is based in Somerville, Mass.
Dream MachineHOPES FOR A GIANT COLLIDER LIE IN A WORLDWIDE APPEAL BY DAVID APPELL
PH
YSIC
S
For the Northern Hemisphere, theflu season typically runs from
November through March. Based oncollected virus samples and
infection activity, the World HealthOrganization decides which
influenza strains to include in avaccine in mid-February. By
mid-March, high-growth strains ofvaccine virus are provided to
manufacturers, and the materialsneeded to test the identity and
potency of the resulting vaccineare supplied in mid-May.
Vaccines become available inclinics in October.
Number of U.S. flu cases perseason: 29 million to 58 million
Number of Americans hospitalizedper season: 114,000
Number of deaths: 36,000
Number of vaccine doses producedthis season: 87.1 million
VIRAL TIMETABLES ALTERNATIVE MEDICINE: Researchers, including Richard Webby, a
virologist at St. Jude Children’s Research Hospital in Memphis, Tenn.,hope to speed influenza vaccine manufacturing by coming up with newoptions to the chicken egg as a virus growth medium.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
20 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
LIN
DA
A. C
ICE
RO
Sta
nfo
rd N
ews
Serv
ice
newsSCAN 20-year outlook of new science facilities. The
report estimates that were the project to be ap-proved and funded, peak spending would oc-cur sometime between 2010 and 2015.
The vision is of one machine built by theworld and shared by the world. “Many peo-ple have been working very hard to make this
more than an empty slogan,” says theoristChris Quigg of the Fermi National Accelera-tor Laboratory in Batavia, Ill., because noone government seems likely to spend the es-timated $5 billion to $7 billion that such a fa-cility would cost.
The plan is to accelerate electrons andpositrons (the antimatter version of the elec-tron) down dual 15-kilometer pipes andsmash them together inside a large detector.The total energy would be up to one trillionelectron volts (TeV). This energy may appearmuch less than the 2-TeV Tevatron at Fermi-lab and the 14-TeV Large Hadron Collider tobe completed at CERN in 2007, but because
the particles in those machines share their en-ergy among their constituent quarks, their ef-fective energy drops by about a factor of 10.By design, the international linear collider willhave higher interaction rates, and because thespins of the particles in its beams are aligned—
something that cannot be done at the Teva-tron or Large Hadron Collider—it will bemuch more precise in dissecting and analyz-ing particle interactions.
The collider could reveal the specifics ofHiggs bosons (particles that imbue all otherparticles with mass) and light supersymmetricparticles (shadowy particles such as the neu-tralino, which may account for the dark mat-ter that constitutes 23 percent of the universe).That knowledge could in turn open the doorto exotica such as extra dimensions and low-energy superstring phenomena. “That’s theexciting thing about the linear collider,” saystheorist Joseph Lykken of Fermilab. “It givesyou a window into this whole other realm ofphysics that we’re really interested in.”
But opening that window requires cold,hard cash. The last time particle physicistsasked for dollars for an accelerator, two bil-lion of them ended up underneath the Texasprairie in now water-filled tunnels meant forthe Superconducting Supercollider. “The sto-ry of its demise is so complicated, it’s fair tosay it died of fluctuations,” Quigg remarks.“Our community hopes to have learned fromthe experience to organize future projects sothey will be less vulnerable to fluctuationsand political tussles.”
In fact, several groups in the U.S., Europeand Japan are committed to the linear collid-er. “We are all behind it,” states AlbrechtWagner, director of the DESY high-energylaboratory in Hamburg, Germany, acknowl-edging that in the end the project’s site will bea political decision, not unlike that now beingmade about the fusion reactor called ITER.
So far the early politics involve technolo-gy recommendations. To accelerate particles,DESY backs a superconducting, lower-radio-frequency cavity; a higher-frequency, room-temperature structure is being championed bya collaboration between the Stanford LinearAccelerator and the KEK Accelerator Labo-ratory in Tsukuba, Japan. Given the historyof grand accelerators, deciding on which ap-proach to take will no doubt be the easy part.
David Appell is based in Lee, N.H.
A linear collider came in at 13th on a list of 28 future science
facilities, behind the internationalfusion reactor project ITER (first),
and the UltraScale ScientificComputing Capability (second),
which aims to increase scientificcomputing capacity 100-fold. Four
projects tied for third: the JointDark Energy Mission; an intense
x-ray laser called the LinacCoherent Light Source; a facility to
mass-produce, characterize andtag tens of thousands of proteins;and the Rare Isotope Accelerator.
Notably, the linear collider rankedahead of several other competing
physics projects, such as asuperneutrino beam and upgrades
to Brookhaven NationalLaboratory’s Relativistic Heavy Ion
Collider. The entire list is atwww.er.doe.gov/Sub/Facilities–for–future/facilities–future.htm
PHYSICSWISH LIST
DOWN THE LINE: The 3.2-kilometer-long tunnels of the Stanford Linear Accelerator Collider would bedwarfed by the proposed International Linear Collider,which would be five times as long.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
22 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
ALI
HAI
DE
R E
PA/
AP P
hot
o
newsSCAN
U .S. soldiers in Iraq face a bewilderingarray of threats. Since American andBritish troops occupied the country last
spring, Iraqi insurgents have downed heli-copters with heat-seeking missiles, detonatedroadside bombs along the routes of armyconvoys and launched mortar rounds at U.S.
bases. One of the biggest frustrations is theelusiveness of the enemy: the insurgents typ-ically slip away before American forces canrespond to an attack.
Now the Pentagon’s R&D arm, the De-fense Advanced Research Projects Agency(DARPA), is trying to provide some high-techassistance. The agency is pushing to deployexperimental systems that could quickly lo-cate the positions of enemy snipers and mor-tar crews. One of the most startling examplesis a ground-based carbon dioxide laser de-signed to pinpoint a sniper by measuring themovements of dust particles in the air causedby the shock wave of a speeding bullet. DARPA
director Anthony J. Tether announced last fallthat the anti-sniper laser, which would re-portedly have a range in the tens of kilome-ters, would be sent to Iraq early this year.
Developed by Mission Research Corpo-ration, a defense contractor based in SantaBarbara, Calif., the system relies on a Dopplerlidar, a laser radar that can measure the ve-locity of moving objects in much the sameway that a radar gun gauges the speed of carson the highway. Because the wavelength ofthe laser light is roughly comparable to the di-ameter of a dust particle—about one to 10 mi-
crons—some of the light will scatter when itencounters airborne dust. The frequency ofthe scattered light will be higher if the dustparticles are moving toward the laser andlower if the particles are moving away. By an-alyzing the returning signals, the Doppler li-dar can determine wind velocities; in fact,these systems already find use in studies of theatmosphere and at airports to detect windshear and other turbulence.
Some defense analysts, however, are skep-tical that such a device could track a bullet.Because the shock wave would be so localizedand short-lived, the system would need tocrisscross the sky with laser beams to pick upsigns of the atmospheric disturbance and de-termine the bullet’s trajectory. Another chal-lenge would be distinguishing between asniper’s gunshot and bullets fired by friendlyforces or by civilians shooting into the air incelebration (a fairly common occurrence inBaghdad and other Iraqi cities). Says Philip E.Coyle, who was the Pentagon’s director oftesting and evaluation during the Clinton ad-ministration: “Before you can let the troopsshoot back, you need a high-confidence sys-tem producing accurate results.”
Although it is unusual for the military tofield experimental prototypes in war zones,DARPA spokesperson Jan Walker notes that itis not unprecedented. For example, the air-borne surveillance system known as JSTARSwas deployed in Bosnia in 1996, and the un-manned Global Hawk reconnaissance aircraftwas rushed into battle in Afghanistan in 2001.But the success rate for new military tech-nologies is not inspiring: during the 1990s, thegreat majority of army systems that went intooperational testing achieved less than halftheir required reliability, and most air forcetests had to be halted because the systemswere simply not ready.
Walker says the Pentagon is confident thatthe anti-sniper laser will prove useful to thesoldiers in Iraq. But Coyle, who is now a se-nior adviser at the Center for Defense Infor-mation, a Washington, D.C., think tank, isless optimistic. “There’s nothing wrong withtrying it to see if it works,” he says. “But of-ten these things don’t pan out.”
The Fog of WarCAN HIGH-TECH SENSORS FIGHT THE INSURGENCY IN IRAQ? BY MARK ALPERT
DE
FEN
SE
Iraq is not the first place where theU.S. military has attempted to usenovel sensors to detect an elusive
enemy. During the Vietnam War,the U.S. Air Force dropped 20,000battery-powered devices into thejungle along the Ho Chi Minh Trail,
the main supply route for the NorthVietnamese army. The devices—
seismic detectors implanted in theground and camouflaged acousticsensors hanging from the trees—
picked up the movements of troopsand supply trucks, and the
transmitted signals were used totarget bombing runs. The air force
claimed that the operation, dubbedIgloo White, destroyed tens ofthousands of trucks, but laterstudies indicated that the kill
figures had been wildly inflated.North Vietnamese soldiers
apparently disabled many of thedevices and deceived others with
tape-recorded truck noises.
HIDDENENEMIES
WHILE PATROLLING the streets in Iraqi cities, U.S.soldiers have proved vulnerable to sniper attacks.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
M A R C H 2 0 0 4
BE
INE
CK
E R
ARE
BO
OK
AN
D M
ANU
SCR
IPT
LIB
RAR
Y, Y
ALE
UN
IVE
RSI
TY
In a Yale University library sits a mapdepicting the New World that pre-dates the landing of Columbus by 60
years—if it isn’t a fake. Although the lineson the so-called Vinland map are faded,those between scientists on the contro-versy are sharp. New salvos regarding itsauthenticity now come from both sides.
The parchment map, about 11 by 16inches large, was uncovered in a Genevabookshop in 1957 with no records of pri-or ownership. To the west of the inscrip-tions of Europe, Africa and the Far Eastare the words “a new land, extremely fer-tile and even having vines.” The writingalso says the crew of Leif Eriksson namedthe land “Vinland.”
In 2002 Jacqueline S.Olin, retired from the Smith-sonian Center for MaterialsResearch and Education inSuitland, Md., and her col-leagues reported results ofcarbon dating indicating thatthe map dates from 1434,give or take 11 years. Thatfinding bolstered three de-cades of speculation linking itto the Council of Basel, con-vened in Switzerland by theCatholic Church from 1431to 1449. There scholars fromaround Europe assembled todiscuss important affairs,such as the rift in the papacy and the pos-sible reunion of the Eastern and WesternChurches. “The fact that it existed in the15th century certainly presents the veryreal possibility of Columbus, or someonein contact with him, having some knowl-edge of the map,” Olin says.
But since the map’s discovery, criticshave called it a clever fake. What lies indispute is not the pre-Columbian age ofthe parchment but that of the map drawnon it. At the same time Olin and col-leagues dated the map’s parchment,chemists Katherine Brown and RobinClark of University College London ar-
gued that the map’s ink dated from after1923. The ink contained jagged yellowcrystals of anatase, a titanium-bearingmineral rarely found in nature that be-came commercially available in 20th-cen-tury printing ink. “The whole points toan elaborate forgery,” Clark states.
Dueling papers appeared again in re-cent months. With medieval methods,Olin made iron gall inks, which were usedbefore the printing press. She found thather inks contained anatase, results she dis-cusses in the December 1, 2003, issue ofAnalytical Chemistry. She adds that theanatase crystals in the map and her inkswere the same size, citing the electron mi-
croscope work of geologist Kenneth M.Towe, retired from the Smithsonian In-stitution. Those crystals found in moderninks should be about 10 times as large.
Towe vociferously disagrees withOlin’s interpretation of his work in a re-port appearing online in January in An-alytical Chemistry. He concludes that themap’s anatase crystals look modern insize. Moreover, he notes that whereas amap drawn with iron gall inks would rea-sonably be expected to contain iron,“there’s hardly any there.”
Olin responds by suggesting that ironmight have disappeared as the inks deteri-
Drawing the LinesIS A PRE-COLUMBUS MAP OF NORTH AMERICA TRULY A HOAX? BY CHARLES CHOI
CH
EM
ISTR
Y
VINLAND MAP contains references to a new world to the west.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
26 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
SIM
ON
PLU
NK
ETT
SO
HO
/LAS
CO
(ESA
an
dN
ASA)
newsSCAN orated. Regarding the anatase crystal sizes, she
concurs with Towe but says many other inkscontain titanium and should be researchedfurther to see what sizes are present. She addsthat the presence of copper, zinc, aluminumand gold in the map’s ink are also consistentwith medieval manufacturing.
Historian Kirsten A. Seaver, a fellow of theRoyal Geographical Society in London, statesthat the map’s writing contains historicalanachronisms such as mention of Bishop Eirikof Greenland of the early 12th century report-ing to superiors, although he would have hadnone, because Greenland had not yet become
part of the Church hierarchy. “This map ab-solutely screams ‘fake,’” Seaver remarks. Infact, she believes she has found the culprit—aGerman Jesuit priest, Father Josef Fischer, aspecialist in mid-15th-century world maps.Her theory is that Fischer created the map inthe 1930s to tease the Nazis, playing on theirclaims of early Norse dominion of the Ameri-cas and on their loathing of Roman CatholicChurch authority. The map, she supposes,vanished during postwar looting. Seaver’sbook on her search will appear this June.
Charles Choi is based in New York City.
On October 19, 2003, a large solar flareerupted from the surface of the sun,drawing scientists’ attention to three
massive sunspot groups that, over the nexttwo weeks, produced a total of 124 flares.Three of them were the biggest flares ever
recorded. Along with these bursts of electro-magnetic radiation came enormous clouds ofplasma mixed with magnetic fields. Known ascoronal mass ejections (CMEs), these unpre-dictable clouds consist of billions of tons ofenergetic protons and electrons. When direct-ed earthward, CMEs can create problems. Atlast count, the fall’s flares and CMEs affectedmore than 20 satellites and spacecraft (not in-cluding classified military instruments),prompted the Federal Aviation Administra-tion to issue a first-ever alert of excessive ra-diation exposure for air travelers, and tem-porarily knocked out power grids in Sweden.
Historically, CMEs have struck the earthwith little or vague warning. If they could beforecast accurately, like tomorrow’s weather,then agencies would have time to prepare ex-pensive instruments in orbit and on theground for the correct size and moment of im-pact. Such precise predictions could soonemerge: last December researchers announcedthe early success of a forecasting instrument,called the Solar Mass Ejection Imager (SMEI),that can track CMEs through space and time.
Launched in January 2003 on a three-yeartest run, SMEI (affectionately known as“schmee”) orbits the planet over the poles,along the earth’s terminator, once every 101minutes. On each orbit, three cameras capture
Storm SpottingA STEP CLOSER TO FORECASTING DISRUPTIVE SOLAR ACTIVITY BY KRISTA WEST
GE
OM
AGN
ETI
CS
Autumn 2003 saw two weeks ofintense solar activity. The most
serious disruptions of the earth’selectronics systems stem fromcoronal mass ejections (CMEs).
October 19Three massive sunspots rotate
to face the earth.
October 22–23First geomagnetic storm, triggered
by a CME, strikes the earth.
October 28The second-largest flare ever
recorded erupts from the sun.
October 28–30First-ever radiation alert goes outto air travelers above 25,000 feet.
October 29Second CME-triggered geomagnetic
storm hits the earth.
November 4The biggest solar flare
ever recorded erupts; fortunately,the sun has rotated enough so
that no disruptive radiation strikes the earth.
BRIGHT LIGHTS,BIG PROBLEMS
SUN BURPS UP a bulb-shaped cloud called a coronalmass ejection, as seen in February 2000 by the sun-watching satellite SOHO. The mask blots outdirect sunlight; the white circle denotes the sun.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
newsSCANimages that, when pieced together, provide a
view of the entire sky with the sun in the mid-dle. The scattering plasma electrons of CMEsappear on SMEI images as bright clouds.
Other sun-watching instruments can im-age CMEs, but they work like still cameras,taking single pictures of the sun. NASA’s So-lar and Heliospheric Observatory (SOHO),for example, can “see” CMEs erupting fromthe sun quickly but is soon blind to the pathof the clouds. SOHO came in handy last fallwhen it caught two large CMEs headed forthe earth, but it could not follow the ejectanor provide an accurate impact time.
Instead of a SOHO-style snapshot cam-era, SMEI works more like a 24-hour sur-veillance system, constantly scanning andtracking. SMEI begins looking about 18 to20 degrees from the sun and continues imag-ing beyond the earth. SMEI can determinethe speed, path and size of a CME, allowingfor refined and reliable impact forecasts.Such information is particularly useful, sci-entists say, in predicting small CME events.Such ejections can take anywhere from oneto five days to reach our planet. Since its
launch, SMEI has detected about 70 CMEs. During last fall’s solar storms, SMEI had
its first big chance to prove worthy of its es-timated $10-million price tag. Managed pri-marily by the Air Force Research Laboratoryat Hanscom Air Force Base in Massachusetts,about 20 air force and university scientistshave been developing SMEI over the past 20years. At the December 2003 American Geo-physical Union meeting in San Francisco,Janet Johnston, SMEI’s program manager,proudly announced that SMEI had success-fully detected two of the autumn’s largestCMEs about 21 and 10 hours, respectively,before they struck the earth.
Unfortunately, scientists didn’t know ofthe detection and tracking potential until af-ter the storms hit the earth. Right now it takesabout 24 hours for SMEI data to reachHanscom because they travel through multi-ple ground-tracking stations. According toDavid F. Webb, a physicist at Boston Collegewho is part of the SMEI team, precise fore-casting demands a reduction in data-trans-mitting time from 24 to six hours. Such a re-duction will require more researchers at
Cryogenic CuttingLIQUID-NITROGEN JET SLICES AND SCOURS ALMOST ANYTHING BY STEVEN ASHLEY
28 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
NIT
RO
CIS
ION
newsSCAN
Late-night television was once awash ina commercial hawking the “amazingGinsu knife” that never needed sharp-
ening. In the infamous ad, the blade carvedthrough tin cans with ease and then deftly cutpaper-thin slices of tomato. Engineers have re-cently produced an innovative industrial cut-ting device with Ginsu knife–like capabilitiesthat uses a supersonic stream of high-pressureliquid nitrogen. The so-called Nitrojet slicesthrough just about anything—steel girders,concrete slabs, stacks of fabric, meat carcass-es—and never gets dull.
Nitrojet technology was originally devel-oped in the 1990s by scientists at the IdahoNational Engineering Laboratory (INEL) as anonthermal method to cut open barrels ofcombustible waste. Ron Warnecke, presidentof TRUtech, an Idaho Falls–based firm thathandles decontamination and decommission-ing efforts for nuclear weapons facilities,stumbled on the still developmental system inthe late 1990s when he was searching for anenvironmentally safe way to clean and cut upplutonium-processing equipment. TRUtechlater licensed the technology and developedINEL’s prototype into a salable product. War-necke has since set up a new company, Ni-troCision, to market the device.
The supercooled nitrogen jet, whichemerges from special nozzles fitted to a hand-held or robotically positioned wand, seems tocleave materials so well because the dense liq-uefied gas enters a solid’s cracks and crevicesand then expands rapidly, breaking it up fromthe inside. The effectiveness of the process forvarious applications depends on the pressure(6,000 to 60,000 pounds per square inch),temperature (300 to –290 degrees Fahrenheit)
and distance to the workpiece chosen by theuser. Lower pressures enable the nozzlestream to strip tough-to-remove coatings offeven delicate surfaces better than almost anyother cleaning process.
Moreover, the cryogenic jet does not cre-ate secondary waste or cross-contamination;as the nontoxic, supercooled “blade” warms,it simply vanishes into the air. Hazardousrefuse created by stripping or cutting can bevacuumed up at the point of impact.
NASA technicians are now employing a Ni-trojet system at the Kennedy Space Center toprecisely peel thermal-protection coatings offthe inside surfaces of the space shuttle’s sol-id-rocket boosters. Water-jet or similar abra-sive-blasting methods would have requiredthe entire internal surface to be processed,Warnecke reports. The U.S. Navy meanwhilehas contracted to use Nitrojet units to removeanticorrosion coatings from ship decks andhulls, antennas and radomes. Others testingthe technology include aerospace firms Boe-ing and Northrop Grumman, semiconductormanufacturers Semitool and Rogers, paintproducer Sherwin-Williams, Merrimac In-dustries (makers of polyurethane parts) andmeat packers Hormel and ConAgra.
Nitrojet systems, which come on skidsmeasuring four feet by four feet by eight feet,start from $200,000 to $300,000 for a low-pressure unit and go to $450,000 for a fullsystem. These figures represent a considerablepremium over the $150,000-plus price tag fora conventional water-jet unit, but advocatesof the technology say its unique capabilitiesare worth the extra cost. But don’t expect it toappear on late-night infomercials, no matterhow many easy payments are offered.
ground-tracking stations to move informa-tion along and to inspect SMEI’s output.
SMEI’s data gathering may also need per-fecting. Lead forecaster Christopher Balch ofthe Space Environment Center in Boulder,Colo., emphasizes that the CME signal muststand out better against other background
light. Once improved, SMEI “could poten-tially fill a gap in our observations,” Balchsays, by allowing scientists to track CMEsprecisely, thereby making “real-time” fore-casts possible.
Krista West is based in Las Cruces, N.M.
TOO
LS
LIKE A KNIFE through warm butter,a high-pressure jet of liquid nitrogenhews through hard materials, thendisappears into thin air.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
30 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
RO
DG
ER
DO
YLE
newsSCAN
The North was once alive with the aboli-tionist spirit and open to the possibilityof integration. Yet this passion yielded
to several forces that marginalized African-Americans in the 20th century.
Before World War I, blacks were relative-ly few in the North, which together with peo-ple’s need to be near their factories and offices,helped to reduce any tendency toward hous-ing segregation. In New York City, for exam-ple, largely black neighborhoods were usual-
ly only a few blocks long and interspersedwith the homes of working-class white fami-lies. The modern ghetto, with its sharply de-fined racial lines, generally did not begin toform until blacks in substantial numbers mi-grated north beginning in 1916. There theyfound themselves competing for jobs andhousing with immigrants from Europe. Thecompetition was often violent, as in the Chica-go riot of 1919, when 38 people were killed.Violence and the threat of violence, togetherwith agreements among white homeownersnot to sell to blacks, increasingly left African-Americans in separate neighborhoods.
Because blacks had fewer choices, land-lords could charge them more than whites.Crowding increased as tenants took in lodgers,
and many landlords allowed their propertiesto become run down. The Federal HousingAdministration and the Veterans Administra-tion condoned redlining, the practice of deny-ing mortgages to those in minority neighbor-hoods, until well into the 1960s.
Despite the problems, several communi-ties, notably Harlem, were vibrant, at leastuntil the manufacturing economy began todecline in the 1970s. Other factors in the de-terioration include the increasing availabili-ty of crack cocaine, the growth of unwedmotherhood, higher crime rates as the babyboomers came of age, and the disruptive ef-fects of urban renewal. Churches, social clubs,newspapers and unions in black communitieswithered, and banks closed their branches, tobe replaced by currency exchanges thatcharged up to $8 for cashing a check.
To measure segregation, economists DavidM. Cutler and Edward L. Glaeser of HarvardUniversity and Jacob L. Vigdor of Duke Uni-versity calculated dissimilarity scores, whichare defined as the proportion of blacks whowould need to move across census-tract linesto achieve the same proportion of blacks inevery tract of a metropolitan area. By con-vention, a dissimilarity index above 0.6 ishigh, whereas an index of less than 0.3 is low.A score of 0 represents perfect integration and1.0 complete segregation.
As the chart shows, the average index forall metropolitan areas rose steadily to reacha peak of 0.74 in 1960 and then declined to0.5 by 2000. But the largest metropolitan ar-eas, particularly in the North, are still on av-erage far above 0.6. Of 291 metropolitan sta-tistical areas, 72 had dissimilarity scores above0.6 in 2000 and 28 had scores below 0.3.Some of the fastest-growing cities, such as LasVegas and Phoenix, had low and decliningscores. Decreasing scores, however, reflect pri-marily the dispersion of more affluent blacksinto previously white neighborhoods. Thenorthern ghettos and their poverty remain, ar-guably, the number-one problem in the U.S.
Urban Injustice: How GhettosHappen. David Hilfiker.
Seven Stories Press, 2002.
How East New York Became a Ghetto. Walter Thabit.
New York University Press, 2003.
S O U R C E : “ T h e R i s e a n d D e c l i n e o ft h e A m e r i c a n G h e t t o , ” b y D a v i d M .
C u t l e r , E d w a r d L . G l a e s e r a n d J a c o bL . V i g d o r i n J o u r n a l o f P o l i t i c a lE c o n o m y , V o l . 1 0 7 , N o . 3 ; J u n e
1 9 9 9 . D a t a p r i o r t o 1 9 5 0 a r e b a s e do n c i t i e s r a t h e r t h a n m e t r o p o l i t a n
a r e a s . A d d i t i o n a l s e g r e g a t i o n d a t aa r e a t h t t p : / / t r i n i t y . a a s . d u k e . e d u /
~ j v i g d o r / s e g r e g a t i o n
FURTHERREADING
Chicago
Detroit
New York City
Washington, D.C.
Average of allmetro areas
Year1890 19601920 2000
Diss
imila
rity
Scor
e
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
SEGREGATION IN U.S. METROPOLITAN AREAS
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
32 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
JON
FE
ING
ER
SH C
orb
is (
top
); M
ARC
TAU
SS P
hot
onic
a (
bot
tom
); I
LLU
STR
ATIO
N B
Y M
ATT
CO
LLIN
S
newsSCAN M E C H A N I C S
Getting into the SwingExperiments designed to study running mostly take anexternal view of the mechanics. Biologists at North-eastern University have peered directly at running mus-cles by measuring blood flow in the legs of the helmet-ed guinea fowl Numida meleagris. Researchers previously suggested that during running, vir-tually all energy fueling the muscles went to generating force when the foot is on the ground(the stance phase). Now they find that bringing the legs forward (the swing phase) consumedroughly a quarter of the energy used by the hind limbs. Because running birds are the sec-ond-best bipedal sprinters after humans, the investigators say their research should providevaluable clues to understanding human locomotion, with potential benefits to rehabilitativemedicine. Their report appears in the January 2 Science. —Charles Choi
B I O L O G Y
Making and Unmaking MemoriesPrions lie at the root of many disorders, such as mad cow disease and fatal insomnia. Butthe prion ability to adopt a secondary shape—and force other proteins into that shape—doesnot always cause cellular malfunctions, as indicated by a protein called CPEB. Experimentsshow that CPEB, whose normal job involves creating other proteins at synapses during mem-ory formation, has an alternative conformation. Its alter ego is still functional, and it can alsoreshape other proteins, as described in the December 26, 2003, issue of Cell. The prionlikenature of CPEB may help lock in long-termmemories, considering that the prion state istypically durable.
Biological activity may also undergird the voluntary suppression of long-term mem-ories, which has remained controversial sinceFreud. In an experiment, volunteers first mem-orized pairs of unrelated nouns, such as “or-deal/roach.” Then, when looking at the first
word of each pair,they were told notto recall its part-ner. As detailed in the January 9Science, when sup-pression success-fully impaired therecall of the sec-ond word, the pre-frontal cortex wasmore active, fol-lowing a patternsimilar to one seen
when that brain region stops physical ac-tions. At the same time, the memory-form-ing hippocampus activated less, suggestingthat the prefrontal cortex controlled its behavior. —Charles Choi
The discovery of mad cows inCanada and in the U.S. last year
continues the global spread ofbovine spongiform encephalo-
pathy (BSE). Assuming that theNorth American cases represent
the same strain of BSE as seen in the U.K., then the risk of getting
the human form of BSE, calledvariant Creutzfeldt-Jakob disease,
appears to be low.
BSE cases identified in the U.K., up to December 2003: 180,343
Number thought to have enteredthe food chain undetected:
1.6 million
Variant Creutzfeldt-Jakob diseasecases in the U.K.: 143
Number worldwide: 153
Number of countries that haddetected native BSE cases by
1986: 11990: 31995: 52000: 122003: 23
Pounds of U.S. beef produced,2002: 27.1 billion
Pounds exported, 2002: 2.45 billion
Pounds of beef consumedannually, per capita: 67.7
Percent consumed as ground beef: 43.2
S O U R C E S : U . K . D e p a r t m e n t f o rE n v i r o n m e n t , F o o d a n d R u r a l
A f f a i r s ; P r o c e e d i n g s o f t h e R o y a lS o c i e t y , N o v e m b e r 7 , 2 0 0 2 ;
U . K . D e p a r t m e n t o f H e a l t h ; C e n t e r sf o r D i s e a s e C o n t r o l a n d P r e v e n t i o n ;
W o r l d O r g a n i z a t i o n f o r A n i m a lH e a l t h ; C a t t l e F a x ; N a t i o n a l
C a t t l e m e n ’ s B e e f A s s o c i a t i o n ; U . S . D e p a r t m e n t o f A g r i c u l t u r e .
DATA POINTS:MADDENING WORLD
SWING PHASE during running uses moreenergy than previously thought.
BURIED: The brain has abiological mechanism tosuppress memories.
P H Y S I C S
Strangeness in Our Midst?The hot early universe or colliding neutronstars may have coughed up so-called strangequark matter, an extremely dense mix of up,down and strange quarks. If they exist, way-faring nuggets of strange matter might piercethe earth every few years and, like stonesdropped in water, trigger seismic ripples intheir wake. Because a strange nugget wouldfar outpace sound underground, seismo-graphs would record it as a simultaneoustremble from many points along a line. Care-ful sifting through one million seismic reportsbetween 1990 and 1993 revealed one set ofreports from November 1993 that has theright properties for a nugget strike, say Vig-dor L. Teplitz and his colleagues at SouthernMethodist University. Corroborating the re-sult would require scrutinizing new readingsin nearly real time. The findings appear in theDecember 2003 Bulletin of the SeismologicalSociety of America. —JR Minkel
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
34 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
ME
GH
AN K
EN
NE
DY
Un
iver
sity
of
Flor
ida
(to
p);
ALF
RE
D P
ASIE
KA
Scie
nce
Ph
oto
Lib
rary
(b
otto
m)
newsSCAN
G E O C H R O N O L O G Y
Turning Back the ClockThe decay of radioactive carbon is the chief way todate ancient samples. The radiocarbon clock drifts,however, because isotopes do not accumulate consis-tently year to year. So researchers calibrate the clockby dating tree rings and other absolute age measures.A group led by Konrad Hughen of the Woods HoleOceanographic Institution has extended the calibra-tion from 26,000 tothe maximum 50,000years ago (radioactivecarbon becomes scarcebeyond that age). Theresearchers matchedup radiocarbon-datedlayers in marine sedi-ment to annual layersin a Greenland ice core.They had previouslyshown that the twosets of layers are syn-chronous from 10,000to 15,000 years ago,and a French group hasobtained evidence for a similar preliminarytrend. The January 9 is-sue of Science has more.
—JR Minkel
H Y D R O G E N S T O R A G E
All Gassed UpStoring elemental hydrogen for useas a clean fuel requires impractical-ly low temperatures or high pres-sures. In search of a better storagemedium, the daughter-father teamof Wendy and David Mao of theUniversity of Chicago and theCarnegie Institution compressedcrystals of hydrogen and water ormethane with a so-called diamondanvil and cooled them with liquidnitrogen. In one instance, the resultwas a hydrogen-water clathrate, orcagelike crystal, that retained its 5.3percent hydrogen by weight whenit returned to atmospheric pressure.The amount of hydrogen caged isreasonably high—today’s metal hy-dride batteries hold about 2 to 3percent—and could easily be re-leased by warming the clathrate.Different additives and pressureand temperature pathways mightmake such storage crystals morepractical. The research appearedonline January 7 in the Proceedingsof the National Academy of Sci-ences USA. —JR Minkel
■ Supersolid: A new state of matterseems to have emerged afterhelium 4 was sufficiently chilledunder pressure. It turned into a solid whose atoms could, like a superfluid, flow without resistance.
Nature, January 15, 2004
■ NASA’s Stardust spacecraft flewwithin 240 kilometers of CometWild-2 to collect microscopicgrains coming off the object. The samples should reach theearth—specifically, Utah—on January 15, 2006.
NASA announcement, January 2, 2004
■ Prostate cancer cells startresisting drugs by making morereceptors for androgens, whichthe cells ordinarily need toproliferate. Blocking thosereceptors could restore drug efficacy.
Nature Medicine online, December 21, 2003
■ Forget about tar levels: The riskof lung cancer for people whosmoked even very low tarcigarettes was the same as for those who puffed theconventional variety.
BMJ, January 10, 2004
BRIEFPOINTS
A S T R O N O M Y
A Super SuperstarThe Palomar telescope has spied what appearsto be the brightest star yet known, a giant sooversized that it defies current theories. Thestar LBV 1806-20 shines up to 40 million timesbrighter than the sun. The previous recordholder, the Pistol Star, was just roughly six mil-lion times as bright. Some 45,000 light-yearsfrom Earth, LBV 1806-20 weighs about 150times as much as the sun, although present the-ory holds that stars of more than 120 solarmasses could not coalesce, because their nu-clear fires should burn off the excess. Thecolossus is surrounded by what the astronomers call “a zoo of freak stars,” such as a raremagnetic neutron star. Rather than collapsing under their own gravity, LBV 1806-20 andits freaky neighbors may have formed when a supernova shock wave crushed a nearby mo-lecular cloud into stars. The scientists presented their findings at the January meeting of theAmerican Astronomical Society. —Charles Choi
ZOOPLANKTON calledforaminifera, when fossilized,are used to calibrateradiocarbon dating.
BIGGEST AND BRIGHTEST: The star LBV 1806-20could swallow at least eight million suns.
Sun
LBV 1806-20
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
Last month this column detailed how a recent lawsuitcharged biotech giant Genentech with attempting to re-tain rights to a technology for more than a decade be-yond the original patent’s expiration date. These days,however, this type of behavior is by no means confined
to the corporate sector. As uni-versity patenting has increaseddramatically in the years sincethe Bayh-Dole Act of 1980,the law that encouraged suchactivity, academic institutionshave taken a lesson or threefrom the corporations whoseconvoluted tactics keep awhite-knuckled lock aroundvaluable patents. Among theivory tower set, Columbia Uni-versity, that august Ivy Leagueinstitution that is now mark-ing its 250th anniversary, maybe lighting the way for othercenters of learning.
A parade of biotech heavy-weights—among them Amgen, Biogen, Genzyme and,yes, Genentech—filed suits against Columbia last yearfor allegedly trying to prolong for an additional 17years what is said to be one of the most lucrative uni-versity patent estates ever. Three biotech patents thatexpired in 2000 brought the academic institution al-most $300 million in royalties and licensing fees dur-ing their lifetime. But Columbia received anotherpatent in 2002 on what the various plaintiffs claim isessentially the same technology covered by those thathad expired: a method for inserting human genes intohamster cells to identify cells that will produce largevolumes of proteins from those genes. And Columbia,which maintains that the new patent covers a differ-ent invention, has already notified previous licensees ofits intention to keep the cash flowing. But the plaintiffs
in the various suits want the new patent invalidated.The patent fight demonstrates that a university is as
able as any corporation to do anything in its power tocontinue milking an intellectual-property cash cow. Indevising a strategy to maintain a grip on its block-buster, Columbia may even be able to teach corporatepatent holders a few lessons. It enlisted Columbiaalumnus Judd Gregg, now a senator from New Hamp-shire, to stick a provision in a few bills in 2000 thatwould extend its patent protection for 15 months.Moreover, even while the school begged legislators foran extension, it was secretly pursuing new patents, afact never revealed to Congress, according to the com-plaint filed by Foley Hoag, the Boston-based law firmretained by Biogen, Genzyme and Baxter Healthcare.The patent in dispute “surfaced” in 2002 (another oneis still pending) after the unsuccessful lobbying effortwas completed.
This classic “submarine” patenting strategy willprobably be remembered for years to come. The fund-ing for the research for the original three patents camefrom the National Institutes of Health. At the time, Co-lumbia had to obtain title to the invention from theNIH. But in doing so, the NIH stipulated that the uni-versity “shall include adequate safeguards against un-reasonable royalties and repressive practices.”
The Columbia imbroglio illustrates that at least foruniversities, the size of revenues expected from patentsdoes matter. The era of university patenting has led tomany fruitful collaborations in which schools licensetheir discoveries to industry. Often university patents re-ceive only modest royalties or fees. But Columbia’spatents were different. The almost $100 million theygarnered in 1999—a large chunk of the money came to-ward the end of the patents’ term—reportedly consti-tuted nearly 25 percent of the university’s research bud-get. The Columbia patents go to prove that when thestakes are high enough, an institution of “higher” learn-ing can get down and connive with the best of them.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 41
JEN
NIF
ER
KAN
E
Staking Claims
Working the System IICorporate greed no longer remains the sole domain of the corporation By GARY STIX
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
Picture yourself watching a one-minute video of two teams ofthree players each. One team wears white shirts and the otherblack shirts, and the members move around one another in asmall room tossing two basketballs. Your task is to count thenumber of passes made by the white team—not easy given theweaving movement of the players. Unexpectedly, after 35 sec-onds a gorilla enters the room, walks directly through the far-rago of bodies, thumps his chest and, nine seconds later, exits.Would you see the gorilla?
Most of us believe we would. In fact, 50 percent of subjectsin this remarkable experiment by Daniel J. Simons of the Uni-versity of Illinois and Christopher F. Chabris of Harvard Uni-versity did not see the gorilla, even whenasked if they noticed anything unusual(see their paper “Gorillas in Our Midst”at http://viscog.beckman.uiuc.edu/djs_lab/). The effect is called inattentionalblindness. When attending to one task—
say, talking on a cell phone while driving—
many of us become blind to dynamicevents, such as a gorilla in the crosswalk.
I’ve incorporated the gorilla video intomy lecture on science and skepticism giv-en at universities around the country. I al-ways ask for a show of hands of thosewho did not see the gorilla during the first viewing. About halfof the more than 10,000 students I encountered last year con-fessed their perceptual blindness. Many were stunned, accus-ing me of showing two different clips. Simons had the same ex-perience: “We actually rewound the videotape to make sure sub-jects knew we were showing them the same clip.”
These experiments reveal our perceptual vainglory, as wellas a fundamental misunderstanding of how the brain works.We think of our eyes as video cameras and our brains as blanktapes to be filled with sensory inputs. Memory, in this model,is simply rewinding the tape and playing it back in the theaterof the mind, in which some cortical commander watches theshow and reports to a higher homunculus what it saw.
This is not the case. The perceptual system and the brain
that analyzes its data are far more complex. As a consequence,much of what passes before our eyes may be invisible to a brainthat is focused on something else. “The mistaken belief that im-portant events will automatically draw attention is exactly whythese findings are surprising; it is also what gives them somepractical implications,” Simons told me. “By taking for grant-ed that unexpected events will be seen, people often are not asvigilant as they could be in actively anticipating such events.”
Driving is an example. “Many accident reports includeclaims like, ‘I looked right there and never saw them,’ ” Simonsnotes. “Motorcyclists and bicyclists are often the victims in suchcases. One explanation is that car drivers expect other cars but
not bikes, so even if they look right at thebike, they sometimes might not see it.” Si-mons recounts a study by NASA researchscientist Richard F. Haines of pilots whowere attempting to land a plane in a sim-ulator with the critical flight informationsuperimposed on the windshield. “Underthese conditions, some pilots failed to no-tice that a plane on the ground wasblocking their path.”
Over the years in this column I havepounded paranormalists pretty hard, sothey may rightly point to these studies
and accuse me of inattentional blindness when it comes to ESPand other perceptual ephemera. Perhaps my attention to whatis known in science blinds me to the unknown.
Maybe. But the power of science lies in open publication,which, with the rise of the Internet, is no longer constrained bythe price of paper. I may be perceptually blind, but not all sci-entists will be, and out of this fact arises the possibility of newpercepts and paradigms. There may be none so blind as thosewho will not see, but in science there are always those whose vi-sion is not so constrained. But first they must convince the skep-tics, and we are trained to look for gorillas in our midst.
Michael Shermer is publisher of Skeptic (www.skeptic.com)and author of The Science of Good and Evil.
42 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
BR
AD H
INE
S; D
ANIE
L J.
SIM
ON
S (g
rou
p s
hot
)
None So BlindPerceptual-blindness experiments challenge the validity of eyewitness testimony and the metaphor of memory as a video recording By MICHAEL SHERMER
Skeptic
SEE anything unusual?
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
Self-assembly has become a critical implement in thetoolbox of nanotechnologists. Scientists and engineerswho explore the nano realm posit that the same typesof forces that construct a snowflake—the natural at-tractions and repulsions that prompt molecules to formintricate patterns—can build useful structures—say,medical implants or components in electronic chips. Sofar much of the work related to self-assembling nano-structures has been nothing more than demonstrationsin university laboratories. To go beyond being a scien-
tific curiosity, these nanotech materials and techniqueswill have to get from benchtop to a $2-billion semi-conductor fabrication facility.
Four years ago two members of the technical staffat the IBM Thomas J. Watson Research Center inYorktown Heights, N.Y., began to contemplate howthey might transform the vision of self-assembly intoa practical reality. The collaborators, Charles Blackand Kathryn Guarini, knew that the grand academicambitions of making an entire set of chip circuits fromself-assembly had to be set aside. Instead the best wayto begin, they thought, might be to replace a singlemanufacturing step. “The idea was that if we couldease the burden in any of the hundreds of steps to makea chip, we should take advantage of that,” Black says.
They first had to select what type of molecules mightself-construct without disrupting routine silicon manu-facturing practices. Polymers were an obvious choice.They make up the “resist” used in photolithography—
the material that, once exposed to ultraviolet or shorter-wavelength light, is washed away to form a circuit pat-tern. During the first two years of their quest, the duospent time learning about polymers and the optimal tem-peratures and thicknesses at which they would self-as-semble. They built on the work of Craig J. Hawker ofthe IBM Almaden Research Center in San Jose, Calif.,and that of former IBMer Thomas P. Russell, a poly-mer scientist at the University of Massachusetts atAmherst. Both had done research on how polymersself-assemble on silicon. With this knowledge, Blackand Guarini even started making things.
44 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
SAM
UE
L VE
LASC
O
Innovations
Nano PatterningIBM brings closer to reality chips that put themselves together By GARY STIX
LAYERING OF MATERIALS LAYERING OF MATERIALS
EXPOSURE TO ULTRAVIOLET LIGHT
HEAT TREATMENT
REMOVAL OF PMMARESIST DEVELOPMENT
Polystyrene PMMAMask
Silicon substrate Silicon substrate
Silicon dioxideSilicon dioxide
Photoresist Diblock copolymer
CONVENTIONALLITHOGRAPHY
SELF-ASSEMBLYLITHOGRAPHY
1
2
3
1
2
3 OLD AND NEW: Conventional lithography exposes a photoresist toultraviolet light. An etchant then removes the exposed part of thephotoresist. Self-assembly patterning occurs when a diblockcopolymer is heated, thereby separating the two polymers in thematerial into defined areas before the PMMA is etched away. Thetemplate of cyclindrical holes is transferred into the silicondioxide before the holes are filled with nanocrystalline siliconused to store data (steps not shown).
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
The two researchers appeared at conferences, giv-ing presentations about honeycomb patterns that hadself-assembled. But that accomplishment consisted oflittle more than PowerPoints, the type of through-the-microscope images found in abundance at any aca-demic conference on nanotechnology. What would thenano patterns be good for? How could they be inte-grated into a fabrication line? Could they best circuit-patterning techniques that had already received hun-dreds of millions of dollars of investment?
Finally, last year, the pair demonstrated how a self-assembled honeycomb pattern might work in a realmanufacturing facility. The material chosen for thedemo was a diblock copoly-mer, one in which two poly-mers—in this case, polystyrene(Styrofoam) and polymethyl-methacrylate (Plexiglas, orPMMA)—are tied together bychemical bonds. When spunonto the surface of a rotatingsilicon wafer, the two poly-mers separate, as if they wereoil and water. Although themolecules stretch out, thechemical bonds keep them at-tached. Subsequent heat treat-ment exacerbates this elongation. In the end, PMMAends up concentrated in small cylinders surrounded onall sides by the polystyrene. The diblock copolymerthus forms on its own into a nearly complete honey-comblike template.
To finish creating the 20-nanometer-wide pores, anorganic etching solvent removes the PMMA. A subse-quent etching step transfers the same honeycomb pat-tern into an underlying layer of more robust silicondioxide. Then a coating of amorphous silicon gets de-posited across the surface of the wafer. A gas etchesaway the silicon except for that deposited in the holes.All that is left are nanocrystalline cylinders surround-ed by silicon dioxide. The final steps place an insulat-ing layer and a block of silicon atop the structure, theblock forming a “gate” that turns the electronic deviceoff and on. Black and Guarini’s honeycomb results ina nanostructure that is part of a working flash-memo-ry device, the kind that retains digital bits even when acamera or a voice recorder is turned off. The nanocrys-talline cylinders form capacitors where data are stored.
Manufacturing engineers are leery of introducingnew technologies unless a researcher can make a verygood case for their adoption. Self-assembly potential-
ly fits the bill. Creating closely spaced holes for a flashmemory would prove exceedingly difficult with ordi-nary lithographic and deposition methods. Formingnanocrystals using conventional techniques creates el-ements of different sizes that are all jumbled together.In contrast, the self-assembled nanocrystals are evenlyspaced and of uniform size, improving their durabilityand their capacity to retain a charge while allowing thecylinders to shrink to smaller than 20 nanometers.
The IBM demonstration served as proof of princi-ple in the strictest sense of the expression. The com-pany has not made commercial flash memories foryears, so the invention could not be applied immedi-
ately to improve its own manu-facturing operations. But thenanocrystals enabled the pair ofresearchers to flaunt this type ofnano patterning. “Politically inthe company maybe it wasn’tthe smartest demonstration wecould have done, but everybodywas supportive and could seethe power of the technology,”Black says.
The understanding gainedof how to integrate nanomanu-facturing with conventional
chipmaking may provide new approaches to fabricat-ing other IBM electronic components. Making holesand filling them could create “decoupling” capacitorsrecessed into the chip substrate that smooth out fluc-tuations in the power supplied to a chip.
Using a variant of nano patterning, a self-assemblingpolymer could also create tiny, finger-shaped siliconprotrusions sticking up from the underlying substrate.These fingers would constitute the “channel” in a tran-sistor through which electrons flow—but one in whichelectrons flow vertically instead of across a chip, as intoday’s devices. The gate to turn the transistor off andon could encircle the silicon finger. The geometry mightprevent electrons from “tunneling,” or leaking, throughthe channel when the transistor is in the off state, a con-stant threat when feature sizes become very small.
Ultimately, self-assembly might play a much biggerrole in fashioning electronic circuits. But the incre-mentalist approach of Black and Guarini may repre-sent the most promising way to get nanotechnologyadopted as a real manufacturing tool. “The greatest ex-citement is that these materials aren’t just in the poly-mer-science laboratory anymore,” Black says. A smallstep for small manufacturing.
46 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
SAM
UE
L VE
LASC
O
Innovations
Silicondioxideinsulatinglayers
NANOCRYSTAL DEVICE
Silicon nanocrystals
Silicon gate
Silicon substrate
FLASH MEMORY: A layer of self-assembled siliconnanocrystals is inserted into an otherwise standarddevice as part of a novel IBM manufacturing process.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
48 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
Insights
Late last spring World Health Organization officialstalked about putting severe acute respiratory syndrome,or SARS, “back in the box” before it could become en-demic in China and the other countries to which it hadspread. The virus infected more than 8,000 peopleworldwide and killed nearly 800 last year. But so far this
season, it had caused just a handful of possible cases bymid-January, with only two confirmed, one the resultof a laboratory accident. If SARS has indeed been tamed,without a vaccine or any effective drug treatment, it willbe a triumph for the good old-fashioned public healthtactics of surveillance and infection control.
“Identify cases, isolate, contact tracing, and whencontacts get sick, [do it] all over again” is the not so se-cret formula for containing disease outbreaks, accord-ing to David L. Heymann, the veteran pathogen fight-er who led WHO’s response to SARS last year as exec-utive director of the agency’s communicable diseasesdivision. Whether it’s SARS, smallpox or polio, the fun-damentals of stopping infectious disease are the same,he says: find it and break its chain of transmission. Heis not declaring victory against SARS just yet, though.Only another full year of surveillance will tell whetherthe virus has become endemic, he says, “so we need tohave the mechanisms in place to detect this one and todetect any new one that emerges, too.”
The 58-year-old American has learned the value ofvigilance over 30 years of battling infectious diseases,both new and old, around the world. Fresh out of theLondon School of Hygiene and Tropical Medicine in1974, he was recruited, along with hundreds of otheridealistic young doctors, by Donald A. Henderson, whowas running WHO’s global smallpox eradication pro-gram. Heymann spent two years in India administeringsmallpox vaccinations. In 1976, thoroughly hooked oninternational public health, he returned to the U.S. tojoin the Centers for Disease Control and Prevention’sepidemic intelligence service.
That year “swine flu” provoked fears of a killer in-fluenza pandemic, prompting the CDC to bolster in-fluenza surveillance. When the agency heard about anunusual respiratory infection spreading at an AmericanLegion convention in Philadelphia, Heymann was senton his first outbreak investigation. Instead of flu, the ill-ness turned out to be a new one, later dubbed Legion- YV
ES
LER
ESC
HE
Loo
kat
Ph
otos
A Strategy of ContainmentPathogens take windows of opportunity, and so must humans, says David L. Heymann, who helped to create a global early-warning and response network By CHRISTINE SOARES
Insights
■ On being called a “roustabout epidemiologist”: “That’s the beauty ofunderstanding a little bit about epidemiology and many differentdiseases—you can jump from one to another. You can figure out whichprinciples you can apply and which you can’t apply and take a fresh look at a new issue.”
■ SARS lesson learned: The world’s health ministers voted unanimously lastMay to allow the World Health Organization to act on information from allsources, not just official reports; all countries must now report anydisease outbreak of “international concern.”
■ At least 34 new pathogens have been identified in the past three decades.
DAVID L. HEYMANN: PATHOGEN PATROL
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
naire’s disease. Just a few weeks later Heymann “got lucky”again, he says. It was Christmastime and he was single, so hewas sent to Zaire (now Congo-Kinshasa) to investigate a high-ly lethal hemorrhagic fever ravaging patients and health careworkers. That virus would be named Ebola.
Heymann spent the next 13 years in West Africa workingfor the CDC and crossing paths again with Henderson, who al-ways found the tireless epidemiologist to be “a really positiveperson, optimistic, very intelligent” and a critical thinker who“could examine what’s being done and how to improve it.”
In 1995, when WHO asked Heymann to create an emerg-ing and infectious disease program, it was clear that the agency“really didn’t have a useful tool in outbreak alert and response,”Heymann remarks. Having chosen to be loaned to WHO, ratherthan climb the CDC senior management ladder in Atlanta, Hey-mann was by then living just outside Geneva,married and a father of three but still jetting offto help contain disease outbreaks. Often WHO’said arrived late because the agency relied onmember nations to voluntarily report domesticoutbreaks, with the exception of yellow fever,cholera and plague, for which reporting wasmandatory. The problem was, the very devel-oping countries where diseases were most like-ly to flare up had little systematic surveillance.By the time the central government realized thatan outbreak was happening, it could havereached crisis proportions.
Once WHO did learn of an outbreak, theagency could only deal directly with nationalgovernments to offer advice and, if invited, as-sistance, albeit with limited resources. But ear-lier in 1995 Heymann had been in Kikwit,Zaire, during a large Ebola outbreak, and he was struck by thenumber of “other actors out there waiting to help.” The RedCross, Doctors Without Borders and additional nongovern-mental organizations could act as eyes and ears for WHO, he re-alized, as well as extra hands during emergencies.
So Heymann and his team set out to create what he calls “anetwork of networks.” It would include laboratories and expertsaround the world pledged to work with WHO when called onand a semiformal array of informants. Also determined to tapinto the digital information stream, Heymann’s group collabo-rated with Canada in 1998 to create a Web-crawling programthat searches for hints of disease outbreaks. “He’s been innov-ative in a number of ways,” says emerging-disease specialistStephen S. Morse of Columbia University, simply by “connect-ing up sources of information—in the intelligence communitythey call it ‘all-source information’—and ‘stovepiping’ existinginformation, making sure it gets to the right people.”
The WHO formally unveiled its Global Alert and ResponseNetwork in 2000, but SARS was the first multicountry outbreakthe coalition faced [see “Caught Off Guard,” by ChristineSoares; News Scan, Scientific American, June 2003]. “Wehad a vision of a world on alert and able to respond to emerg-ing and other infectious diseases,” Heymann says. “This was itsinternational rollout, and it worked.” Scientists from 17 coun-tries worked on SARS, he notes, “and when you have real-timeinformation, you can make evidence-based decisions and WHOcan play that role.”
Heymann, too, has a new role, having been charged withWHO’s current attempt to completely eradicate an old diseasefrom the world—this time, polio. He took over the job last Julyfrom Jong Wook Lee when the latter became WHO’s directorgeneral, and in January, Heymann made a bold public promise
to stop the transmission of wild poliovirus in allcountries by the end of this year. The move wascalculated to draw world attention and to puton the spot the leaders of the nations where po-lio is still endemic. “We have to do it,” Hey-mann says of the self-imposed deadline. “If wedon’t, we might have to admit that it might notbe feasible to do. The only thing that may belacking now is political will.”
In its 16th year, the eradication program hasalready cost $4.6 billion. Just six countries havewild poliovirus transmission within their bor-ders, but political squabbles have bogged downimmunization efforts in some areas. Polio is alsomuch harder to ferret out, notes Henderson,who served as WHO’s adviser for polio eradi-cation in the Americas. Unlike smallpox, whichproduces dramatic symptoms in all victims, po-
lio causes a distinctive “acute flaccid paralysis” in only one ofevery 200 cases. “You just didn’t know where it was until youfound that first case,” Henderson explains. “I wish him well,” hesighs. “If anyone can do it, it’s David.”
Henderson, Morse and other observers are less confidentthat international support for WHO’s efforts to bolster globaldisease surveillance will continue now that the program’scharismatic leader is gone. “If it doesn’t go on without me, it waspretty poorly conceived, and I think there’s no question that itwill,” Heymann declares. Besides, he enjoys starting things morethan maintaining them and relishes the chance to reinvigoratethe polio program.
The challenge is rejuvenating him in turn, Heymann says, bygetting him into the field more often. “Somebody told me oncethat you have idealism candles that burn, and those candles slow-ly go out, but you can rekindle them. The fire burns brightly againwhen you get out and see, really, the need in this world.”
50 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
CO
UR
TESY
OF
DAV
ID L
. H
EYM
ANN
Insights
INDIA, 1974: David L. Heymannreceives a smallpox vaccination to demonstrate its safety.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
At 8:15 P.M. Pacific time on January 3, the Spirit rover,tucked inside its protective capsule, separated from itsinterplanetary mother ship and prepared to enter the
atmosphere of Mars. For weeks, mission engineers and scien-tists had been listing in grim detail everything that could gowrong. Explosive bolts might not blow on time; strong windsmight slam the capsule against the ground; the lander mightsettle with its nose down, wedged helplessly between rocks; ra-dio links might fail. As the final days ticked by, a dust storm onthe planet erupted, reducing the density of the upper atmo-sphere. To compensate, controllers reprogrammed the para-chute to deploy earlier. Eight hours before the capsule’s entry,deputy mission manager Mark Adler said, “We’re sending acomplicated system into an unknown environment at very highspeed. I feel calm. I feel ready. I can only conclude it’s becauseI don’t have a full grasp of the situation.”
This candid doom-mongering was reassuring. If the teamhad said there was nothing to worry about, it would have beentime to start worrying. Between 1960 and 2002 the U.S., Rus-sia and Japan sent 33 missions to the Red Planet. Nine made it.By the standards of planetary exploration, the failure rate is notunusually high: of the first 33 missions to the moon, only 14
succeeded. But the blunders that damned the Mars Climate Or-biter in 1999—neglecting to convert imperial to metric units,then failing to diagnose the error when the spacecraft kept drift-ing off course—are hard to live down. And just a week beforeSpirit reached Mars, the British Beagle 2 lander bounded intothe Martian atmosphere never to be heard from again.
Controllers at NASA’s Jet Propulsion Laboratory (JPL) havea tradition of opening a bag of peanuts for good luck, and themoment had come to do so. At 8:29 P.M., Spirit started its mete-oric descent. (To be precise, that is when the confirmation signalreached Earth. By then, Spirit had already landed on Mars; theonly question was whether it had landed in one piece or in many.)Within two minutes, the lander had survived the peak atmo-spheric heating and maximum g-force. After another two min-utes, it deployed its chute and emerged from its capsule. Twominutes later its cushion of air bags inflated and controllers an-nounced, “We have signs of bouncing on the surface of Mars.”
The control room became a blur of cheering and hugging.It didn’t take long, though, for people to wonder whether theyhad cheered and hugged too soon. The radio signal had flat-lined. Rob Manning, the leader of the group that devised thelanding sequence, recalls: “The signal disappeared. That
THE SPIRIT OF
LOW-GAIN ANTENNA
DIRECTION OF FIRSTLONG JOURNEY
HIGH-GAIN ANTENNA
A
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
caused us some pause. I was trying to act calm. It was nerve-wracking.” Up until then, he says, the entry had felt just likeone of the team’s many test runs. “It was only when the signalstarted going away that I said, ‘Uh-oh, this is not a rehearsal.’”
Engineers had warned that Spirit might go silent for 10 min-utes or so until it rolled to a stop. A tumbling lander does notmake a good transmission platform. But the 10th minute cameand went without contact, then the 11th and the 12th. Peopleswiveled in their chairs, crossed their arms, chewed gum. A thinjittery line, representing radio static, ran across the bottom ofcontrollers’ computer screens. Manning says he was watchingthe bottom of his screen so intently that it took him a momentto notice when the line jumped to the top. At 8:52 P.M., or 2:51P.M. local time at the landing site, Spirit proclaimed its safe ar-rival on the Red Planet.
Squyres’s OdysseyLIKE SAILORS ROUNDING Cape Horn, scientists and en-gineers willingly put themselves in the capricious hands of fatefor a reason: to put life on our planet into context, either as asingular phenomenon or as an exemplar of a universal process.Steve Squyres, principal investigator of the rover’s scientific in-
struments, has been trying to get to Mars for 17 years. The Cor-nell University professor has something of a wunderkind repu-tation. He did his Ph.D. from start to finish in three years and,during the 1980s, became an expert on half the solid bodies ofthe solar system, from the icy satellites of Jupiter to the volcanicplains of Venus to the water-cut highlands of Mars. But he cameto feel that his career was missing something.
“The real advances in our business come from people whobuild instruments and put them on spacecraft and send themto the planets,” he says. “I worked on Voyager; I worked onMagellan. I didn’t think of those missions, I didn’t design thoseinstruments, I didn’t calibrate them. I just parachuted in at theend, scooped up some data and went off and wrote a bunch ofpapers. It was a very enjoyable, satisfying way to do a career,in a lot of respects, but I did feel that I was profiting by the ef-forts of others. For just once—and it is going to be just once;this is an experience neither to be missed nor repeated—for justonce I wanted to do one where at the end I could say, You
EASTERN PANORAMA from the Spirit landing site runs from due north at theleft to due south at the right. The first major goal of the rover is to reach acrater about 250 meters to the northeast. Later it could drive toward the EastHills, which lie three to four kilometers away and are about 100 meters high.
NASA’s rover fights the curse of the Angry Red Planet BY GEORGE MUSSER
EXPLORATION
EAST HILL COMPLEX
B C D E F G
NAS
A/JP
L/C
OR
NE
LL U
NIV
ER
SITY
SOUTH MESAS 1 & 2COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
54 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
NAS
A/JP
L/M
ALIN
SP
ACE
SC
IEN
CE
SYS
TEM
S; N
ASA
PLA
NE
TAR
Y D
ATA
SYST
EM
(to
p v
iew
)
13˚S
14˚S
15˚S
16˚S
17˚S
1531
1131
731
331
–69
–469
–869
–1269
–1669
–2069
–2469
–2869
_�
_�
_�
_�
_�
_�
_�
_�
_�
_�
_�
_�
�
m
173˚E 174˚E 175˚E 176˚E 177˚E 178˚E
GUSEV CRATER
POSSIBLE DELTA REMNANTS
MA’ADIM VALLIS
POSSIBLE OUTFLOW (WALL BREACH)
HUMANITY’S NEW BEACHHEAD ON MARS
5 km
20 km
SPIRIT’S LANDING SITE, Gusev Crater, is only the fourth place onMars that humans have seen in any detail. The crater lies on theboundary between the southern highlands and northern plains. It is one of half a dozen possible lake beds that scientists have
identified on the Red Planet. The landing sites of the ill-fated Beagle2 and of Opportunity, Spirit’s twin, may also have been ancient lakes.The earlier Mars Pathfinder rover roamed the mouth of a largeoutflow channel. The Viking landers set down on featureless plains.
VIKING 1 (CHRYSE PLANITIA)
OPPORTUNITY(MERIDIANI PLANUM)
MARS PATHFINDER(ARES VALLIS)
VIKING 2(UTOPIA PLANITIA)
BEAGLE 2 (ISIDIS PLANITIA)
SPIRIT (GUSEV CRATER )
GUSEV CRATER is just north of Ma’adim Vallis, a canyon 900kilometers long. The regional view (a) shows topography(colors) and strips of high-resolution images. The high densityof craters implies an ancient terrain, perhaps four billion yearsold. Mosaics of high- and low-resolution images (b, c) zoom inon the landing site. The ellipses represent the targetedlanding area (which changed slightly over time); the yellowlines are sight lines from the rover’s initial position.
NORTH HILL
NORTHWEST HILL
SOUTHWEST HILL
SOUTH SOUTHWEST HILL
SOUTH MESAS
EAST HILL COMPLEX (A-G)
5 km
DUST DEVILTRAILS POSSIBLE
CRATER LAKE
a
b
c
LANDINGSITE
LANDING SITE
0˚
90˚S
60˚S
30˚S
90˚N
60˚N
30˚N
180˚W 90˚W 0˚ 90˚E 180˚E
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 55
ALFR
ED
T.
KAM
AJIA
N;
SOU
RC
E:
NAS
A/C
OR
NE
LL U
NIV
ER
SITY
(to
p);
NAS
A/JP
L/C
OR
NE
LL U
NIV
ER
SITY
(ce
nte
r le
ft a
nd
rig
ht;
bot
tom
lef
t);
NAS
A/JP
L/AR
IZO
NA
STAT
E U
NIV
ER
SITY
/CO
RN
ELL
UN
IVE
RSI
TY (
bot
tom
rig
ht)
LANDING SEQUENCE of the Spirit rover followed the pattern pioneered by MarsPathfinder in 1997. Spirit entered the atmosphere at 5.4 kilometers a second.Drag on the heat shield reduced its speed to 430 meters a second, theparachute slowed it to 70 meters a second, and rockets brought it to rest seven meters above the ground. (The rockets did not bring it all the way down, because that would have required extremely precise distance measurementsand finely tuned rocket control.) Protected by air bags, Spirit bounced 28 timesand came to rest about 300 meters southeast of the point of first impact.
CRUISE-STAGE SEPARATION
ATMOSPHERIC ENTRY
PARACHUTE DEPLOYMENT
BRIDLE CUT
HEAT-SHIELDJETTISONING
BOUNCING
PETAL OPENING
ROVER PREPARATION
EAST HILLS E (left) and F (right) were imaged several hours apart, showing how dust affects thevisibility. The atmosphere above Gusev is dustier than predicted; consequently, the rover is warmerbut has less solar power. Hill E is 3.1 kilometers away, and F is 4.2 kilometers away.
SMOOTH ROCK SURFACES may have beenpolished by windblown sand grains. This is one of the first color images taken by Spirit.
THERMAL SCAN shows the area from the East Hill Complex to Sleepy Hollow. Dust is warmer (red)because it has a low thermal inertia, which means it heats up quickly in the sun. Rocks, with their higherthermal inertia, stay cooler (blue). Other data from the infrared spectrometer reveal magnesiumcarbonate and hydrated minerals, but no one yet knows what it means for the history of water at Gusev.
DRAG MARKS, left by the air bags as they wereretracted, indicate a cohesive soil—perhapselectrostatically charged dust or a weaklycemented “duricrust” like that seen by Viking.
LATE MORNING EARLY AFTERNOON
RETROROCKET FIRING
ROLL-OFF
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
know, okay, that was something that I helped make happen.”In 1987 Squyres put together a team, built a camera and
proposed it to NASA for what became the Mars Pathfinder mis-sion. It had the wrong dimensions and was disqualified. He alsojoined one of the instrument teams for the Mars Observerspacecraft. Shortly after it lifted off in September 1992, itsbooster rocket fired to break out of Earth orbit, and the fragili-ty of spaceflight intruded. The radio signal went dead. Sittingin the auditorium at launch control, Squyres put his head in hishands and said, “I think we may have lost it. I think we mayhave lost it.” Forty minutes later the spacecraft reappeared. Itvanished for good when it got to Mars the following year.
In 1993 Squyres and his team proposed another instrumentpackage and were again turned down. As they were develop-ing yet another set of plans, for a full-blown mobile geology labcalled Athena, news broke that a meteorite discovered inAntarctica might contain hints of past life on Mars. The hooplareenergized Mars exploration. The Pathfinder mission in 1997showed what a rover could do, and in November of that yearNASA gave the go-ahead to Athena. Squyres found himself theleader of 170 scientists and 600 engineers.
Two years later NASA lost the Mars Climate Orbiter and theMars Polar Lander. Although Squyres’s team was not directlyinvolved, the fiascoes convulsed the entire Mars program. In re-sponse to an investigation panel, which put the blame largelyon a caustic mix of underfunding and overconfidence, the agen-cy increased the budget for the rovers; they eventually cost $820million. Redesigned and refocused, Spirit and its twin, Oppor-
tunity, finally blasted off last summer. “To get through some-thing like what we went through, you have to be optimistic bynature,” Squyres says. “To be prepared for every eventuality,you also have to be pessimistic by nature.”
Freeze-Dried Planet AS THE TWO Mars Exploration Rovers (MERs) were com-ing together, Martian science went through an upheaval. TheMariner and Viking missions of the 1960s and 1970s revealeda cold, dry and lifeless world, but one etched with remnants ofpast vigor: delicate valley networks from the distant past andvast flood channels from the intermediate past. Researchers ex-pected that when new space probes assayed the planet, theywould find water-related minerals: carbonates, clays, salts.
Over the past six and a half years, the Mars Global Surveyorand Mars Odyssey orbiters—bearing duplicates of the instru-ments that the ill-fated Mars Observer carried—have lookedfor and detected essentially none of those minerals. They havefound layers of olivine, a mineral that liquid water should havedegraded. And yet the orbiters have also seen fresh gullies, oldlake beds and shorelines, and an iron oxide mineral, gray hema-tite (as opposed to red hematite, otherwise known as rust), thattypically forms in liquid water. The planet holds extensive reser-voirs of ice and bears the marks of recent geologic and glacialactivity. Scientists are more baffled than ever.
“There’s a fairly raging debate about how the environmentof early Mars differed from now,” says Matt Golombek, theJPL planetary geologist who led the Pathfinder science team andis a member of the Mars Exploration Rover team. “MER is re-ally the first attempt to go to the surface and try to verify whatthe environment was really like.”
The notoriously risk-averse Viking planners sent their twolanders to the most boring places on Mars. (To be fair, you’d
WESTERN PANORAMA runs from due south at the left to due north at theright. The prominent light-colored area is Sleepy Hollow, a shallowdepression about nine meters in diameter and located about 12 metersaway. Dark marks on the dusty surface of the hollow may be places wherethe rover bounced before settling down.
SOUTH SOUTHWEST HILLSOUTH MESAS 1 & 2 SOUTHWEST HILL COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
probably do the same if you had a $3.5-billion, easily toppledspacecraft and knew almost nothing about the terrain.) Path-finder, though bolder, was really just a test flight. Beyond a de-sire to study as many different rocks as possible, Golombek’s teamdidn’t much care where it went. Spirit and Opportunity are thefirst landers to visit places that scientists actively wanted to go.
From orbit, Spirit’s new home, Gusev Crater, looks like alake bed. It has fine layering, deltalike deposits and sinuous ter-racing, and it sits at the northern end of Ma’adim Vallis, one ofthe largest valleys on the planet. Opportunity has gone for thegray hematite, which is concentrated in Meridiani Planum. PhilChristensen, a planetary geologist at Arizona State University,recently studied the topography of the hematite outcrops andconcluded that the mineral forms a thin, flat layer—as thoughMeridiani, like Gusev, was once a lake bed.
Only on the surface can these hypotheses be tested. For in-stance, because wind cannot transport sand grains larger thanhalf a centimeter, the discovery of bigger grains would implyanother agent of erosion, probably water. When hematite crys-tallizes in lake water (as opposed to, say, a hot spring), thechemical reaction often involves the mineral goethite, whichspectrometers on the rovers can look for. Piece by piece, datumby datum, the rovers should help resolve how Mars can be bothso Earth-like and so alien.
Mars under the EarthlingsABOUT THREE HOURS after Spirit landed, at 11:30 P.M.
Pacific time on January 3, the data started to pour in, relayed by the Odyssey orbiter. For observers used to earlier missions,when images slowly built up line by line like a curtain risingon another world, it was startling. The first pictures flashed upon the screen, and Gusev Crater leapt into the control room.
The main cameras sit on a mast 1.5 meters tall, so the view
closely matches what you’d see if you stood on the planet. Butit still takes some getting used to. Jim Bell, a Cornell scientistwho has worked on the color panoramic camera, Pancam, since1994, says: “One thing that I learned through all the testing wedid is when you experience a place through the eyes of a rover,and then go yourself, it’s pretty different. The sense of depth isvery different, because you’re looking at this flat projection ofthe world, and there’s nothing in it for human reference.There’s no trees, no fire hydrants—you’re missing all the cueswe have all around us that tell us how far away things are.”
Even so, the first images have an eerily familiar quality,showing rocks, hollows, hills and mesas. “It’s beautiful in thesame way the desert is beautiful,” aerospace engineer JulieTownsend says. “It’s a beautiful vacantness, the beauty of anundisturbed landscape.”
But space exploration is like plucking the petals of a daisy:it works, it works not, it works, it works not. You never knowhow it will end. Early morning Pacific time on January 21, con-trollers were preparing Spirit to analyze its first rock, namedAdirondack. They instructed the rover to test part of the infraredspectrometer, and Spirit sent the robotic equivalent of “roger.”But then it went silent. For two days, controllers tried nearly adozen times to reach it. When they finally reestablished contact,the situation was serious. Though in no imminent danger, Spirithad rebooted itself more than 60 times trying to shake off a faultit could not diagnose. Pete Theisinger, the project manager, says,“The chances it will be perfect again are not good.” But he adds,“The chances that it will not work at all are also low.” And that,in the business of planetary science, is a victory.
George Musser, a staff writer, was a graduate student ofSteve Squyres’ in the early 1990s. For updates on the Spiritand Opportunity missions, see www.sciam.com
NORTHWEST HILLSLEEPY HOLLOW
NAS
A/JP
L/C
OR
NE
LL U
NIV
ER
SITY
DIRECTION OF ROLL-OFF
SASHIMI ADIRONDACK
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
58 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
SCO
TT G
OLD
SMIT
H ROBOTSROBOTS
A New Race of
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 59w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 59
DECEMBER 2, 2003: The Red Team prepares its robotic vehicle,Sandstorm, for its maiden voyage. As Nick Miller, one of severaldozen Carnegie Mellon University undergraduates on the team,drives the robot around a test loop between abandoned steel mills inPittsburgh, onboard computers (in metal box) record the test path.Five days later the robot drives the loop with no one at the wheel.
Around the U.S., engineers arefinishing one-year crash projectsto create robots able to dash 200 miles through the MojaveDesert in a day, unaided byhumans. Scientific American tailedthe odds-on favorite team for 10 months and found that majorinnovations in robotics are notenough to win such a contest.Obsession is also required
BY W. WAYT GIBBS
PITTSBURGH, DECEMBER 10, 2003: A cold rain blows sideways through the
night into the face of Chris Urmson as he frets over Sandstorm, the robotic vehicle
idling next to him on an overgrown lot between two empty steel mills. Urmson checks
a tarp protecting the metal cage full of computers and custom electronics that serves
as the sensate head of the chimeric robot, which has the body of an old Marine Corps
Humvee. His ungloved hands shivering and his body aching from three sleep-deprived
days and nights of work in the field, Urmson stares glumly at the machine and weighs
his options. None of them are good.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
He and his teammates had vowed months ago that by mid-night tonight Sandstorm would complete a 150-mile journey onits own. It seemed a reasonable goal at the time: after all, 150miles on relatively smooth, level ground would be but a babystep toward the 200-mile, high-speed desert crossing that the ro-bot must be ready for on March 13, 2004, if it is to win the U.S.Department of Defense’s Grand Challenge race, as well as the$1-million prize and the prestige that accompanies an extraor-dinary leap in mobile robotics.
But after 20 hours of nonstop debugging, Sandstorm’s nav-igational system is still failing in mystifying ways. Two days agothe machine was driving itself for miles at a time. Last night itcrashed through a fence, and today it halts after just a few lapsaround the test path. The dozen or so team members here arewet, cold and frazzled, hunched over laptops in a makeshift lean-to or hunkered down in a van. The 28-year-old Urmson hashardly seen his wife and two-month-old baby for weeks. Con-
tinuing under these wretched conditions seems pointless.On the other hand, an hour ago he and the rest of the group
huddled around William “Red” Whittaker, the leader of the RedTeam—and Urmson’s Ph.D. adviser at Carnegie Mellon Uni-versity (CMU)—and acceded to his decision that they wouldcontinue fixing and testing through the night and into the dayand through the night again, if need be, until Sandstorm com-pleted the 150-mile traverse they had promised. For theumpteenth time, Red repeated the team’s motto: “We say whatwe’ll do, and we do what we say.” Their reputations, theirmorale—and for the students, their final-exam grades—are onthe line.
But at the moment, Whittaker is not around, so Urmson, asthe team’s technical director, is in charge. He looks at the rivuletsstreaming over the tarp, considers how many weeks of workcould be undone by one leak shorting the circuits inside, andaborts the test, sending everyone home to their beds.
60 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
LUC
Y R
EAD
ING
; SO
UR
CE
S: M
DE
P,
BLM
, SR
A/D
ARP
A (m
ap
);
CO
UR
TESY
OF
THE
RE
D T
EAM
, C
ARN
EG
IE M
ELL
ON
UN
IVE
RSI
TY (
top
lef
t a
nd
bot
tom
lef
t);
ALE
X G
UTI
ER
RE
Z (t
op r
igh
t);
JASO
N D
IVE
NE
RE
(b
otto
m r
igh
t)
DARPA ANNOUNCED in February 2003 that it was organizing adesert race for self-navigating robotic vehicles to be held onMarch 13, 2004. The race was named the Grand Challengebecause its requirements—cross 200 miles of unfamiliar, roughterrain in 10 hours or less, without any human assistance—fellwell beyond the capabilities of any robot yet designed.
THE PRIZE: $1 million to the team whose vehicle completes thecourse in the shortest time less than 10 hours.
THE RULES: The robotic racers must be fully autonomous; duringthe race they cannot receive signals of any kind (except a stopcommand) from humans. The vehicles must stay on the groundand within the boundaries of the course. No robot may
intentionally interfere with another. The race will begin with astaggered start; a qualifying event will determine who goes first.If no vehicle wins in 2004, the race will be repeated each yearuntil there is a winner or the funding runs out (after 2007).
THE COURSE: Two hours before the race begins, DARPA officialswill give each team a CD-ROM containing a series of GPScoordinates, called waypoints, spaced 150 to 1,000 feet apart.The width of the route between waypoints will also vary: in somesections of the course, racers will have to remain within a 10-foot-wide corridor, whereas in other sections they will be able toroam more freely. Depending on how officials mix and matchfrom various potential routes through the Mojave Desert (map),the course may be as short as 150 miles or as long as 210 miles.
58395
15127
95
62
247
4015Needles
Baker
Las Vegas
Death Valley National Park
Ft. Irwin Army Base
Twentynine Palms Marine
Corps Base
JoshuaTree National Park
Mojave NationalPreserve
Barstow
LosAngeles
Primm
Victorville
Searchlight
Henderson
RACE OFFICIALS have warned participants to expect sandytrails, narrow underpasses, power line towers and hairpinturns. The Red Team is creating a test course in Pittsburgh thatincludes all of these hazards.
All-terrain area Federal park Military base
POSSIBLE RACE ROUTESUnpavedPaved
The Grand Challenge Race
30 miles
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
The next day brings hell to pay. Like an angry coach at half-time, Whittaker castigates the team for giving up and for miss-ing other self-imposed goals. “A great deal of what we agreedto do got lost as the team focused monotonically on the 150-mile objective,” he rebukes. “The vehicle body didn’t get paint-ed; the Web site didn’t get updated; the sensor electronicsweren’t completed. And do we win the race if we don’t have bet-ter shock isolation than we have now?” Heads shake. “No, we’lllose the race. Is the condition of this shop consistent with whowe are?” he asks, waving at the tools and parts scattered overevery flat surface. Eyes avert. He clenches his jaw.
“Yesterday we lost that sense deep inside of what we’re allabout,” Whittaker continues. “What we have just been throughwas a dress rehearsal of race day. This is exactly what the 13thof March will be like. We’re in basic training; this is all aboutcranking it up a notch. Come March, we will be the machine,an impeccable machine.”
Whittaker concludes his pep talk and asks for a show ofhands of all those willing to devote every minute of the next fourdays to another grueling attempt to complete a 15-hour, 150-mile autonomous traverse. Fourteen hands shoot up. Sometimebetween the first team meeting eight months ago and today, eachperson in the room had passed his own point of no return.
A Grand Challenge IndeedA P R I L 3 0 , 2 0 0 3 : In a conference room at CMU’s Robotics In-stitute, a tall man rises to his feet. He wears the blue blazer andtan chinos of an academic but has the bravado of a heavyweightwho used to box for the marines. “Welcome to the first meetingof the Red Team,” he booms. “I’m Red Whittaker, director ofthe Fields Robotics Center, and I am committed to leading thisteam to victory in Las Vegas next year.”
Whittaker attended the conference last February at whichofficials from the Defense Advanced Research Projects Agency(DARPA) announced their first-ever prize contest, a robot racefrom Barstow, Calif., to Las Vegas [see box on opposite page].DARPA set up the competition to spur progress toward a vehiclethat could enter a battlefield with minimal human supervision.“It could be delivering supplies or taking out wounded. It couldalso be a tank,” says Anthony J. Tether, the agency’s director.
A different vision moved Whittaker to be among the first ofmore than 100 teams that would sign up to enter the race. Tohim, the principal attractions are the public attention it will bringto robotics and the difficulty of the task, which he often com-pares to Lindbergh’s first transatlantic flight. “The race defiesprevailing technology, and many hold that the challenge prize isunwinnable in our time,” he wrote in an e-mail on March 13to potential volunteers and sponsors.
Building an autonomous robot would not be the hard part.With colleagues at the Robotics Institute, Whittaker has creat-ed self-driving vehicles that haul boulders, harvest crops, mapunderground mines, and hunt for meteorites in Antarctica.What makes the Grand Challenge aptly named is its speed—thespeed at which the robot must move over rough, unfamiliar ter-rain and the haste with which it must be built.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 61
SCO
TT G
OLD
SMIT
H (
top
an
d m
idd
le);
ALE
X G
UTI
ER
RE
Z (b
otto
m)
NOVEMBER 29: Team leader Red Whittaker helps to tackle major problems witha gimbal meant to give the robot a steady gaze despite bounces and bumps.
DECEMBER 2: Sandstorm takes its first independent steps, driving fourmiles in 30 minutes. It reaches a leisurely top speed of 15 miles an hour.
DECEMBER 8: After navigating well for four hours and 46 miles, Sandstormveers off course and into a fence. The next night it rams through the fence.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
“In order to win, Sandstorm will have to average better than10 meters per second [22 miles per hour],” CMU engineer ScottThayer points out. That is roughly 10 times the speed of the pro-totype robots that DARPA has acquired through a four-year,$22-million program to develop unmanned ground vehicles.
“Just getting it to move that fast will be a profoundly chal-lenging problem,” Thayer says. “Maintaining those speeds safe-ly for almost 10 hours straight is just mind-boggling.” He ven-tures that “it will take a fundamental innovation to win. Andthe professional roboticists like me may be the last to come upwith a breakthrough like that. After doing this for decades, wetend to think more incrementally. So who knows—one personwith a dune buggy may win it.”
Blueprint for the Red TeamJ U N E 2 4 : “The last time we met, we considered a tricycle withgiant wheels seven feet in diameter,” Whittaker reports at theteam’s third meeting. “We also looked at a four-wheel-drive,four-wheel-steered vehicle with a chassis that can change shape.
We gave these hard technical looks, but each is too bold a tech-nical step for a yearlong program.”
Three months into that year, the team has not yet decidedwhether to base its robot on a tortoise, such as a militaryHumvee, or on a hare, such as a professional pickup truck or alow-slung Chenowth combat buggy. Whittaker presents a math-ematical analysis of how each vehicle would perform on a coursecomposed mainly of dirt roads and rough trails. “A tough con-sistent vehicle could go 250 miles in 9.3 hours; a sprinter wouldtake 10.6 hours,” he concludes. The choice seems clear, yet itwill be September before they will raise the door on the Plane-tary Robotics Building, where the team has set up shop, andpush in a 1986 Hummer M998.
But the group—which now numbers more than 50, thanksto the dozens of CMU graduate and undergraduate studentsworking on the project for credit—has prepared a 58-page tech-nical paper describing how Sandstorm will track its position,plan its route, and detect and avoid hazards in its way. AlexGutierrez, one of the graduate students at the core of the team,
Planning to Win
62 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
W.
WAY
T G
IBB
S; S
OU
RC
ES:
USG
S, T
HE
RE
D T
EAM
THE RED TEAM concluded early on that the most feasible wayto win the race is to give the Sandstorm robot an extremelydetailed and accurate navigational plan to guide it over the raceroute. The exact course will be held secret until two hours beforethe starting gun, however. So the team has spent thousands ofhours assembling maps, models and aerial imagery of the entirepotential race area, which spans 400 times the area shown inthis illustration. The engineers overlay, align and hand-correctseveral distinct views of the terrain.
From the U.S. Geological Survey, the team obtainedrelatively rough three-dimensional profiles of the land andaerial photography that can distinguish objects as small as one meter. To these they add custom-made road andvegetation maps, then fuse these layers of information into
an enormous geographic database several terabytes in size. A computer program can use this database to calculate the
“cost” for Sandstorm to traverse every square meter in theregion. Some areas, such as cliffs or course boundaries, have aninfinite cost because they would disable or disqualify the racer.Dry lake beds, in contrast, might have a cost of zero.
On race day, the actual course data (simulated below ascircles and blue lines) will be sent through a high-speed link tothe Red Team’s control center. There a fleet of computers willuse the cost map to compute the optimal route. A dozen or moretrained volunteers will then divide the route into sections andwill tweak the computed plan as needed so that it does notmistakenly send Sandstorm into harm’s way. The finalnavigation instructions (yellow dots) will be beamed to therobot shortly before the race begins.
TERRAIN ELEVATION MODEL
AERIAL IMAGERY
ROADS ANDVEGETATION
COMPOSITE ROUTE MAP
3 miles
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
hands out copies to executives from SAIC, Boeing, Caterpillar,Seagate and other corporate partners as they enter the room.
“First we will work for eight months to create the best pos-sible maps of the course terrain,” Whittaker explains. “WhenDARPA hands out the race route, two hours before the racestarts, we will use those maps to calculate the optimal route anddo a simulated flight though it” [see box on opposite page]. Theresulting list of thousands of GPS coordinates will be copied tocomputers on the robot, giving it “little seeds of corn to aim forevery meter or so,” Whittaker says. “Sandstorm will just goalong like Pac-Man, gobbling up these little virtual dots.”
The budget now sums to an astonishing bottom line:$3,539,491. Nearly $2.5 million of that is for personnel ex-penses that will probably never get paid. The $725,000 for thevehicle itself is not optional, however, and so far only Caterpil-lar and a local foundation have written checks. But many oth-ers are donating valuable equipment and expertise.
Applanix, for example, delivered a $60,000 position-track-ing system that not only will allow Sandstorm to know whereit is as it bounces along the desert but also will help it to solveone of the toughest problems in mobile robotics: watchingwhere it is going with a steady gaze. “It will know what theworld outside looks like through lasers, what it looks like inradar, and what it looks like through a stereo, or two-eyed, cam-era—provided by our good friends at SAIC,” Whittaker de-clares. Each of these sensors will be mounted on motorized plat-forms connected to the Applanix system in a tight feedbackloop. These gimbals, as engineers call them, will compensate forthe motion of the vehicle much like the neck and eye muscles ofa human driver [see box on next two pages].
Many of the competing teams have similar plans. One com-posed of undergraduates at the California Institute of Technol-ogy is forgoing radar and relying heavily on four video camerasmounted to the front of their modified Chevrolet Tahoe. TheRed Team’s Navtech radar is worth its $47,000 price because“it works through dust, which can blind the other sensors,”Whittaker says. For that very reason, Ohio State University’sTeam Terramax is mounting two radars—plus six cameras andfour laser scanners—on the robot it is building from a huge six-wheeled Oshkosh truck.
More sensors are not necessarily better. Each one streamsdata like a fire hose; too many can choke a robot’s computers.As the vehicle jolts and shakes, overlapping scans may confusemore than they inform. And merging sensor data of differenttypes is notoriously tricky. Laser scanners produce “pointclouds,” radars emit rectangular blips, a stereo camera gener-ates a so-called disparity map. “If you aren’t careful,” says JayGowdy, a CMU scientist on the Red Team, “you can end upcombining the weaknesses of each sensor instead of combiningtheir strengths.”
Reality Checks InN O V E M B E R 6 : Whittaker, Urmson and Philip Koon, one of twoengineers that Boeing Phantom Works has embedded with theteam, sit down for the weekly teleconference with the team’s
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 63
SCO
TT G
OLD
SMIT
H (
top
); T
AKE
HIT
O E
TAN
I (m
idd
le);
JAS
ON
DIV
EN
ER
E (
bot
tom
)
DECEMBER 9: As a cold rain begins to fall outside the workshop, teammembers scramble to prepare the vehicle for another long night of testing.
DECEMBER 10: Sleep-deprived and frustrated, Chris Urmson and KevinPeterson struggle to debug the robot’s hardware and software.
DECEMBER 18: Engineers from Boeing Phantom Works join part of the teamin the Mojave Desert to test an innovative radar system for Sandstorm.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
How Sandstorm Works JUST BEFORE THE RACE BEGINS, the Red Team will calculate
the best route and send a detailed itinerary (in the form ofgeographic coordinates for every meter of the course) to theSandstorm robot. The vehicle will try to follow this virtual trail of
breadcrumbs from the starting line to the finish as closely as itcan, while detecting and avoiding any unexpected obstacles,such as a disabled racer in the road ahead. To succeed, the robotmust solve four challenging problems.
Applanix navigation
computer
GPS trace of 119-mile test(red lines are sensor glitches)
3-Ghz dual-processor Xeon computers
Four-processor Itanium2 computer
Ultrapreciseodometer
Long-rangescanning laser
Air knife
Stereo videocamera
Short-rangescanning laser
Short-rangescanning laser
Radiatorguard
Electronics box(E-box)
Heavy-dutyshock absorbers
Rotating radarantenna
GPS antennas
b c
Trail
Vegetation
LONG-RANGE LASER10- to 350-meter range;default focus at 50 metersSTEREO CAMERA
30-meter range
RADAR10- to 100-meter range
SHORT-RANGE LASER10- to 30-meter range
2. PERCEIVING AN OBSTACLESandstorm uses four kinds of sensors to look for obstacles (a).A long-range laser traces the profile of the terrain 50 times asecond. Successive profiles build up to form a 3-D model (b).Shorter-range lasers also cover all sides of the vehicle. A stereo camera sends video to a dedicated computer thatestimates the slope and roughness of the ground. A rotatingradar antenna will pick up obstructions (c) even when dust orglare blinds the other sensors.
64 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
1. TRACKING ITS POSITIONAn Applanix navigation computer contains two GPS receivers,three fiber-optic gyroscopes, three silicon accelerometers andan ultraprecise odometer, which it uses to pin down the robot’sposition to within 50 centimeters and to measure its orientationin space to 0.4 degree. The system updates the robot’s sense of where it is 200 times a second.
5-kilowattMechron dieselgenerator
Car
Coil springs andshock absorbers
a
Dedicated video
processor
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
partners. “We were maybe 50–50 on our goals this week—thisis the first time we have really missed the mark,” Whittaker an-nounces. The radar was hung up in customs en route from theU.K. After more than 100 hours of work, the mapping grouphas completed less than 4 percent of the area they aim to cover.And money is getting tight. “At the moment, we’re short about$950,000 and burning through eight grand a day,” Whittakerreports. He hopes to sell advertising space on the robot’s hoodand fin for half a million dollars but has found no buyers.
Two weeks later the team meets to confront other problems.A superprecise optical odometer built to slide on the robot’s axledoes not fit together properly. “And this is troubling,” Whit-taker says as he points to a large spike on a graph of how thecomputer cage—they call it the E-box—bounced around as thevehicle ran over a railroad tie at five miles an hour. “That readsseven g’s, which is very bad,” he continues. Hard disks will crashand chips may pop from their sockets unless the E-box is iso-lated from all shocks greater than about three g’s. They must fig-ure out a better way to suspend the E-box within the chassis.
“Engineering is always a series of failures to get to success,”points out Bryon Smith, one of the few seasoned roboticists onthe team. “It takes iteration after iteration to get it right.” But it-erations take time. The 100 days that Whittaker scheduled fordevelopment are almost up, and the team has yet to install andwire all the onboard computers, construct the gimbals, finish thesoftware or mount the sensors.
“This vehicle hasn’t rolled so much as a foot under its owncontrol,” Whittaker says. “You have promised to get 150 mileson that beast in two weeks. Just so we’re clear on the ambitionhere: DARPA’s Spinner vehicle program, based right here atCMU, has a team of pros and a budget of $5 million and is nowin its second year. So far the furthest it has driven is 15 miles.Okay, anyone who thinks it is not appropriate for us to go for150 miles by December 10, raise your hand.” No one does.“There it is,” he smiles. “We’re now heading into that violentand wretched time of birthing this machine and launching it onits maiden voyage.”
D E C E M B E R 1 : “There were a bunch of us here all day onThanksgiving and through the weekend—me, Alex, Philip, Yu[Kato] and several others. But it was worth it,” Urmson says. Soends any semblance of normal life as these young engineers aredrawn into their leader’s constructive obsession. “Around 3 or4 A.M. Sunday morning, as all the pieces started coming togeth-er and getting connected, it felt damn good,” Whittaker adds,casting critical looks at those who spent the holiday with theirfamilies.
The robot now has several of its sensory organs attached anda rudimentary nervous system working. Smith and Kato haveassembled the three-axis gimbal that will aim and steady thestereo camera and long-range laser only to discover “verystrange behavior with the fiber-optic gyroscopes” that measurethe device’s motion, Smith reports. Whittaker listens intentlyto the details. “The gimbal is an essential device to win the race,”he reminds the team. “Its main purpose is to suppress jitter.Right now when we turn it on, it induces jitter.” For the next
4. ENDURING THE DUST AND BUMPSBack roads through the Mojave are rough, so the team hasequipped the Humvee with racing shocks and springs, a radiatorguard and run-flat wheels. To protect the computers, the electronicsbox is suspended on tripods of spring-reinforced shock absorbers
and strapped in place by superstrong bungee cords. A dozen“ruggedized” hard disks inside will operate in redundant
pairs. As Sandstorm bounces over a washboard dirt roadat 30 miles an hour, it must hold its forward sensorssteady. Red Team engineers built a computer-
controlled stabilizer, or gimbal (above), that bothaims and steadies the camera and long-range laser.
The gimbal uses three fiber-optic gyroscopes andthree precise actuators to measure and compensate
for the vehicle’s pitch, roll and yaw. The radar issimilarly bolted to a one-axis gimbal.
Fiber-opticgyroscope
Stereo videocamera
3. REVISING ITS ROUTEEven the best maps are not up to the minute. So three onboard Xeon
computers will use data from each sensor to update the “cost”assigned to each square meter in the area. A paved road carries a cost ofzero; a cliff or competing racer warrants an infinite cost. Several times asecond, a fourth Itanium2 computer checks whether the “breadcrumbtrail” (d, yellow dots) passes through high-cost territory ( red areas). Ifso, the planner program prices alternative routes (blue arcs) and shiftsthe breadcrumbs to the shortest safe path (e).
d e
Obstacle
Alternateroutes
Preplanned route
Correctedroute
Harmonic driveactuator
Long-rangelaser scanner
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 65
BR
YAN
CH
RIS
TIE
DE
SIG
N (
illu
stra
tion
s);
SOU
RC
E:
THE
RE
D T
EAM
Controlcomputer
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
week, Kato will hardly leave the shop as he valiantly attemptsto correct through software a fundamental flaw in the gyroscopehardware.
At 7:51 the next evening, after handling well in a piloted testrun, Sandstorm is allowed to take its own wheel. It is drivingblind, simply following a recorded list of GPS waypoints thattrace an oval loop. The computer is doing the steering, but Urm-son is on board as a “human kill switch,” to hit the emergencystop button if something goes wrong. Four miles and half anhour later programmer Kevin Peterson clicks a button on hislaptop, a command travels wirelessly to the robot, and Sand-storm brakes to a halt. “Very well done,” Whittaker congratu-lates, and sends the vehicle back to the shop for another nightof modifications.
“From now on we need everybody here 24 hours so that assoon as the vehicle returns from the field, people jump on it andstart working,” Whittaker says in the morning. “It is exciting tosee a robot first spring into action. But the point is to make thiskind of driving boring. A 150-mile traverse in the next five days,while taking sensor data: that’s the final exam, and it’s pass/fail.”
D E C E M B E R 8 : The Red Team has set up camp by the emptyblast furnaces to watch the robot make its 15-hour nonstop, un-guided journey. They record a figure-eight test path, but the ma-chine gets confused at the crossing point; sometimes it goes leftand sometimes right. So they go back to the oval loop.
But before the test can begin, a short circuit sends current
surging through a wireless “E-stop” receiver that DARPA hasprovided so that race officials can disable any robot that goesberserk. With that receiver fried, the team has no fail-safe wayto force Sandstorm to stop—only a piece of software. Petersonand Martin Stolle, two of the team’s software gurus, urge Whit-taker not to rely on the software.
Urmson arrives with a servomotor borrowed from a radio-controlled airplane and proceeds to jury-rig a wireless killswitch. But that transmitter also shorts out. “So now we havejust Martin’s software stop,” Whittaker sighs. “Martin, howmany hours do you have on your controller?” he asks.
“We’ve tested it for about half an hour,” Martin replies.Moreover, he warns, if the onboard computer fails, “we will loseall control, and the vehicle will just plow ahead until it hits anobstacle larger than a Humvee.”
Urmson huddles the team together. “We can go ahead, butwe all need to understand and agree that—”
“Everyone understands it, and I’m accountable,” Red inter-rupts. “It’s not a question of pros or cons; we’re going to do it.”The sun has set, and the slush on the track is refreezing. Whit-taker insists that two team members stay in the open to keepwatch as the robot drives 792 laps around its short test loop.
With a puff of gray smoke, Sandstorm zooms forward. As itrounds the first two turns and enters a straightaway, sparks ap-pear in the undercarriage. It skids to a stop on command, andteam members sprint out with a fire extinguisher. The cause is
66 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
TEAM
PH
ANTA
SM (
ATV)
; TE
AM C
YBE
RR
IDE
R (
du
ne
bu
gg
y);
AXI
ON
RAC
ING
(SU
V);
TEAM
TE
RR
AMAX
(m
ilit
ary
veh
icle
)
Modified All-Terrain VehiclesPros: Inexpensive; off-road suspensionsare standard; can stop, turn andaccelerate quickly; small size provides amargin of error on narrow trails.Cons: Sensors are low and thus limited intheir range of view; high risk of critical
damage in a collision; very limited ability to generate electricalpower; small fuel tanks; overturn easily.Teams: ENSCO, Phantasm ( pictured), Virginia Tech
Modified Sport-Utility VehiclesPros: Easily acquired; good groundclearance; large enclosed interior forelectronics; powerful engines; highmounting points for sensors.Cons: Expensive; high rollover risk;complex electrical system; suspension
is designed for paved roads rather than trails.Teams: Arctic Tortoise, Axion Racing (pictured), Caltech, Digital Auto Drive, Insight Racing, Navigators, Overbot, Palos Verdes Road Warriors
Dune BuggiesPros: Very low center of gravity preventsoverturning; frame and suspension arecustomized for desert racing;lightweight, agile and fast.Cons: Sensors are low and vulnerable tocollisions and dust; small wheels; low
mass and electrical budgets limit onboard computing.Teams: AI Motorvator, CyberRider ( pictured), LoGHIQ, Sciautonics(which is fielding two robots)
Modified Military VehiclesPros: Very high ground clearance,stability and crash tolerance; powerful engines and large chassis can easily carry a large payload ofelectronics and computers; high vantagepoint for sensors.
Cons: Expensive and hard to obtain; parts are difficult to find;stiff suspension creates problems for sensors; wide turningradius; relatively slow acceleration and braking. Teams: The Red Team, Terramax (pictured)
The CompetitorsMORE THAN 100 TEAMS registered for the Grand Challenge; 86 sent technical applications to DARPA, which approved 45. DARPA
officials later culled the field to 25 vehicles, which fall into roughly four categories. No more than 20 will be allowed to race.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
innocuous: someone had forgotten to refill a gas cylinder thatkeeps the parking brake released, so it was driving with its brakeon. They push the vehicle back onto the course, only to find thatthe batteries have failed.
And so it went for the next several days, with one thing af-ter another going wrong. While Smith and Kato managed toconquer the bugs in the gimbal and get one of its three armsworking for 15 hours, gremlins bedeviled the rest of the RedTeam, sending Sandstorm careening into, and later through, achain-link fence. In the wee hours of December 13, the robotwas just clearing 119 miles when it headed for the hills and hadto be stopped. They persisted for two more days, through asnowstorm and bitter cold, persisted and failed.
Sprinting to the Starting LineD E C E M B E R 2 1 : “We didn’t do the 150,” Whittaker acknowl-edges, as the diehards meet to take stock. “But it was a hell of afour days. It was our battle cry, and it was magnificent.”
On Christmas Eve a new shock isolation design for the E-box is tested. It works, as do all three arms on the gimbal. Christ-mas Day brings—what else?—test, fail, rework, repeat.
Within two weeks, as industry partners fly in for the last fullteam meeting on January 6, the robot is ready for its public un-veiling before politicians and television cameras. Behind closeddoors, Whittaker acknowledges that “in the last six monthswe’ve fallen behind a month. Following GPS waypoints, the ve-hicle is now rock-solid, to the point where you can turn yourback on it.” Sandstorm has graduated from a paved lot to anopen field, where it now safely drives by itself at more than 30miles an hour.
But although the machine can see the world, it cannot yet
reason enough to avoid obstacles. Even with 10 of the mostpowerful processors that Intel makes installed in Sandstorm, thecomputers formulate their plans about a third too slowly.
In February the robot and its creators will head to the desert.“We need to put 10,000 miles of testing on it,” Whittaker says.“This fancy stuff could shake apart because it’s all prototype.Just inside the E-box there are 5,000-odd components, a failurein any one of which could screw us up. Any team could beat us.”
And if the Red Team wins? The best thing about building anew race of robots, Whittaker said one frigid night in Decem-ber as we watched Sandstorm do its laps beneath a nearly per-fect full moon, is not the act of creation. “What’s most fun is ex-ploring the space of possibilities you have opened with your in-vention. I’m thinking about proposing a mission to NASA tolaunch a lunar rover that could circumnavigate the pole of themoon, searching for ice.” Other team members have suggestedbuilding a robot to run the Iditarod in Alaska or to serve as anambulance in Antarctica.
More likely, however, the $1-million prize will go unclaimedthis year and the contest will repeat in 2005. “If no one wins thisrace and we recommit for next year, who’s in?” Whittaker asksat the end of the meeting. Up go a roomful of hands.
Senior writer W. Wayt Gibbs has been in Pittsburgh coveringthe progress of the Red Team since March 2003.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 67
W.
WAY
T G
IBB
S
JANUARY 20: Sandstorm grows faster, smarter and more robust almost every day. Yet Whittaker still gives it only 40 percent odds of finishing the race.
The Red Team Web site: redteamracing.org
For links to other teams: www.darpa.mil/grandchallenge/teams.htm
For more information on the Grand Challenge race:www.darpa.mil/grandchallenge/
M O R E T O E X P L O R E
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
Global warming is real, and Nevertheless, practical actions, which
ICEBERG BREAKS OFF the San Rafael Glacier in Chile. Global disintegration of ice masses has the potentialto raise sea level by several meters or more. Thegrim consequences of a rising sea level set a lowthreshold for how much the planet can warm withoutdisrupting human society.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
TIME BOMB
the consequences are potentially disastrous. would also yield a cleaner, healthier atmosphere, could slow, and eventually stop, the process
BY JAMES HANSEN
GlobalWarming
Defusingthe
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
became strikingly apparent to me one summerafternoon in 1976 on Jones Beach, Long Island. Arriving at midday, my wife, son and I founda spot near the water to avoid the scorchinghot sand. As the sun sank in the late after-noon, a brisk wind from the ocean whipped
up whitecaps. My son and I had goose bumps as we ran alongthe foamy shoreline and watched the churning waves.
That same summer Andy Lacis and I, along with other col-leagues at the NASA Goddard Institute for Space Studies, had es-timated the effects of greenhouse gases on climate. It was wellknown by then that human-made greenhouse gases, especiallycarbon dioxide and chlorofluorocarbons (CFCs), were accu-mulating in the atmosphere. These gases are a climate “forcing,”a perturbation imposed on the energy budget of the planet. Likea blanket, they absorb infrared (heat) radiation that would oth-erwise escape from the earth’s surface and atmosphere to space.
Our group had calculated that these human-made gaseswere heating the earth’s surface at a rate of almost two wattsper square meter. A miniature Christmas tree bulb dissipatesabout one watt, mostly in the form of heat. So it was as if hu-mans had placed two of these tiny bulbs over every square me-ter of the earth’s surface, burning night and day.
The paradox that this result presented was the contrast be-tween the awesome forces of nature and the tiny lightbulbs.Surely their feeble heating could not command the wind andwaves or smooth our goose bumps. Even their imperceptibleheating of the ocean surface must be quickly dissipated to greatdepths, so it must take many years, perhaps centuries, for theultimate surface warming to be achieved.
This seeming paradox has now been largely resolvedthrough study of the history of the earth’s climate, which re-veals that small forces, maintained long enough, can cause large
climate change. And, consistent with the historical evidence,the earth has begun to warm in recent decades at a rate pre-dicted by climate models that take account of the atmosphericaccumulation of human-made greenhouse gases. The warmingis having noticeable impacts as glaciers are retreating world-wide, Arctic sea ice has thinned, and spring comes about oneweek earlier than when I grew up in the 1950s.
Yet many issues remain unresolved. How much will climatechange in coming decades? What will be the practical conse-quences? What, if anything, should we do about it? The debateover these questions is highly charged because of the inherenteconomic stakes.
Objective analysis of global warming requires quantitativeknowledge of three issues: the sensitivity of the climate systemto forcings, the forcings that humans are introducing, and thetime required for climate to respond. All these issues can bestudied with global climate models, which are numerical sim-ulations on computers. But our most accurate knowledge aboutclimate sensitivity, at least so far, is based on empirical datafrom the earth’s history.
The Lessons of HistoryOVER THE PAST few million years the earth’s climate hasswung repeatedly between ice ages and warm interglacial pe-riods. A 400,000-year record of temperature is preserved in theAntarctic ice sheet, which, except for coastal fringes, escapedmelting even in the warmest interglacial periods. This record[see box on opposite page] suggests that the present interglacialperiod (the Holocene), now about 12,000 years old, is alreadylong of tooth.
The natural millennial climate swings are associated withslow variations of the earth’s orbit induced by the gravity ofother planets, mainly Jupiter and Saturn (because they are soheavy) and Venus (because it comes so close). These perturba-tions hardly affect the annual mean solar energy striking theearth, but they alter the geographical and seasonal distributionof incoming solar energy, or insolation, as much as 20 percent.The insolation changes, over long periods, affect the buildingand melting of ice sheets.
Insolation and climate changes also affect uptake and releaseof carbon dioxide and methane by plants, soil and the ocean.Climatologists are still developing a quantitative understandingof the mechanisms by which the ocean and land release carbondioxide and methane as the earth warms, but the paleoclimatedata are already a gold mine of information. The most criticalinsight that the ice age climate swings provide is an empiricalmeasure of climate sensitivity.
The composition of the ice age atmosphere is known pre-cisely from air bubbles trapped as the Antarctic and Greenlandice sheets and numerous mountain glaciers built up from an-nual snowfall. Furthermore, the geographical distributions ofthe ice sheets, vegetation cover and coastlines during the ice ageare well mapped. From these data we know that the change of
paradox in the notion of human-made global warming
70 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
RIC
ER
GE
NB
RIG
HT
Cor
bis
(p
rece
din
g p
ag
es)
■ At present, our most accurate knowledge about climatesensitivity is based on data from the earth’s history, andthis evidence reveals that small forces, maintained longenough, can cause large climate change.
■ Human-made forces, especially greenhouse gases, sootand other small particles, now exceed natural forces, andthe world has begun to warm at a rate predicted byclimate models.
■ The stability of the great ice sheets on Greenland andAntarctica and the need to preserve global coastlines seta low limit on the global warming that will constitute“dangerous anthropogenic interference” with climate.
■ Halting global warming requires urgent, unprecedentedinternational cooperation, but the needed actions arefeasible and have additional benefits for human health,agriculture and the environment.
Overview/Global Warming
A
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
climate forcing between the ice age and today was about 6.5watts per square meter. This forcing maintains a global tem-perature change of 5 degrees Celsius (9 degrees Fahrenheit), im-plying a climate sensitivity of 0.75 ± 0.25 degrees C per wattper square meter. Climate models yield a similar climate sensi-tivity. The empirical result is more precise and reliable, how-ever, because it includes all the processes operating in the realworld, even those we have not yet been smart enough to includein the models.
The paleodata provide another important insight. Changesof the earth’s orbit instigate climate change, but they operate byaltering atmosphere and surface properties and thus the plane-tary energy balance. These atmosphere and surface propertiesare now influenced more by humans than by our planet’s orbitalvariations.
Climate-Forcing Agents TodayTHE LARGEST change of climate forcings in recent centuriesis caused by human-made greenhouse gases. Greenhouse gas-es in the atmosphere absorb heat radiation rather than lettingit escape into space. In effect, they make the proverbial blan-ket thicker, returning more heat toward the ground rather thanletting it escape to space. The earth then is radiating less ener-gy to space than it absorbs from the sun. This temporary plan-
etary energy imbalance results in the earth’s gradual warming. Because of the large capacity of the oceans to absorb heat,
it takes the earth about a century to approach a new balance—
that is, for it to once again receive the same amount of energyfrom the sun that it radiates to space. And of course the balanceis reset at a higher temperature. In the meantime, before itachieves this equilibrium, more forcings may be added.
The single most important human-made greenhouse gas iscarbon dioxide, which comes mainly from burning fossil fuels(coal, oil and gas). Yet the combined effect of the other human-made gases is comparable. These other gases, especially tro-pospheric ozone and its precursors, including methane, are in-gredients in smog that damage human health and agriculturalproductivity.
Aerosols (fine particles in the air) are the other main hu-man-made climate forcing. Their effect is more complex. Some“white” aerosols, such as sulfates arising from sulfur in fossilfuels, are highly reflective and thus reduce solar heating of theearth; however, black carbon (soot), a product of incompletecombustion of fossil fuels, biofuels and outdoor biomass burn-ing, absorbs sunlight and thus heats the atmosphere. Thisaerosol direct climate forcing is uncertain by at least 50 per-cent, in part because aerosol amounts are not well measuredand in part because of their complexity.
Eemian interglacial Holocene
Carb
on D
ioxi
de(p
arts
per
mill
ion) 280
240
200
700
600
500
400
Met
hane
(par
ts p
er b
illio
n)Te
mpe
ratu
re C
hang
e (º
C)(r
elat
ive
to p
ast m
illen
nium
)
2
0
–2
–4
–6
–8
Thousands of Years before Present
400 350 300 250 200 150 100 50 0
JEN
CH
RIS
TIAN
SEN
; SO
UR
CE
: J.
R.
PE
TIT
ET
AL.
IN N
ATU
RE,
VO
L. 3
99
, P
AGE
S 4
29
–4
36
; JU
NE
3,
19
99
(g
rap
hs)
;C
SIR
O/S
IMO
N F
RAS
ER
Ph
oto
Res
earc
her
s, I
nc.
(p
hot
ogra
ph
)
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 71
400,000 YEARS OF CLIMATE CHANGEANTARCTIC ICE has preserved a 400,000-year record of temperature and of levels of carbon dioxide andmethane in the atmosphere. Scientists study gasestrapped in air bubbles in the ice—generally using icecores ( photograph) extracted from the ice sheet andtransported to a laboratory. The historical recordprovides us with two critical measures: Comparison ofthe current interglacial period (the Holocene) with themost recent ice age (20,000 years ago) gives anaccurate measure of climate sensitivity to forcings.The temperature in the previous interglacial period (the Eemian), when sea level was several metershigher than today, defines an estimate of the warmingthat today’s civilization would consider to bedangerous anthropogenic interference with climate.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
72 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
JEN
CH
RIS
TIAN
SEN
; SO
UR
CE
: JA
ME
S H
ANSE
N
Aerosols also cause an indirect climate forcing by altering theproperties of clouds. The resulting brighter, longer-lived cloudsreduce the amount of sunlight absorbed by the earth, so the in-direct effect of aerosols is a negative forcing that causes cooling.
Other human-made climate forcings include replacement offorests by cropland. Forests are dark even with snow on theground, so their removal reduces solar heating.
Natural forcings, such as volcanic eruptions and fluctua-tions of the sun’s brightness, probably have little trend on atimescale of 1,000 years. But evidence of a small solar bright-ening over the past 150 years implies a climate forcing of a fewtenths of a watt per square meter.
The net value of the forcings added since 1850 is 1.6 ± 1.0watts per square meter. Despite the large uncertainties, thereis evidence that this estimated net forcing is approximately cor-rect. One piece of evidence is the close agreement of observedglobal temperature during the past several decades with climatemodels driven by these forcings. More fundamentally, the ob-served heat gain by the world ocean in the past 50 years is con-sistent with the estimated net climate forcing.
Global WarmingGLOBAL AVERAGE surface temperaturehas increased about 0.75 degree C during theperiod of extensive instrumental measure-ments, which began in the late 1800s. Mostof the warming, about 0.5 degree C, occurredafter 1950. The causes of observed warmingcan be investigated best for the past 50 years,because most climate forcings were observedthen, especially since satellite measurementsof the sun, stratospheric aerosols and ozonebegan in the 1970s. Furthermore, 70 percentof the anthropogenic increase of greenhousegases occurred after 1950.
The most important quantity is the plan-etary energy imbalance [see box on page75]. This imbalance is a consequence of thelong time that it takes the ocean to warm. We conclude that theearth is now out of balance by something between 0.5 and onewatt per square meter—that much more solar radiation is be-ing absorbed by the earth than is being emitted as heat to space.Even if atmospheric composition does not change further, theearth’s surface will therefore eventually warm another 0.4 to0.7 degree C.
Most of the energy imbalance has been heat going into theocean. Sydney Levitus of the National Oceanic and Atmospher-ic Administration has analyzed ocean temperature changes ofthe past 50 years, finding that the world ocean heat content in-creased about 10 watt-years per square meter in the past 50years. He also finds that the rate of ocean heat storage in recentyears is consistent with our estimate that the earth is now outof energy balance by 0.5 to one watt per square meter. Notethat the amount of heat required to melt enough ice to raise sealevel one meter is about 12 watt-years (averaged over the plan-et), energy that could be accumulated in 12 years if the planetis out of balance by one watt per square meter.
The agreement with observations, for both the modeledtemperature change and ocean heat storage, leaves no doubtthat observed global climate change is being driven by naturaland anthropogenic forcings. The current rate of ocean heatstorage is a critical planetary metric: it not only determines theamount of additional global warming already in the pipeline,but it also equals the reduction in climate forcings needed tostabilize the earth’s present climate.
The Time BombTHE GOAL OF the United Nations Framework Conventionon Climate Change, produced in Rio de Janeiro in 1989, is tostabilize atmospheric composition to “prevent dangerous an-thropogenic interference with the climate system” and to
GREENHOUSE GASES AEROSOLS
Carb
on d
ioxi
de
Met
hane
Chlo
roflu
oroc
arbo
ns
Nitr
ous
oxid
e
Ozon
e
Blac
k ca
rbon
iRe
flect
ve a
eros
ols
Land
cov
er c
hang
es Sun
Forc
ing
Agen
tW
atts
per
Squ
are
Met
er 1
0
–1
Clou
d dr
ople
t cha
nges
small forces, maintained long enough, can cause large climate change
CLIMATE FORCINGS
A CLIMATE FORCING is a mechanism that alters the globalenergy balance. A forcing can be natural—fluctuations inthe earth’s orbit, for example—or human-made, such asaerosols and greenhouse gases. Human-made climateforcings now dominate natural forcings. Carbon dioxide isthe largest forcing, but air pollutants (black carbon, ozone,methane) together are comparable. (Aerosol effects are notknown accurately.)
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
achieve that goal in ways that do not disrupt the global econo-my. Defining the level of warming that constitutes “dangerousanthropogenic interference” is thus a crucial but difficult partof the problem.
The U.N. established an Intergovernmental Panel on Cli-mate Change (IPCC) with responsibility for analysis of globalwarming. The IPCC has defined climate-forcing scenarios, usedthese for simulations of 21st-century climate, and estimated theimpact of temperature and precipitation changes on agricul-ture, natural ecosystems, wildlife and other matters. The IPCCestimates sea-level change as large as several tens of centimetersin 100 years, if global warming reaches several degrees Celsius.The group’s calculated sea-level change is due mainly to ther-mal expansion of ocean water, with little change in ice-sheetvolume.
These moderate climate effects, even with rapidly increas-ing greenhouse gases, leave the impression that we are not closeto dangerous anthropogenic interference. I will argue, howev-er, that we are much closer than is generally realized, and thusthe emphasis should be on mitigating the changes rather thanjust adapting to them.
The dominant issue in global warming, in my opinion, issea-level change and the question of how fast ice sheets can dis-integrate. A large portion of the world’s people live within a fewmeters of sea level, with trillions of dollars of infrastructure.The need to preserve global coastlines sets a low ceiling on thelevel of global warming that would constitute dangerous an-thropogenic interference.
The history of the earth and the present human-made plan-etary energy imbalance together paint a disturbing pictureabout prospects for sea-level change. Data from the Antarctictemperature record show that the warming of the past 50 yearshas taken global temperature back to approximately the peak
of the current interglacial (the Holocene). There is some addi-tional warming in the pipeline that will take us about halfwayto the highest global temperature level of the previous inter-glacial (the Eemian), which was warmer than the Holocene,with sea level estimated to have been five to six meters higher.One additional watt per square meter of forcing, over andabove that today, will take global temperature approximatelyto the maximum level of the Eemian.
The main issue is: How fast will ice sheets respond to glob-al warming? The IPCC calculates only a slight change in the icesheets in 100 years; however, the IPCC calculations include onlythe gradual effects of changes in snowfall, evaporation and melt-ing. In the real world, ice-sheet disintegration is driven by high-ly nonlinear processes and feedbacks. The peak rate of deglacia-tion following the last ice age was a sustained rate of melting of more than 14,000 cubic kilometers a year—about one meter of sea-level rise every 20 years, which was maintained for sev-eral centuries. This period of most rapid melt coincided, as well
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 73
RAN
DY
HAR
RIS
HUMAN-MADE climate forcings, mainly greenhouse gases, heat the earth’ssurface at a rate of about two watts per square meter—the equivalent of two tiny one-watt bulbs burning over every square meter of the planet.The full effect of the warming is slowed by the ocean, because it canabsorb so much heat. The ocean’s surface begins to warm, but before itcan heat up much, the surface water is mixed down and replaced by colderwater from below. Scientists now think it takes about a century for theocean to approach its new temperature.
JAMES HANSEN is director of the NASA Goddard Institute for SpaceStudies and a researcher at the Columbia University Earth Insti-tute. He received his Ph.D. in physics and astronomy from theUniversity of Iowa, where he studied under James Van Allen.Hansen is best known for his testimony to congressional com-mittees in the 1980s that helped to raise awareness of the glob-al warming issue.
THE
AU
THO
R
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
as can be measured, with the time of most rapid warming.Given the present unusual global warming rate on an already
warm planet, we can anticipate that areas with summer melt andrain will expand over larger areas of Greenland and fringes ofAntarctica. Rising sea level itself tends to lift marine ice shelvesthat buttress land ice, unhinging them from anchor points. Asice shelves break up, this accelerates movement of land ice to theocean. Although building of glaciers is slow, once an ice sheetbegins to collapse, its demise can be spectacularly rapid.
The human-induced planetary energy imbalance providesan ample supply of energy for melting ice. Furthermore, this en-ergy source is supplemented by increased absorption of sunlightby ice sheets darkened by black-carbon aerosols, and the pos-itive feedback process as meltwater darkens the ice surface.
These considerations do not mean that we should expectlarge sea-level change in the next few years. Preconditioning ofice sheets for accelerated breakup may require a long time, per-haps many centuries. (The satellite ICESat, recently launched byNASA, may be able to detect early signs of accelerating ice-sheetbreakup.) Yet I suspect that significant sea-level rise could be-gin much sooner if the planetary energy imbalance continues
74 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
ON A SLIPPERY SLOPE to disaster, a stream of snowmelt cascades down a moulin on the Greenland ice sheet during a recent summer. The moulin, a near-vertical shaft worn in the ice by surface water, carries water to the base of the ice sheet. There the water is a lubricating fluid thatspeeds motion and disintegration of the ice sheet. Ice sheet growth is a slow, dry process, inherently limited by the snowfall rate, butdisintegration is a wet process, driven by positive feedbacks, and oncewell under way it can be explosively rapid.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 75
JEN
CH
RIS
TIAN
SEN
; SO
UR
CE
: JA
ME
S H
ANSE
N
to increase. It seems clear that global warming beyond somelimit will make a large sea-level change inevitable for futuregenerations. And once large-scale ice-sheet breakup is underway, it will be impractical to stop. Dikes may protect limitedregions, such as Manhattan and the Netherlands, but most ofthe global coastlines will be inundated.
I argue that the level of dangerous anthropogenic influenceis likely to be set by the global temperature and planetary ra-diation imbalance at which substantial deglaciation becomespractically impossible to avoid. Based on the paleoclimate evi-dence, I suggest that the highest prudent level of additionalglobal warming is not more than about one degree C. Thismeans that additional climate forcing should not exceed aboutone watt per square meter.
Climate-Forcing ScenariosTHE IPCC defines many climate-forcing scenarios for the 21stcentury based on multifarious “story lines” for populationgrowth, economic development and energy sources. It estimatesthat added climate forcing in the next 50 years is one to threewatts per square meter for carbon dioxide and two to four wattsper square meter with other gases and aerosols included. Eventhe IPCC’s minimum added forcing would cause dangerous an-thropogenic interference with the climate system based on ourcriterion.
The IPCC scenarios may be unduly pessimistic, however.First, they ignore changes in emissions, some already under way,because of concerns about global warming. Second, they assumethat true air pollution will continue to get worse, with ozone,methane and black carbon all greater in 2050 than in 2000.Third, they give short shrift to technology advances that can re-duce emissions in the next 50 years.
An alternative way to define scenarios is to examine currenttrends of climate-forcing agents, to ask why they are changingas observed, and to try to understand whether reasonable ac-tions could encourage further changes in the growth rates.
The growth rate of the greenhouse-gas climate forcing peakedin the early 1980s at almost 0.5 watt per square meter perdecade but declined by the 1990s to about 0.3 watt per squaremeter per decade. The primary reason for the decline was re-duced emissions of chlorofluorocarbons, whose production wasphased out because of their destructive effect on stratosphericozone.
The two most important greenhouse gases, with chlorofluo-rocarbons on the decline, are carbon dioxide and methane. Thegrowth rate of carbon dioxide surged after World War II, flat-tened out from the mid-1970s to the mid-1990s, and rose mod-erately in recent years to the current growth rate of about twoparts per million per year. The methane growth rate has declineddramatically in the past 20 years, by at least two thirds.
Year
W/m
2
(cha
nge
over
tim
e)
1950 1960 1970 1980 1990 2000
1.0
0
–1.0
REFLECTED ENERGY (from atmosphere and surface) 101 W/m2
100 W/m2 because of natural processes 1 W/m2 because of human-made aerosols
RADIATED HEAT (from land and ocean sinks) 238 W/m2
240 W/m2 because of natural processes –2 W/m2 because of human-made greenhouse gases, which return heat to the surface
1 W/m2 of excess energy, which warms the oceans and melts glaciers and ice sheets
TOTAL INCOMING SOLAR ENERGY 340 W/m2
Atmosphere
NET RESULT 1 W/m2
339 W/m2TOTAL OUTGOING ENERGY
Reflected Energyfrom Atmosphere
Reflected Energyfrom Earth
HeatReturnedto Earth
Incoming Energy from Sun
Radiated Heat
IMBALANCE OVER PAST HALF-CENTURY
we are much closer to dangerous anthropogenic interference than is generally realized
EARTH’S ENERGY IMBALANCETHE EARTH’S ENERGY is balanced when the outgoing heatfrom the earth equals the incoming energy from the sun. Atpresent the energy budget is not balanced (diagram andtable). Human-made aerosols have increased reflection ofsunlight by the earth, but this reflection is more than offsetby the trapping of heat radiation by greenhouse gases. Theexcess energy—about one watt per square meter—warmsthe ocean and melts ice. The simulated planetary energyimbalance (graph) is confirmed by measurements of heatstored in the oceans. The planetary energy imbalance is acritical metric, in that it measures the net climate forcingand foretells future global warming already in the pipeline.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
These growth rates are related to the rate of global fossil-fueluse. Fossil-fuel emissions increased by more than 4 percent ayear from the end of World War II until 1975 but subsequent-ly by only about 1 percent a year. The change in fossil-fuelgrowth rate occurred after the oil embargo and price increasesof the 1970s, with subsequent emphasis on energy efficiency.Methane growth has also been affected by other factors, in-cluding changes in rice farming and increased efforts to capturemethane at landfills and in mining operations.
If recent growth rates of these greenhouse gases continued,the added climate forcing in the next 50 years would be about1.5 watts per square meter. To this must be added the changecaused by other forcings, such as atmospheric ozone andaerosols. These forcings are not well monitored globally, but it is
known that they are increasing in some countries while decreas-ing in others. Their net effect should be small, but it could addas much as 0.5 watt per square meter. Thus, if there is no slow-ing of emission rates, the human-made climate forcing could in-crease by two watts per square meter in the next 50 years.
This “current trends” growth rate of climate forcings is atthe low end of the IPCC range of two to four watts per squaremeter. The IPCC four watts per square meter scenario requires4 percent a year exponential growth of carbon dioxide emis-sions maintained for 50 years and large growth of air pollution;it is implausible.
Nevertheless, the “current trends” scenario is larger thanthe one watt per square meter level that I suggested as our cur-rent best estimate for the level of dangerous anthropogenic in-
Crabo
D nio
xide
(p
art
spe
llim r
i on )
400
375
350
325
1940 1960 1980 2000 2020
Met
hane
(p
art
spe
rblliio
n) 1,750
1,500
1,250
1,000
Alternativescenario
IPCCprojection
IPCCprojection
Atmosphericdata
AtmosphericdataIce core data
Ice core data
Alternativescenario
Year
U.S. ENERGY CONSUMPTION
1980 1990 2000 2010 2020
Ener
yg
(qua
llirdio
nB
UTpe
rye
ra)
250
200
150
100
50
0
1975 projection(government and industry)
2002 projection(U.S. Energy Information
Administration)
CoalOil and gasNuclearRenewables
Year
GLOBAL CARBON DIOXIDE AND METHANE AMOUNTS
76 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
JEN
CH
RIS
TIAN
SEN
; SO
UR
CE
: JA
ME
S H
ANSE
N (
gra
ph
s);
MAR
K B
OU
LTO
N P
hot
o R
esea
rch
ers,
In
c. (
ph
otog
rap
h)
the emphasis should be on mitigating thechanges rather than just adapting to them
REDUCING EMISSIONS
OBSERVED AMOUNTS of carbon dioxide and methane (top twographs) fall below IPCC estimates, which have provedconsistently pessimistic. Although the author’s alternativescenario agrees better with observations, continuation on thatpath requires a gradual slowdown in carbon dioxide andmethane emissions. Improvements in energy efficiency(bottom graph) have allowed energy use in the U.S. to fall belowprojections in recent decades, but more rapid efficiency gainsare needed to achieve the carbon dioxide emissions of thealternative scenario, unless nuclear power and renewableenergies grow substantially.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 77
JAM
ES
HAN
SEN
fluence. This raises the question of whether there is a feasiblescenario with still lower climate forcing.
A Brighter FutureI HAVE DEVELOPED a specific alternative scenario that keepsadded climate forcing in the next 50 years at about one watt persquare meter. It has two components: first, halt or reverse growthof air pollutants, specifically soot, atmospheric ozone and meth-ane; second, keep average fossil-fuel carbon dioxide emissions inthe next 50 years about the same as today. The carbon dioxideand non–carbon dioxide portions of the scenario are equally im-portant. I argue that they are feasible and at the same time pro-tect human health and increase agricultural productivity.
In addressing air pollution, we should emphasize the con-stituents that contribute most to global warming. Methane of-fers a great opportunity. If human sources of methane are re-duced, it may even be possible to get the atmospheric methaneamount to decline, thus providing a cooling that would par-tially offset the carbon dioxide increase. Reductions of black-carbon aerosols would help counter the warming effect of re-ductions in sulfate aerosols. Atmospheric ozone precursors, be-sides methane, especially nitrogen oxides and volatile organiccompounds, must be reduced to decrease low-level atmosphericozone, the prime component of smog.
Actions needed to reduce methane, such as methane cap-ture at landfills and at waste management facilities and duringthe mining of fossil fuels, have economic benefits that partiallyoffset the costs. In some cases, methane’s value as a fuel entire-ly pays for the cost of capture. Reducing black carbon wouldalso have economic benefits, both in the decreased loss of lifeand work-years (minuscule soot particles carry toxic organiccompounds and metals deep into lungs) and in increased agri-cultural productivity in certain parts of the world. Primesources of black carbon are diesel fuels and biofuels (wood andcow dung, for example). These sources need to be dealt withfor health reasons. Diesel could be burned more cleanly withimproved technologies; however, there may be even better so-lutions, such as hydrogen fuel, which would eliminate ozoneprecursors as well as soot.
Improved energy efficiency and increased use of renewableenergies might level carbon dioxide emissions in the near term.Long-term reduction of carbon dioxide emissions is a greaterchallenge, as energy use will continue to rise. Progress is need-ed across the board: continued efficiency improvements, morerenewable energy, and new technologies that produce little orno carbon dioxide or that capture and sequester it. Next-gen-eration nuclear power, if acceptable to the public, could be animportant contributor. There may be new technologies before2050 that we have not imagined.
Observed global carbon dioxide and methane trends [seebox on opposite page] for the past several years show that thereal world is falling below all IPCC scenarios. It remains to beproved whether the smaller observed growth rates are a fluke,soon to return to IPCC rates, or are a meaningful difference.In contrast, the projections of my alternative scenario and the
observed growth rates are in agreement. This is not surprising,because that scenario was defined with observations in mind.And in the three years since the alternative scenario was defined,observations have continued on that path. I am not suggesting,however, that the alternative scenario can be achieved with-out concerted efforts to reduce anthropogenic climate forcings.
How can I be optimistic if climate is closer to the level ofdangerous anthropogenic interference than has been realized?If we compare the situation today with that 10 to 15 years ago,we note that the main elements required to halt climate changehave come into being with remarkable rapidity. I realize thatit will not be easy to stabilize greenhouse-gas concentrations,but I am optimistic because I expect that empirical evidence forclimate change and its impacts will continue to accumulate andthat this will influence the public, public-interest groups, in-dustry and governments at various levels. The question is: Willwe act soon enough?
For an expanded version of this article, including more dataand additional sources, see www.sciam.com/ontheweb
BUT WHAT ABOUT . . .
“Last winter was so cold! I don’t notice any global warming!”Global warming is ubiquitous, but its magnitude so far is only
about one degree Fahrenheit. Day-to-day weather fluctuationsare roughly 10 degrees F. Even averaged over a season thisnatural year-to-year variability is about two degrees F, so globalwarming does not make every season warmer than a fewdecades ago. But global warming already makes the probabilityof a warmer than “normal” season about 60 percent, rather thanthe 30 percent that prevailed from 1950 to 1980.
“The warming of the past century is just anatural rebound from the little ice age.”Any rebound from the European little ice age, which peaked in
1650–1750, would have been largely complete by the 20thcentury. Indeed, the natural long-term climate trend today wouldbe toward a colder climate were it not for human activities.
“Isn’t human-made global warming savingus from the next ice age?”Yes, but the gases that we have added to the atmosphere are
already far more than needed for that purpose.
“The surface warming is mainly urban ‘heatisland’ effects near weather stations.”Not so. As predicted, the greatest warming is found in remote
regions such as central Asia and Alaska. The largest areas ofsurface warming are over the ocean, far from urban locations[see maps at www.giss.nasa.gov/data/update/gistemp].Temperature profiles in the solid earth, at hundreds of boreholesaround the world, imply a warming of the continental surfacesbetween 0.5 and one degree C in the past century.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
ILLU
STR
ATIO
N B
Y JA
NA
BR
EN
NIN
G (
ph
otog
rap
h u
sed
for
ill
ust
rati
on p
urp
oses
on
ly)
Drug abuse produces long-term changes in the reward circuitry of the brain.Knowledge of the cellular and molecular details of theseadaptations could lead to newtreatments for the compulsivebehaviors that underlie addiction
White lines on a mirror. A needle and spoon. For many users, the sight ofa drug or its associated paraphernalia can elicit shudders of anticipa-
tory pleasure. Then, with the fix, comes the real rush: the warmth, the clarity,the vision, the relief, the sensation of being at the center of the universe. Fora brief period, everything feels right. But something happens after repeatedexposure to drugs of abuse—whether heroin or cocaine, whiskey or speed.
The amount that once produced euphoria doesn’t work as well, andusers come to need a shot or a snort just to feel normal; without it, theybecome depressed and, often, physically ill. Then they begin to use thedrug compulsively. At this point, they are addicted, losing controlover their use and suffering powerful cravings even after the thrillis gone and their habit begins to harm their health, finances andpersonal relationships.
Neurobiologists have long known that the euphoria in-duced by drugs of abuse arises because all these chemicalsultimately boost the activity of the brain’s reward system:
78 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
By Eric J. Nestler and Robert C. Malenka
TheAddicted
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
ADDICTION ARISES in part because habit-forming drugs cause the brain’s circuit forassessing reward to deem the drugs more desirable than anything else in life.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
a complex circuit of nerve cells, or neu-rons, that evolved to make us feel flush af-ter eating or sex—things we need to do tosurvive and pass along our genes. At leastinitially, goosing this system makes us feelgood and encourages us to repeat what-ever activity brought us such pleasure.
But new research indicates thatchronic drug use induces changes in thestructure and function of the system’sneurons that last for weeks, months oryears after the last fix. These adapta-tions, perversely, dampen the pleasur-able effects of a chronically abused sub-stance yet also increase the cravings thattrap the addict in a destructive spiral ofescalating use and increased fallout atwork and at home. Improved under-standing of these neural alterationsshould help provide better interventionsfor addiction, so that people who havefallen prey to habit-forming drugs canreclaim their brains and their lives.
Drugs to Die ForTHE REALIZATION that various drugsof abuse ultimately lead to addictionthrough a common pathway emergedlargely from studies of laboratory animalsthat began about 40 years ago. Given theopportunity, rats, mice and nonhumanprimates will self-administer the samesubstances that humans abuse. In theseexperiments, the animals are connected toan intravenous line. They are then taughtto press one lever to receive an infusionof drug through the IV, another lever toget a relatively uninteresting saline so-lution, and a third lever to request a foodpellet. Within a few days, the animals are hooked: they readily self-adminis-ter cocaine, heroin, amphetamine and
many other common habit-forming drugs.What is more, they eventually display
assorted behaviors of addiction. Individ-ual animals will take drugs at the expenseof normal activities such as eating andsleeping—some even to the point that theydie of exhaustion or malnutrition. For the
most addictive substances, such as co-caine, animals will spend most of theirwaking hours working to obtain more,even if it means pressing a lever hundredsof times for a single hit. And just as hu-man addicts experience intense cravingswhen they encounter drug paraphernaliaor places where they have scored, the an-imals, too, come to prefer an environmentthat they associate with the drug—an areain the cage in which lever pressing alwaysprovides chemical compensation.
When the substance is taken away,the animals soon cease to labor for chem-ical satisfaction. But the pleasure is notforgotten. A rat that has remained clean—
even for months—will immediately returnto its bar-pressing behavior when givenjust a taste of cocaine or placed in a cageit associates with a drug high. And certain
80 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
■ Drugs of abuse—cocaine, alcohol, opiates, amphetamine—all commandeer thebrain’s natural reward circuitry. Stimulation of this pathway reinforcesbehaviors, ensuring that whatever you just did, you’ll want to do again.
■ Repeated exposure to these drugs induces long-lasting adaptations in thebrain’s chemistry and architecture, altering how individual neurons in thebrain’s reward pathways process information and interact with one another.
■ Understanding how chronic exposure to drugs of abuse reshapes an addict’sbrain could lead to novel, more broadly effective ways to correct the cellularand molecular aberrations that lie at the heart of all addiction.
Overview/The Evolution of Addiction
THE BRAIN UNDER THE INFLUENCECHRONIC USE of addictivesubstances can change thebehavior of a key part of thebrain’s reward circuit: thepathway extending from thedopamine-producing nervecells (neurons) of the ventraltegmental area (VTA) todopamine-sensitive cells inthe nucleus accumbens.Those changes, induced inpart by the molecular actionsdepicted at the right and inthe graph, contributesignificantly to thetolerance, dependence andcraving that fuel repeateddrug use and lead to relapseseven after long periods ofabstention. The coloredarrows on the brain indicatesome of the pathways linkingthe nucleus accumbens andVTA with other regions thatcan help to make drug usershighly sensitive to remindersof past highs, vulnerable torelapses when stressed, andunable to control their urgesto seek drugs.
for example, isdispatched to the VTA, where it quietsdopamine release anddepresses the rewardcircuit, causing a userto need more drug tofeel high
Dopamine-sensitive cell in nucleus accumbens
To VTA
Dynorphin
Ventral tegmentalarea (VTA)
Prefrontalcortex
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
psychological stresses, such as a period-ic, unexpected foot shock, will send ratsscurrying back to drugs. These same typesof stimuli—exposure to low doses ofdrug, drug-associated cues or stress—trig-ger craving and relapse in human addicts.
Using this self-administration set-up and related techniques, researchersmapped the regions of the brain that me-diate addictive behaviors and discoveredthe central role of the brain’s reward cir-cuit. Drugs commandeer this circuit, stim-ulating its activity with a force and per-sistence greater than any natural reward.
A key component of the reward cir-cuitry is the mesolimbic dopamine sys-tem: a set of nerve cells that originate inthe ventral tegmental area (VTA), nearthe base of the brain, and send projectionsto target regions in the front of the brain—
most notably to a structure deep beneaththe frontal cortex called the nucleus ac-cumbens. Those VTA neurons communi-cate by dispatching the chemical messen-ger (neurotransmitter) dopamine fromthe terminals, or tips, of their long pro-jections to receptors on nucleus accum-bens neurons. The dopamine pathwayfrom the VTA to the nucleus accumbensis critical for addiction: animals with le-sions in these brain regions no longershow interest in substances of abuse.
Rheostat of RewardREWARD PATHWAYS are evolutionar-ily ancient. Even the simple, soil-dwellingworm Caenorhabditis elegans possessesa rudimentary version. In these worms,inactivation of four to eight key dopa-mine-containing neurons causes an ani-
mal to plow straight past a heap of bac-teria, its favorite meal.
In mammals, the reward circuit ismore complex, and it is integrated withseveral other brain regions that serve tocolor an experience with emotion and di-rect the individual’s response to rewardingstimuli, including food, sex and social in-teraction. The amygdala, for instance,helps to assess whether an experience ispleasurable or aversive—and whether itshould be repeated or avoided—and helpsto forge connections between an experi-ence and other cues; the hippocampusparticipates in recording the memories ofan experience, including where and whenand with whom it occurred; and thefrontal regions of the cerebral cortex co-ordinate and process all this informationand determine the ultimate behavior of theTE
RE
SE W
INSL
OW
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 81
1Dopamine signalingalso leads to
production of the proteindelta FosB (∆FosB)
2∆FosB repressesdynorphin synthesis
and activates specificgenes (different from thoseswitched on by CREB)
3The activatedgenes give rise
to proteins involvedin sensitizingresponses to drugsand to reminders of past drug use
4The protein CDK5, forexample, may promote
structural changes that couldmake nucleus accumbensneurons persistently sensitiveto drugs and drug-related cues
Delta FosB: A Key to Craving
2Those rises rapidlyactivate a protein
called CREB. Then CREBbound to DNA activatesspecific genes
1Dopamine signaling leadsto increases in cyclic AMP
(cAMP) and calcium ion (Ca2+) concentrations
CREB: A Source of Tolerance
Nucleus
Dynorphin gene
No dynorphin Gene activatedby ∆FosB
CDK5
Genes activated by CREB
Dopamine-producingnerve cell of VTA
Dopaminereceptor
Dopamine
cAMP
Ca2+
CREB
1 2 3 4 5
WHETHER a user is tolerant to adrug or, conversely, sensitized toit depends in part on the levels ofactive CREB and ∆FosB innucleus accumbens cells.Initially CREB dominates, leadingto tolerance and, in the drug’sabsence, discomfort that onlymore drug can cure. But CREBactivity falls within days whennot boosted by repeated hits. Incontrast, ∆FosB concentrationsstay elevated for weeks after thelast drug exposure. As CREBactivity declines, the dangerouslong-term sensitizing effects of∆FosB come to dominate.
Exposure to drug
Last exposure
Activitylevel
Days
TIMING MAKES A DIFFERENCE
∆FosB
∆FosBCREB
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
individual. The VTA-accumbens path-way, meanwhile, acts as a rheostat of re-ward: it “tells” the other brain centershow rewarding an activity is. The more re-warding an activity is deemed, the morelikely the organism is to remember it welland repeat it.
Although most knowledge of thebrain’s reward circuitry has been derivedfrom animals, brain-imaging studies con-ducted over the past 10 years have re-vealed that equivalent pathways controlnatural and drug rewards in humans. Us-ing functional magnetic resonance imag-ing (fMRI) or positron emission tomog-raphy (PET) scans (techniques that mea-sure changes in blood flow associatedwith neuronal activity), researchers havewatched the nucleus accumbens in cocaineaddicts light up when they are offered asnort. When the same addicts are shown avideo of someone using cocaine or a pho-tograph of white lines on a mirror, the ac-cumbens responds similarly, along withthe amygdala and some areas of the cor-tex. And the same regions react in com-pulsive gamblers who are shown imagesof slot machines, suggesting that the VTA-accumbens pathway has a similarly criti-cal role even in nondrug addictions.
Dopamine, PleaseHOW IS IT POSSIBLE that diverse ad-dictive substances—which have no com-mon structural features and exert a vari-
ety of effects on the body—all elicit sim-ilar responses in the brain’s reward cir-cuitry? How can cocaine, a stimulantthat causes the heart to race, and heroin,a pain-relieving sedative, be so oppositein some ways and yet alike in targetingthe reward system? The answer is that alldrugs of abuse, in addition to any othereffects, cause the nucleus accumbens toreceive a flood of dopamine and some-times also dopamine-mimicking signals.
When a nerve cell in the VTA is excit-ed, it sends an electrical message racingalong its axon—the signal-carrying “high-way” that extends into the nucleus ac-cumbens. The signal causes dopamine tobe released from the axon tip into the tinyspace—the synaptic cleft—that separatesthe axon terminal from a neuron in thenucleus accumbens. From there, the do-pamine latches onto its receptor on the ac-cumbens neuron and transmits its signalinto the cell. To later shut down the sig-nal, the VTA neuron removes the dopa-mine from the synaptic cleft and repack-ages it to be used again as needed.
Cocaine and other stimulants tem-porarily disable the transporter proteinthat returns the neurotransmitter to theVTA neuron terminals, thereby leavingexcess dopamine to act on the nucleus ac-cumbens. Heroin and other opiates, onthe other hand, bind to neurons in theVTA that normally shut down the dopa-mine-producing VTA neurons. The opi-
ates release this cellular clamp, thus free-ing the dopamine-secreting cells to pourextra dopamine into the nucleus accum-bens. Opiates can also generate a strong“reward” message by acting directly onthe nucleus accumbens.
But drugs do more than provide thedopamine jolt that induces euphoria andmediates the initial reward and rein-forcement. Over time and with repeatedexposure, they initiate the gradual adap-tations in the reward circuitry that giverise to addiction.
An Addiction Is BornTHE EARLY STAGES of addiction arecharacterized by tolerance and depen-dence. After a drug binge, an addictneeds more of the substance to get thesame effect on mood or concentrationand so on. This tolerance then provokesan escalation of drug use that engendersdependence—a need that manifests itselfas painful emotional and, at times, phys-ical reactions if access to a drug is cut off.Both tolerance and dependence occur because frequent drug use can, ironical-ly, suppress parts of the brain’s reward circuit.
At the heart of this cruel suppressionlies a molecule known as CREB (cAMPresponse element-binding protein). CREBis a transcription factor, a protein thatregulates the expression, or activity, ofgenes and thus the overall behavior of
82 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
SPOTS OF COLOR in brain scans of cocaine addicts (above) confirm animalstudies indicating that drug intake can induce profound immediate activitychanges in many brain regions, including those shown; brightest spots show themost significant change. While being scanned, the subjects rated their feelingsof rush and craving on a scale of zero to three—revealing that the VTA and thesublenticular extended amygdala are important to the cocaine-induced rush andthat the amygdala and the nucleus accumbens influence both the rush and thecraving for more drug, which becomes stronger as the euphoria wears off (graph).
nerve cells. When drugs of abuse are ad-ministered, dopamine concentrations inthe nucleus accumbens rise, inducing do-pamine-responsive cells to increase pro-duction of a small signaling molecule,cyclic AMP (cAMP), which in turn acti-vates CREB. After CREB is switched on,it binds to a specific set of genes, trigger-ing production of the proteins thosegenes encode.
Chronic drug use causes sustained ac-tivation of CREB, which enhances ex-pression of its target genes, some of whichcode for proteins that then dampen thereward circuitry. For example, CREBcontrols the production of dynorphin, anatural molecule with opiumlike effects.Dynorphin is synthesized by a subset ofneurons in the nucleus accumbens thatloop back and inhibit neurons in theVTA. Induction of dynorphin by CREBthereby stifles the brain’s reward circuit-ry, inducing tolerance by making thesame-old dose of drug less rewarding.The increase in dynorphin also con-tributes to dependence, as its inhibition ofthe reward pathway leaves the individual,in the drug’s absence, depressed and un-able to take pleasure in previously enjoy-able activities.
But CREB is only a piece of the story.This transcription factor is switched offwithin days after drug use stops. SoCREB cannot account for the longer-last-ing grip that abused substances have on
the brain—for the brain alterations thatcause addicts to return to a substanceeven after years or decades of abstinence.Such relapse is driven to a large extent bysensitization, a phenomenon whereby theeffects of a drug are augmented.
Although it might sound counterin-tuitive, the same drug can evoke both tol-erance and sensitization. Shortly after ahit, CREB activity is high and tolerancerules: for several days, the user wouldneed increasing amounts of drug to goosethe reward circuit. But if the addict ab-stains, CREB activity declines. At thatpoint, tolerance wanes and sensitizationsets in, kicking off the intense cravingthat underlies the compulsive drug-seek-ing behavior of addiction. A mere taste ora memory can draw the addict back. Thisrelentless yearning persists even afterlong periods of abstention. To under-stand the roots of sensitization, we haveto look for molecular changes that lastlonger than a few days. One candidateculprit is another transcription factor:delta FosB.
Road to RelapseDELTA FOSB APPEARS to functionvery differently in addiction than CREBdoes. Studies of mice and rats indicatethat in response to chronic drug abuse,delta FosB concentrations rise graduallyand progressively in the nucleus accum-bens and other brain regions. Moreover,because the protein is extraordinarily sta-ble, it remains active in these nerve cellsfor weeks to months after drug adminis-tration, a persistence that would enableit to maintain changes in gene expressionlong after drug taking ceased.
Studies of mutant mice that produceexcessive amounts of delta FosB in thenucleus accumbens show that prolongedinduction of this molecule causes animalsto become hypersensitive to drugs. Thesemice were highly prone to relapse afterthe drugs were withdrawn and latermade available—a finding implying thatdelta FosB concentrations could wellcontribute to long-term increases in sen-sitivity in the reward pathways of hu-mans. Interestingly, delta FosB is alsoproduced in the nucleus accumbens inmice in response to repetitious nondrugrewards, such as excessive wheel runningand sugar consumption. Hence, it mighthave a more general role in the develop-ment of compulsive behavior toward awide range of rewarding stimuli.
Recent evidence hints at a mechanismfor how sensitization could persist evenafter delta FosB concentrations return tonormal. Chronic exposure to cocaineand other drugs of abuse is known to in-duce the signal-receiving branches of nu-cleus accumbens neurons to sprout addi-tional buds, termed dendritic spines, thatbolster the cells’ connections to other neu-rons. In rodents, this sprouting can con-tinue for some months after drug takingceases. This discovery suggests that deltaFosB may be responsible for the added
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 83
ERIC J. NESTLER and ROBERT C. MALENKA study the molecular basis of drug addiction.Nestler, professor in and chair of the department of psychiatry at the University of TexasSouthwestern Medical Center at Dallas, was elected to the Institute of Medicine in 1998.Malenka, professor of psychiatry and behavioral sciences at the Stanford University Schoolof Medicine, joined the faculty there after serving as director of the Center for the Neurobi-ology of Addiction at the University of California, San Francisco. With Steven E. Hyman, nowat Harvard University, Nestler and Malenka wrote the textbook Molecular Basis of Neu-ropharmacology (McGraw-Hill, 2001).
THE
AU
THO
RS
MICROGRAPHS of nucleus accumbens neurons in animals exposed to nonaddictive drugs displaydendritic branches with normal numbers of signal-receiving projections called spines (left andcenter). But those who become addicted to cocaine sprout additional spines on the branches,which consequently look bushier (right). Presumably, such remodeling makes neurons moresensitive to signals from the VTA and elsewhere and thus contributes to drug sensitivity. Recentfindings suggest that delta FosB plays a part in spine growth.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
spines. Highly speculative extrapolationfrom these results raises the possibilitythat the extra connections generated bydelta FosB activity amplify signaling be-tween the linked cells for years and thatsuch heightened signaling might causethe brain to overreact to drug-relatedcues. The dendritic changes may, in theend, be the key adaptation that accountsfor the intransigence of addiction.
Learning AddictionTHUS FAR WE HAVE focused on drug-induced changes that relate to dopaminein the brain’s reward system. Recall, how-ever, that other brain regions—namely,the amygdala, hippocampus and frontalcortex—are involved in addiction andcommunicate back and forth with the
VTA and the nucleus accumbens. All thoseregions talk to the reward pathway by re-leasing the neurotransmitter glutamate.When drugs of abuse increase dopaminerelease from the VTA into the nucleus ac-cumbens, they also alter the responsive-ness of the VTA and nucleus accumbensto glutamate for days. Animal experi-ments indicate that changes in sensitivityto glutamate in the reward pathway en-hance both the release of dopamine fromthe VTA and responsiveness to dopaminein the nucleus accumbens, thereby pro-moting CREB and delta FosB activity andthe unhappy effects of these molecules.Furthermore, it seems that this altered glu-tamate sensitivity strengthens the neu-ronal pathways that link memories ofdrug-taking experiences with high reward,
thereby feeding the desire to seek the drug.The mechanism by which drugs alter
sensitivity to glutamate in neurons of thereward pathway is not yet known withcertainty, but a working hypothesis canbe formulated based on how glutamateaffects neurons in the hippocampus.There certain types of short-term stimulican enhance a cell’s response to glutamateover many hours. The phenomenon,dubbed long-term potentiation, helpsmemories to form and appears to be me-diated by the shuttling of certain gluta-mate-binding receptor proteins from in-tracellular stores, where they are not func-tional, to the nerve cell membrane, wherethey can respond to glutamate releasedinto a synapse. Drugs of abuse influencethe shuttling of glutamate receptors in the
84 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
COCAINE AND RELATEDSTIMULANTS block dopamineuptake or increase dopaminerelease by the terminals ofVTA cells and thus increasedopamine signaling in thenucleus accumbens
DIFFERENT DRUGS, SAME ULTIMATE EFFECTDRUGS OF ABUSE hit various targets in the brain, but all directly orindirectly enhance the amount of dopamine signaling in the nucleusaccumbens, thereby promoting addiction. Knowledge of the targetsraises ideas for therapy (see box on opposite page).
Nucleusaccumbens
neuron
Opiatereceptor
Glutamate
Glutamatereceptor
Dopaminetransporter
Dopamine
Cocaine
Dopaminereceptor
Projection fromcortex, amygdalaor hippocampus
MANY DRUGS, includingcocaine, amphetamine(speed), morphine andalcohol, can alter theresponses of nucleusaccumbens and VTA cells to glutamate in long-lastingways. Those changescontribute to drug cravingsby heightening memoriesof past drug experienceseven after the substance is no longer used
ALCOHOL AND OPIATES(opium, heroin andtheir relatives)enhance dopaminerelease by quietingneurons that wouldotherwise inhibitdopamine-secretingneurons
NICOTINE inducesVTA cells to releasedopamine into thenucleus accumbens
Dopamine-releasing VTA
neuron
CREB
Inhibitory neuron
in VTA
∆FosB Opiumlikeneurotransmittermade by neurons
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
reward pathway. Some findings suggestthat they can also influence the synthesisof certain glutamate receptors.
Taken together, all the drug-inducedchanges in the reward circuit that we havediscussed ultimately promote tolerance,dependence, craving, relapse and thecomplicated behaviors that accompanyaddiction. Many details remain mysteri-ous, but we can say some things with as-surance. During prolonged drug use, andshortly after use ceases, changes in theconcentrations of cyclic AMP and the ac-tivity of CREB in neurons in the rewardpathway predominate. These alterationscause tolerance and dependence, reducingsensitivity to the drug and rendering theaddict depressed and lacking motivation.With more prolonged abstention, changes
in delta FosB activity and glutamate sig-naling predominate. These actions seemto be the ones that draw an addict backfor more—by increasing sensitivity to thedrug’s effects if it is used again after alapse and by eliciting powerful responsesto memories of past highs and to cues thatbring those memories to mind.
The revisions in CREB, delta FosBand glutamate signaling are central to ad-diction, but they certainly are not thewhole story. As research progresses, neu-roscientists will surely uncover other im-portant molecular and cellular adapta-tions in the reward circuit and in relatedbrain areas that will illuminate the truenature of addiction.
A Common Cure?BEYOND IMPROVING understandingof the biological basis of drug addiction,the discovery of these molecular alter-ations provides novel targets for the bio-chemical treatment of this disorder. Andthe need for fresh therapies is enormous.In addition to addiction’s obvious phys-ical and psychological damage, the con-dition is a leading cause of medical ill-ness. Alcoholics are prone to cirrhosis ofthe liver, smokers are susceptible to lungcancer, and heroin addicts spread HIVwhen they share needles. Addiction’s tollon health and productivity in the U.S. hasbeen estimated at more than $300 billiona year, making it one of the most seriousproblems facing society. If the definitionof addiction is broadened to encompassother forms of compulsive pathologicalbehavior, such as overeating and gam-bling, the costs are far higher. Therapiesthat could correct aberrant, addictive re-actions to rewarding stimuli—whethercocaine or cheesecake or the thrill of win-ning at blackjack—would provide anenormous benefit to society.
Today’s treatments fail to cure mostaddicts. Some medications prevent the
drug from getting to its target. Thesemeasures leave users with an “addictedbrain” and intense drug craving. Othermedical interventions mimic a drug’s ef-fects and thereby dampen craving longenough for an addict to kick the habit.These chemical substitutes, however,may merely replace one habit with an-other. And although nonmedical, reha-bilitative treatments—such as the popu-lar 12-step programs—help many peoplegrapple with their addictions, partici-pants still relapse at a high rate.
Armed with insight into the biology ofaddiction, researchers may one day beable to design medicines that counter orcompensate for the long-term effects ofdrugs of abuse on reward regions in thebrain. Compounds that interact specifi-cally with the receptors that bind to glu-tamate or dopamine in the nucleus ac-cumbens, or chemicals that preventCREB or delta FosB from acting on theirtarget genes in that area, could potential-ly loosen a drug’s grip on an addict.
Furthermore, we need to learn to rec-ognize those individuals who are mostprone to addiction. Although psycholog-ical, social and environmental factorscertainly are important, studies in sus-ceptible families suggest that in humansabout 50 percent of the risk for drug ad-diction is genetic. The particular genes in-volved have not yet been identified, but ifsusceptible individuals could be recog-nized early on, interventions could be tar-geted to this vulnerable population.
Because emotional and social factorsoperate in addiction, we cannot expectmedications to fully treat the syndrome ofaddiction. But we can hope that futuretherapies will dampen the intense biolog-ical forces—the dependence, the crav-ings—that drive addiction and will there-by make psychosocial interventions moreeffective in helping to rebuild an addict’sbody and mind.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 85
Incentive-Sensitization and Addiction. Terry E. Robinson and Kent C. Berridge in Addiction,Vol. 96, No. 1, pages 103–114; January 2001.Molecular Basis of Long-Term Plasticity underlying Addiction. Eric J. Nestler in Nature ReviewsNeuroscience, Vol. 2, No. 2, pages 119–128; February 2001.Addiction: From Biology to Drug Policy. Second edition. A. Goldstein. Oxford University Press, 2001.National Institute on Drug Abuse Information on Common Drugs of Abuse:www.nida.nih.gov/DrugPages/
M O R E T O E X P L O R E
TREATMENTPOSSIBILITIESHypothetical anticocaine agentmight reduce dopaminesignaling in the nucleusaccumbens by interfering withcocaine’s ability to blockdopamine uptake by VTA neuron terminals.
Hypothetical broad-spectrumagent would mute dopamine’seffects by preventing CREB or ∆FosB from accumulating orfrom activating the target genesof these molecules.
Hypothetical broad-spectrumagent might interfere with theunhelpful changes in glutamatesignaling that occur in nucleusaccumbens cells with chronicdrug use.
Opiate antagonists (such asnaltrexone), already on themarket, block opiate receptors.They are used against alcoholismand cigarette smoking becausealcohol and nicotine triggerrelease of the brain’s own opiumlike molecules.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
A LACK OF
RUMBLING DOES NOT
NECESSARILY MAKE
AN EARTHQUAKE HARMLESS.SOME OF THE QUIET TYPES
COULD PRESAGE DEVASTATING
TSUNAMIS OR LARGER, GROUND-SHAKING SHOCKS
EarthquakesEarthquakesEarthquakes By Peter Cervelli
The Threat of
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
In early November 2000 the Big Island of Ha-waii experienced its largest earthquake in morethan a decade. Some 2,000 cubic kilometers of the southern slope of Kilauea volcano lurched to-ward the ocean, releasing the energy of a magnitude5.7 shock. Part of that motion took place under anarea where thousands of people stop every day tocatch a glimpse of one of the island’s most spectac-ular lava flows. Yet when the earthquake struck, noone noticed—not even seismologists.
How could such a notable event be overlooked?As it turns out, quaking is not an intrinsic part of allearthquakes. The event on Kilauea was one of thefirst unambiguous records of a so-called silent earth-quake, a type of massive earth movement unknownto science until just a few years ago. Indeed, I wouldnever have discovered this quake if my colleagues atthe U.S. Geological Survey’s Hawaiian Volcano Ob-servatory had not already been using a network ofsensitive instruments to monitor the volcano’s ac-tivity. When I finally noticed that Kilauea’s southflank had shifted 10 centimeters along an under-ground fault, I also saw that this movement had tak-en nearly 36 hours—a turtle’s pace for an earth-quake. In a typical tremor, opposite sides of the faultrocket past each other in a matter of seconds—
quickly enough to create the seismic waves thatcause the ground to rumble and shake.
But just because an earthquake happens slowlyand quietly does not make it insignificant. My co-in-vestigators and I realized immediately that Kilauea’s
silent earthquake could be a harbinger of disaster. Ifthat same large body of rock and debris were to gainmomentum and take the form of a gigantic land-slide—separating itself from the rest of the volcanoand sliding rapidly into the sea—the consequenceswould be devastating. The collapsing materialwould push seawater into towering tsunami wavesthat could threaten coastal cities along the entire Pa-cific Rim. Such catastrophic flank failure, as geolo-gists call it, is a potential threat around many islandvolcanoes worldwide.
Unexpected StirFORTUNATELY, the discovery of silent earth-quakes is revealing more good news than bad. Thechances of catastrophic flank failure are slim, andthe instruments that record silent earthquakes mightmake early warnings possible. New evidence forconditions that might trigger silent slip suggests boldstrategies for preventing flank collapse. Occurrencesof silent earthquakes are also being reported in ar-eas where flank failure is not an issue. There silentearthquakes are inspiring ways to improve forecastsof their ground-shaking counterparts.
The discovery of silent earthquakes and their linkto catastrophic flank collapse was a by-product of
GIANT LANDSLIDE(upper left) spawnedby a silent earthquakecould generate afearsome tsunamihundreds of metershigh (below).
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
efforts to study other potential naturalhazards. Destructive earthquakes andvolcanoes are a concern in Japan and theU.S. Pacific Northwest, where tectonicplates constantly plunge deep into theearth along what are called subductionzones. Beginning in the early 1990s, ge-ologists began deploying large networksof continuously recording Global Posi-tioning System (GPS) receivers in theseregions and along the slopes of active vol-canoes, such as Kilauea. By receiving sig-
nals from a constellation of more than 30navigational satellites, these instrumentscan measure their own positions on theplanet’s surface at any given time to with-in a few millimeters.
The scientists who deployed theseGPS receivers expected to see both theslow, relentless motion of the planet’sshell of tectonic plates and the relativelyquick movements that earthquakes andvolcanoes trigger. It came as some sur-prise when these instruments detectedsmall ground movements that were notassociated with any known earthquake oreruption. When researchers plotted theground movements on a map, the patternthat resulted very much resembled onecharacteristic of fault movement. In oth-er words, all the GPS stations on one sideof a given fault moved several centimetersin the same general direction. This pattern
would have been no surprise if it had tak-en a year or longer to form. In that case,scientists would have known that a slowand steady process called fault creep wasresponsible. But at rates of up to cen-timeters a day, the mystery events werehundreds of times as fast as that. Beyondtheir relative speediness, these silent earth-quakes shared another attribute withtheir noisy counterparts that distin-guished them from fault creep: they arenot steady processes but instead are dis-
crete events that begin and end suddenly.That sudden beginning, when it takes
place on the slopes of a volcanic island,creates concern about a possible cata-strophic flank event. Most typical earth-quakes happen along faults that havebuilt-in brakes: motion stops once thestress is relieved between the two chunksof earth that are trying to move past eachother. But activity may not stop if gravi-ty becomes the primary driver. In theworst-case scenario, the section of thevolcano lying above the fault becomes sounstable that once slip starts, gravitypulls the entire mountainside downhilluntil it disintegrates into a pile of debrison the ocean floor.
The slopes of volcanoes such as Ki-lauea become steep and vulnerable to thiskind of collapse when the lava from re-peated eruptions builds them up more
rapidly than they can erode away. Dis-covering the silent earthquake on Kilaueasuggests that the volcano’s south flank ison the move—perhaps on its way toeventual obliteration.
For now, friction along the fault is act-ing like an emergency brake. But gravityhas won out in many other instances inthe past. Scientists have long seen evi-dence of ancient collapses in sonar imagesof giant debris fields in the shallow waterssurrounding volcanic islands around the
world, including Majorca in the Mediter-ranean Sea and the Canary Islands in theAtlantic Ocean. In the Hawaiian Islands,geologists have found more than 25 indi-vidual collapses that have occurred overthe past five million years—the blink of aneye in geologic time.
In a typical slide, the volume of ma-terial that enters the ocean is hundreds oftimes as great as the section of Mount St.Helens that blew apart during the 1980eruption—more than enough to havetriggered immense tsunamis. On theHawaiian island of Lanai, for instance,geologists discovered evidence of waveaction, including abundant marine shellfragments, at elevations of 325 meters.Gary M. McMurtry of the University ofHawaii at Manoa and his colleagues con-clude that the most likely way the shellscould have reached such a lofty locationwas within the waves of a tsunami thatattained the astonishing height of 300meters along some Hawaiian coastlines.Most of the tallest waves recorded inmodern times were no more than onetenth that size.
Preparing for the WorstAS FRIGHTENING AS such an eventmay sound, this hazard must be under-stood in the proper context. Catastroph-ic failure of volcanic slopes is very rare ona human timescale—though far morecommon than the potential for a large as-teroid or comet to have a damaging col-
88 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
DAV
ID F
IER
STE
IN (
pre
ced
ing
pa
ges
)
■ Not all earthquakes shake the ground, it turns out. The so-called silent typesare forcing scientists to rethink their understanding of the way quake-pronefaults behave.
■ In rare instances, silent earthquakes that occur along the flanks of seasidevolcanoes may cascade into monstrous landslides that crash into the sea andtrigger towering tsunamis.
■ Silent earthquakes that take place within fault zones created by one tectonicplate diving under another may increase the chance of ground-shaking shocks.
■ In other locations, however, silent slip may decrease the likelihood ofdestructive quakes, because they release stress along faults that mightotherwise seem ready to snap.
Overview/Slippery Slope
Tsunami-generating VOLCANIC COLLAPSES may occur once every 10,000 years.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
lision with the earth. Collapses largeenough to generate a tsunami occursomewhere in the Hawaiian Islands onlyabout once every 100,000 years. Somescientists estimate that such events occurworldwide once every 10,000 years. Be-cause the hazard is extremely destructive
when it does happen, many scientistsagree that it is worth preparing for.
To detect deformation within unsta-ble volcanic islands, networks of contin-uous GPS receivers are beginning to bedeployed on Réunion Island in the IndianOcean, on Fogo in the Cape Verde Is-
lands, and throughout the Galápagosarchipelago, among others. Kilauea’s net-work of more than 20 GPS stations, forexample, has already revealed that thevolcano experiences creep, silent earth-quakes as well as large, destructive typi-cal earthquakes. Some scientists propose,
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 89
UNDERWATER FIELDS of rock and debris (examplesoutlined in red) reveal that massive sections of Hawaiianvolcanoes have crumbled into the ocean many times in thepast. Some geologists suspect that one collapse (blackoutline) off the western flank of Mauna Loa volcano kickedup a gigantic tsunami that deposited shattered shell androck as high as 800 meters along the nearby coast. Acomputer simulation (left) reveals that the same landslidemay have produced waves up to 300 meters high. For abrief time, Maui could have been divided in two and theocean floor west of Molokai could have been exposed.
GIANT LANDSLIDES AND TERRIFYING TSUNAMIS
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
however, that Kilauea may currently beprotected from catastrophic collapse byseveral underwater piles of mud androck—probably debris from old flankcollapses—that are buttressing its southflank. New discoveries about the way Ki-lauea is slipping can be easily generalizedto other island volcanoes that may nothave similar buttressing structures.
Whatever the specific circumstancesfor an island, the transition from silentslip to abrupt collapse would involve asudden acceleration of the mobile slope.In the worst case, this acceleration wouldproceed immediately to breakneck ve-locities, leaving no chance for early de-tection and warning. In the best case, theacceleration would occur in fits andstarts, in a cascade of silent earthquakes
slowly escalating into regular earth-quakes, and then on to catastrophe. Acontinuous GPS network could easily de-tect this fitful acceleration, well beforeground-shaking earthquakes began tooccur and, with luck, in plenty of time fora useful tsunami warning.
If the collapse were big enough, how-ever, a few hours’ or even days’ warningmight come as little comfort because itwould be so difficult at that point toevacuate everyone. This problem raisesthe question of whether authorities mightever implement preventive measures. Theproblem of stabilizing the teetering flanksof oceanic volcanoes is solvable—in prin-ciple. In practice, however, the effort re-quired would be immense. Consider sim-ple brute force. If enough rock were re-
moved from the upper reaches of an un-stable volcanic flank, then the gravita-tional potential energy that is driving thesystem toward collapse would disappearfor at least several hundred thousandyears. A second possible method—low-ering an unstable flank slowly through aseries of small earthquakes—would bemuch cheaper but fraught with geologicunknowns and potential dangers. To doso, scientists could conceivably harnessas a tool to prevent collapse the verything that may be currently driving silentearthquakes on Kilauea.
Nine days before the most recentsilent earthquake on Kilauea, a torrentialrainstorm dropped nearly a meter of wa-ter on the volcano in less than 36 hours.Geologists have long known that waterleaking into faults can trigger earth-quakes, and nine days is about the sameamount of time that they estimate it takeswater to work its way down throughcracks and pores in Kilauea’s fracturedbasaltic rock to a depth of five kilome-ters—where the silent earthquake oc-curred. My colleagues and I suspect that
90 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
DAV
ID F
IER
STE
IN
PETER CERVELLI is a research geophysicist at the U.S. Geological Survey’s Hawaiian Vol-cano Observatory, which sits along the rim of Kilauea Caldera on the Big Island. As leaderof the observatory’s crustal deformation project, Cervelli is responsible for interpreting datafrom a network of nearly 50 instruments that measure the tilt, strain and subtle movementswithin the island’s two most active volcanoes, Mauna Loa and Kilauea. Cervelli discoveredthe silent earthquake that struck Kilauea’s south flank in November 2000 while he wasworking on his Ph.D., which he received from Stanford University in 2001.
THE
AU
THO
R
PERCOLATING WATER may trigger silent earthquakes if it finds away into a vulnerable fault. Highly pressurized by the burden ofoverlying rock, water can push apart the two sides of the fault
(inset), making it easier for them to slip past each other (redarrows). This kind of silent slip can occur within subductionzones and volcanic islands.
RAINWATER may seep down from the earth’s surfaceinto shallow faults, such as those that separate anunstable slope from the rest of a volcano.
WATER squeezed out of hydrous minerals in a slabof ancient seafloor may enter faults created as theslab dives underneath another tectonic plate.
WATER-FILLED fault
SUBDUCTION ZONE VOLCANIC ISLAND
Unstable slope
Subducting seafloor
Continental crust
Sea level
Water
Rain
Rainwater
20 KM
40 KM
FaultFault
THE MECHANICS OF SILENT EARTHQUAKES
5 KM
10 KM
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
the burden of the overlying rock pressur-ized the rainwater, forcing the sides of thefault apart and making it much easier forthem to slip past each other.
This discovery lends credence to thecontroversial idea of forcefully injectingwater or steam into faults at the base ofan unstable flank to trigger the stress-re-lieving earthquakes needed to let it downslowly. This kind of human-induced sliphappens at very small scales all the timeat geothermal plants and other locationswhere water is pumped into the earth.
But when it comes to volcanoes, the ex-treme difficulty lies in putting the rightamount of fluid in the right place so asnot to inadvertently generate the verycollapse that is meant to be avoided.Some geophysicists considered this strat-egy as a way to relieve stress along Cali-fornia’s infamous San Andreas fault, butthey ultimately abandoned the idea forfear that it would create more problemsthan it would solve.
Wedges of WaterAPART FROM CALLING attention tothe phenomenon of catastrophic collapseof the flank of a volcano, the discovery ofsilent earthquakes is forcing scientists toreconsider various aspects of fault mo-tion—including seismic hazard assess-ments. In the U.S. Pacific Northwest, in-vestigators have observed many silentearthquakes along the enormous Casca-dia fault zone between the North Amer-ican plate and the subducting Juan deFuca plate. One curious feature of thesesilent earthquakes is that they happen atregular intervals—so regular, in fact, thatscientists are now predicting their occur-rence successfully.
This predictability most likely stemsfrom the fact that water flowing from be-low subduction zones may exert signifi-cant control over when and where thesefaults slip silently. As the subductingplate sinks deeper into the earth, it en-
counters higher and higher temperaturesand pressures, which release the signifi-cant amount of water trapped in water-rich minerals that exist within the slab.The silent earthquakes may then takeplace when a batch of fluid from the slabis working its way up—as the fluid pass-es, it will unclamp the fault zone a littlebit, perhaps allowing some slow slip.
What is more, Garry Rogers andHerb Dragert of the Geological Survey ofCanada reported last June that thesesilent tremors might even serve as pre-
cursors to some of the region’s large,ground-shaking shocks. Because the slowslips occur deep and at discrete intervals,they regulate the rate at which stress ac-cumulates on the shallower part of thefault zone, which moves in fits and starts.In this shallow, locked segment of thefault, it usually takes years or even cen-turies to amass the stress required to setoff a major shock. Rogers and Dragertsuggest, however, that silent slip maydramatically hasten this stress buildup,thereby increasing the risk of a regularearthquake in the weeks and months af-ter a silent one.
Silent earthquakes are forcing scien-tists to rethink seismic forecasts in otherparts of the world as well. Regions ofJapan near several so-called seismicgaps—areas where fewer than expectedregular earthquakes occur in an other-wise seismically active region—arethought to be overdue for a destructiveshock. But if silent slip has been relieving
stress along these faults without scientistsrealizing it, then the degree of dangermay actually be less than they think.Likewise, if silent slip is discovered alongfaults that were considered inactive up tonow, these structures will need carefulevaluation to determine whether they arealso capable of destructive earthquakes.
If future study reveals silent earth-quakes to be a common feature of mostlarge faults, then scientists will be forcedto revisit long-held doctrines about allearthquakes. The observation of many
different speeds of fault slip poses a realchallenge to theorists trying to explainthe faulting process with fundamentalphysical laws, for example. It is now be-lieved that the number and sizes of ob-served earthquakes can be explainedwith a fairly simple friction law. But canthis law also account for silent earth-quakes? So far no definitive answer hasbeen found, but research continues.
Silent earthquakes are only just be-ginning to enter the public lexicon. Thesesubtle events portend an exponential in-crease in our understanding of the howand why of fault slip. The importance ofdeciphering fault slip is difficult to over-state because when faults slip quickly,they can cause immense damage, some-times at a great distance from the source.The existence of silent earthquakes givesscientists a completely new angle on theslip process by permitting the detailedstudy of fault zones through every stageof their movement.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 91
Sudden Aseismic Fault Slip on the South Flank of Kilauea Volcano, Hawaii. Peter Cervelli, Paul Segall, Kaj Johnson, Michael Lisowski and Asta Miklius in Nature, Vol. 415, pages 1014–1017;February 28, 2002.
Episodic Tremor and Slip on the Cascadia Subduction Zone: The Chatter of Silent Slip.Garry Rogers and Herb Dragert in Science, Vol. 300, pages 1942–1943; June 20, 2003.
Giant Landslides, Mega-Tsunamis, and Paleo-Sea Level in the Hawaiian Islands. G. M. McMurtry,P. Watts, G. J. Fryer, J. R. Smith and F. Imamura in Marine Geology. Available online atwww.sciencedirect.com/science/journal/00253227
Visit the U.S. Geological Survey Hawaiian Volcano Observatory at http://hvo.wr.usgs.gov
M O R E T O E X P L O R E
Some SILENT EARTHQUAKES HAPPEN at suchregular intervals that they can be predicted successfully.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
92 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
The
By Partha Dasgupta and Eric Maskin
All VOTING SYSTEMS have drawbacks. But by taking intorank candidates, one system gives the
Most American and Frenchcitizens—indeed, those of democracies the worldover—spend little time contemplating their voting sys-tems. That preoccupation is usually left to political andelectoral analysts. But in the past few years, a largesegment of both these countries’ populations havefound themselves utterly perplexed. People in Francewondered how a politician well outside the politicalmainstream made it to the final two-candidate runoff
in the presidential election of 2002. In the U.S., manyvoters asked why the most popular candidate lost theelection of 2000.
We will leave discussions of hanging chads, butter-fly ballots, the electoral college and the U.S. SupremeCourt to political commentators. But based on re-search by ourselves and colleagues, we can address amore fundamental issue: What kinds of systems, bethey for electing national leaders or student council
Fairest
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 93
account how voters truest REFLECTION OF THE ELECTORATE’S VIEWS
presidents, go furthest toward truly representing thewishes of the voters? We argue that one particular sys-tem would be best in this sense—and it would be sim-ple and practical to implement in the U.S., France andmyriad other countries.
The Importance of Being RankedIN MOST NATIONAL presidential electoral systems,a voter chooses only his or her favorite candidate
rather than ranking them all. If just two candidatescompete, this limitation makes no difference. But withthree or more candidates, it can matter a great deal.
The French presidential election of 2002 provides acase in point. In the first round, voters could vote forone of nine candidates, the most prominent being theincumbent Jacques Chirac of the Gaullist party, the So-cialist leader Lionel Jospin and the National Front can-didate Jean-Marie Le Pen. The rules dictated that
Vote of all
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
if no candidate obtained an outright majority, the two candi-dates with the largest numbers of votes would face each other ina runoff. Chirac finished first (with 19.9 percent of the vote). Thereal surprise, however, lay in second place: the far-right-wingerLe Pen took it (with 16.9 percent), while Jospin—who, withChirac, had been heavily favored to reach the runoff—finishedthird (with 16.2 percent). In the second round, Chirac handilydefeated Le Pen.
Despite Jospin’s third-place finish, most available evidencesuggests that in a one-to-one contest against Le Pen, he wouldhave easily won. It is even plausible that he could have defeatedChirac had he made it to the second round. Yet by having vot-ers submit only their top choice, the French electoral system can-not take account of such important information. Furthermore,it permits extremist candidates such as Le Pen—candidates whohave no real chance of winning—to have an appreciable effecton the outcome.
The 2000 U.S. presidential election exposed similar short-comings. To make this point most clearly, we will pretend thatthe election procedure was simpler than it actually was. We willconsider just the four main candidates, and we will assume thatthere is no difference between the popular vote and the electoralcollege vote. (There have been many complaints about the elec-toral college, but even if it were replaced by popular vote, seri-ous problems would remain.) We will also assume that there areonly four kinds of voters: those who prefer Ralph Nader to AlGore, Gore to George W. Bush, and Bush to Pat Buchanan (the“Nader” voters); those with the ranking Gore, Bush, Nader,Buchanan (the “Gore” voters); those with the ranking Bush,Buchanan, Gore, Nader (the “Bush” voters); and those with theranking Buchanan, Bush, Gore, Nader (the “Buchanan” voters).
To be concrete, suppose that 2 percent of the electorate areNader voters, 49 percent Gore voters, 48 percent Bush voters,and 1 percent Buchanan voters. If voters each choose one can-didate, Gore will receive 49 percent and Bush 48 percent of the
total (the actual percentages were 48.4 percent and 47.9 percent,respectively). Given that no candidate receives a majority (thatis, more than 50 percent), how is the winner to be determined?Gore receives a plurality (the most votes short of 50 percent), soperhaps he should win.
On the other hand, the American Constitution stipulatesthat, absent a majority of the electoral votes, the House of Rep-resentatives should determine the winner. With a Republicanmajority in 2000, the House would presumably have gone forBush. Clearly, having U.S. voters name solely their favorite can-didate does not result in an outcome that is obviously right.
As in the French election, such ambiguity can be resolvedby having voters submit complete rankings. Even though Goreis the favorite of only 49 percent of the electorate, the rankingsshow that a clear majority of 51 percent—the Gore and Nadervoters combined—prefer Gore to either Bush or Buchanan. SoGore is the winner according to an electoral system called true
majority rule (or simple majority rule), in which voters submitrankings of all the candidates and the winner is the one whobeats each opponent in head-to-head competition based onthese rankings.
Rankings can also be used in other electoral systems. Con-sider, for instance, “rank-order voting”—a procedure oftenused to elect committee officers that has been proposed to solvethe problems inherent in the American and French presidentialelectoral systems. If four candidates are running, each voter as-signs four points to his or her favorite, three to the next favorite,two to the next, and one to the least favorite. The winner is thecandidate with the biggest total. The method appears to havebeen invented by Jean-Charles Borda, an 18th-century Frenchengineer, and is sometimes known as the Borda count.
Imagine that 100 million people vote in the U.S. election.Based on our earlier assumptions, we know that 49 million ofthem will rank Gore first. So Gore will receive 196 millionpoints—that is, 49 million times four points—from the Gorevoters. The Nader voters place him second, so he picks up sixmillion points from them. Finally, the Bush and Buchanan vot-ers place him third, for an additional 98 million points. Hisgrand total is 300 million points. If we make the correspond-ing computations for the others, we find that Nader gets 155million points and Buchanan 199 million. Strikingly, Bush gets346 million, even though a majority of the electorate preferGore [see scenario A in box on opposite page]. Only 2 percentof the electorate ranks Bush lower than second place, which isgood enough to elect him under rank-order voting.
Thus, true majority rule and rank-order voting result in dra-matically different outcomes. Considering this sharp contrast,it may seem hard to say which method is better at capturing theessence of voters’ views. But we propose to do just that. We canevaluate these two systems—and any other—according to some
94 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
TOB
Y TA
LBO
T AP
Ph
oto;
PH
OTO
ILLU
STR
ATIO
N B
Y JO
HN
NY
JOH
NSO
N (
pre
ced
ing
pa
ges
)
■ There is no such thing as a perfect voting system: everykind has one flaw or another.
■ Nevertheless, one method could solve some of theproblems that arose during recent elections in France andthe U.S. Called true majority rule, this system incorporatesinformation about the ranking of candidates, permitting amore accurate representation of voters’ views.
■ Our theoretical work shows that true majority rule moreoften avoids the flaws that arise for other voting methods.And, significantly, it could be easily implemented incountries the world over.
Overview/Getting Voting Right
True majority rule and rank-order voting result in DRAMATICALLY DIFFERENT outcomes.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
fundamental principles that any electoral method should sat-isfy. Kenneth J. Arrow of Stanford University originated thisaxiomatic approach to voting theory in a 1951 monograph, awork that has profoundly shaped the voting literature.
Most voting analysts would agree that any good electoralmethod ought to satisfy several axioms. One is the consensusprinciple, often called the Pareto principle after Italian sociol-ogist Vilfredo Pareto. It states that if everyone agrees that can-didate A is better than B, then B will not be elected. This ax-iom does not help discriminate between true majority rule andrank-order voting, however, because both methods satisfy it—that is, both will end up with B losing. Moreover, the princi-
ple does not apply very often: in our U.S. election example,there is no unanimous preference for any one candidate overanother.
Another important axiom holds that all voters should countequally—the “one-person, one-vote,” or equal-treatment, prin-ciple. Voting theorists call it the principle of anonymity: whoyou are should not determine your influence on the election.True majority rule and rank-order voting also both satisfyanonymity.
A third criterion, however, does differentiate between thetwo. Neutrality, as this axiom is called, has two components.The first is symmetry, which means that the electoral rulesshould not favor one candidate over the other. The second re-quires that the voters’ choice between candidates A and B shouldnot depend on their views about some third candidate C. Whatwould happen in our U.S. example if the Bush voters’ rankingshifted to become Bush, Gore, Buchanan, Nader (instead ofBush, Buchanan, Gore, Nader)? From the standpoint of true ma-jority rule, nothing important would change: the majority stillprefer Gore to Bush. But look at what happens under rank-order voting: Gore now receives 348 million points, while Bush’stotal remains 346 million [see scenario B in box at left]. Gorenow wins instead of Bush.
Obviously, rank-order voting can violate neutrality. Voters’preferences between Gore and Buchanan, a candidate whostands no chance of getting elected, determine the choice be-tween Bush and Gore—and the outcome of the election. In con-trast, true majority rule always satisfies neutrality. This last as-sertion may puzzle those readers who recall that in the actualelection, discussion abounded about whether votes for Naderwould affect the race between Bush and Gore. Indeed, in retro-spect it appears that Nader—perhaps with help from the infa-mous butterfly ballot in Florida and even from Buchanan—mayhave siphoned off enough Gore votes to tip the election to Bush.But this effect was possible only because the U.S. election sys-tem is not actually true majority rule but its own unique system.
Majority Rule and the French ElectionLET’S LOOK AT WHAT would happen to the French electionof 2002 under true majority rule—which, for simplicity’s sake,we will henceforth refer to as majority rule. Imagine Chirac,Jospin and Le Pen are the only candidates, and the electoratedivides into three groups. Everyone in the first group, 30 per-cent of voters, has the ranking Jospin, Chirac, Le Pen. In thesecond group, 36 percent of the electorate, the ranking isChirac, Jospin, Le Pen. In the remaining 34 percent, voters rankLe Pen over Jospin over Chirac. Chirac and Le Pen—with 36
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 95
SLIM
FIL
MS
PARTHA DASGUPTA and ERIC MASKIN frequently collaborate intheir research, including recent work on auction theory. Dasgup-ta is Frank Ramsey Professor of Economics at the University ofCambridge and past president of the Royal Economic Society.Maskin is Albert O. Hirschman Professor of Social Science at the In-stitute for Advanced Study in Princeton, N.J., and past presidentof the Econometric Society.
THE
AU
THO
RS
CANDIDATE POINTS VOTE TOTALSRANKING ASSIGNED (in millions)
GORE VOTERS 49% (of 100 million votes)Gore 4 4 X 49 = 196Bush 3 3 X 49 = 147Nader 2 2 X 49 = 98Buchanan 1 1 X 49 = 49
NADER VOTERS 2%Nader 4 4 X 2 = 8Gore 3 3 X 2 = 6Bush 2 2 X 2 = 4Buchanan 1 1 X 2 = 2
BUSH VOTERS 48%Bush 4 4 X 48 = 192Buchanan 3 3 X 48 = 144Gore 2 2 X 48 = 96Nader 1 1 X 48 = 48
BUCHANAN VOTERS 1%Buchanan 4 4 X 1 = 4Bush 3 3 X 1 = 3Gore 2 2 X 1 = 2Nader 1 1 X 1 = 1
Gore Total: 300Bush Total: 346
BUSH VOTERS 48%Bush 4 4 X 48 = 192Gore 3 3 X 48 = 144Buchanan 2 2 X 48 = 96Nader 1 1 X 48 = 48
Gore Total: 348Bush Total: 346
RANK-ORDER VOTING: SAMPLE SCENARIOSIN THIS ELECTORAL SYSTEM, candidates are ranked and thecorresponding points are tallied. Interestingly, even if acandidate were the true majority winner, as Gore is inscenario A, he or she would not necessarily win the rank-order vote. But a slight change in ranking, as happens withthe Bush voters in scenario B, can make an enormousdifference. In this case, it would lead to Gore winning.
Scenario A
Scenario B
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
and 34 percent of the vote, respectively—would move forwardinto a runoff, where Chirac would easily prevail because 66 per-cent of voters prefer him to Le Pen.
The same outcome would result under yet another system,called instant-runoff voting (IRV), which is practiced in Irelandand Australia and which, like rank-order voting, has been ad-vocated as an alternative to the French and U.S. systems. InIRV, simply put, rankings are used by election officials to suc-cessively eliminate the lowest-ranking candidates (and to in-corporate their percentages into the voters’ next-ranked choic-es) until only two candidates remain.
But the French and IRV systems conflict with majority rule.If you examine the configuration of voters’ rankings, you seethat Jospin actually commands an enormous majority: 64 per-cent of the electorate prefer him to Chirac, and 66 percent pre-
fer him to Le Pen. Majority rule dictates that Jospin should winby a landslide [see box above].
Recall that under majority rule a voter can make a politi-cal statement without harming the chances of any electable can-didate. Someone who preferred Jospin to Chirac and knew thatLe Pen had no chance of winning but wished to rank him firstas a gesture of protest could do so without fear of knockingJospin out of the race. (Except, of course, in the highly unlike-ly event that a majority of other voters made the same gesture.)The analogous point can be made about a voter who preferredGore to Bush but wished to lend symbolic support to Nader.
Yet despite these virtues, majority rule has a flaw. It can vi-olate another well-accepted voting principle: transitivity. Tran-sitivity requires that if candidate A is chosen over B, and B is cho-
sen over C, then A should be chosen over C. Now, ignoringBuchanan, pretend that 35 percent of the electorate prefer Goreto Bush to Nader, 33 percent rank Bush above Nader aboveGore, and 32 percent go for Nader above Gore above Bush. Six-ty-seven percent of voters rank Gore above Bush, 68 percentrank Bush above Nader, and 65 percent rank Nader aboveGore. In other words, no matter which candidate is chosen, atleast 65 percent of voters prefer somebody else! In this case, ma-jority rule produces no winner.
This possibility, called the Condorcet paradox, was identi-fied in the late 18th century by Marie-Jean-Antoine-Nicholasde Caritat, the Marquis de Condorcet, a colleague and arch-critic of Borda. The three rankings—Gore over Bush over Nad-er, Bush over Nader over Gore, and Nader over Gore overBush—are collectively called a Condorcet cycle.
Our comparison of majority rule and rank-order voting ap-pears to have resulted in a dead heat: majority rule satisfiesevery principle on our list except transitivity, and rank-ordervoting satisfies all but neutrality. This conundrum leads us toconsider whether some other electoral system exists that satis-fies all the principles. Arrow’s celebrated impossibility theoremsays no. It holds that any electoral method must sometimes vi-olate at least one principle [see “Rational Collective Choice,”by Douglas H. Blair and Robert A. Pollak; Scientific Amer-ican, August 1983].
Beyond ImpossibilityBUT ARROW’S THEOREM is unduly negative. It requires thatan electoral method must satisfy a given axiom, no matter whatvoters’ rankings turn out to be. Yet some rankings are quite un-likely. In particular, the Condorcet paradox—the bugaboo ofmajority rule—may not always be a serious problem in practice.After all, voters’ rankings do not come out of thin air. They of-ten derive from ideology.
To see what implications ideology holds for majority rule,think about each candidate’s position on a spectrum ranging
from the political left to the right. If we move from left to right,we presumably encounter the 2000 presidential candidates inthe order Nader, Gore, Bush, Buchanan. And if ideology drivesvoters’ views, then any voter who ranks Nader above Gore islikely to rank Gore above Bush and Bush above Buchanan. Sim-ilarly, any voter who ranks Bush above Gore can be anticipat-ed to rank Gore above Nader. We would not expect to find avoter with the ranking Bush, Nader, Gore, Buchanan.
In a pioneering paper published in the 1940s, the late Dun-can Black of the University College of North Wales showed thatif voters’ rankings are ideologically driven in the above man-ner—or at least if there are not too many nonideological voters—
majority rule will satisfy transitivity. This discovery made pos-sible a great deal of work in political science because, by posit-
96 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
JAC
K G
UE
Z AF
P/C
orb
is (
Josp
in);
DAN
IEL
JAN
IN A
FP/C
orb
is (
Ch
ira
c);
FRAN
ÇO
IS G
UIL
LOT
Reu
ters
New
Med
ia/C
orb
is (
Le P
en)
FRENCH ELECTION OF 2002CANDIDATE PERCENTAGE OF VOTERSRANKING CHOOSING THIS RANKING
When more than two choices present themselves, voters should SUBMIT A RANKING of candidates.
Jospin
Chirac
Le Pen
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
ing ideological rankings of candi-dates on the part of voters, re-searchers could circumvent theCondorcet paradox and make clearpredictions about the outcome ofmajority rule.
Of course, voters may not al-ways conform to such a tidy left-right spectrum. But other situationsalso ensure transitivity. For anoth-er example, look again at theFrench election. Although Chiracand Jospin led the two major par-ties, it seems fair to say that they didnot inspire much passion. It was theextremist candidate, Le Pen, whoaroused people’s repugnance or en-thusiasm: evidence suggests that ahuge majority of voters ranked himthird or first among the three topcandidates; few ranked him second.One can argue about whether suchpolarization is good or bad forFrance. But it is unquestionablygood for majority rule. If votersagree that one candidate of three isnot ranked second, transitivity isguaranteed. This property, calledvalue restriction, was introduced in1966 by Amartya Sen of HarvardUniversity.
In our research on voting, wesay that a voting system works wellfor a particular class of rankings ifit satisfies the four axioms when allvoters’ rankings belong to thatclass. For instance, majority ruleworks well when all rankings areideologically driven. It also works well when all rankings are“value restricted.” Indeed, we have found that whenever anyvoting system works well, so does majority rule. Furthermore,majority rule works well in some cases in which other systemsdo not. We call this the majority dominance theorem.
To illustrate, we will imagine a three-way race betweenGore, Bush and Nader. Suppose that every voter in fact ranksthe candidates as either Gore, Bush, Nader or Bush, Gore, Nad-er. With voters’ rankings belonging to this two-element class,rank-order voting satisfies its nemesis: the principle of neutral-ity (because voters’ views on Nader do not affect whether Bushor Gore wins a rank-order election). Yet majority rule alsoworks well here, because it satisfies its nemesis, transitivity.
But rank-order voting no longer works well if the situationbecomes slightly more complicated. If we add a third ranking—
Gore, Nader, Bush—majority rule is still transitive. These threerankings together do not constitute a Condorcet cycle. Rank-
order voting, however, no longersatisfies neutrality. Suppose 51 per-cent rank Bush above Gore aboveNader. If the remaining 49 percentrank Gore above Nader aboveBush, Gore will win. If the remain-der instead have the ranking Gore,Bush, Nader, however, then Bushwins—even though this group of 49percent has the same ranking ofGore and Bush in either case.
Majority rule still fails to workwell sometimes, as the Condorcetparadox shows, though less oftenthan other voting rules do. And insuch cases, it has to be modified toidentify a winner. There are manyways this can be done. Perhaps thesimplest modification is as follows:If no one obtains a majority againstall opponents, then among thosecandidates who defeat the most op-ponents in head-to-head compar-isons, select as winner the one withthe highest rank-order score.
Improving FutureElectionsTHE WAY most countries picktheir presidents is faulty. Both the2000 U.S. and 2002 French presi-dential elections were appreciablyaffected—perhaps decisively—bycandidates who had no realisticchance of winning. These candi-dates were able to wield influencebecause, in each case, only a voter’stop-ranked candidate was counted.
We believe that when more than two choices present them-selves, voters should submit a ranking of candidates and thatmajority rule—as we have discussed it—should determine thewinner. Such a method would not be perfect; no method is. Butas the majority dominance theorem shows, it would come clos-er to an accurate representation of the voters’ wishes than anyother system does.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 97
Social Choice and Individual Values. Kenneth J. Arrow. Wiley, 1951. (Yale University Press, 1990.)
The Theory of Committees and Elections. Duncan Black. CambridgeUniversity Press, 1958. (Kluwer Academic Publishers, 1998.)
Collective Choice and Social Welfare. Amartya Kumar Sen. Holden-Day,1970. (North-Holland, 1984.)
On the Robustness of Majority Rule and Unanimity Rule. Partha Dasgupta and Eric Maskin. Available atwww.sss.ias.edu/papers/papers/econpapers.html
M O R E T O E X P L O R E
BALLOT FROM SOUTH AFRICA’S first free elections in1994. Sixty-two percent of the electorate chose NelsonMandela and the African National Congress Party.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
98 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
KE
NT
SNO
DG
RAS
S P
reci
sion
Gra
ph
ics;
SO
UR
CE
S: S
WAT
CH
GR
OU
P A
ND
TIM
EX
QUARTZ WATCHES
Step on a hunk of quartz in the dirt, and you’ll thinknothing of it. But carve one of the stone’s crystals intoa tiny tuning fork, and you’ll have the key componentfor the watch ticking on your wrist.
Almost 90 percent of today’s watches are elec-tronic. Batteries provide the power to turn the handsor fire the liquid-crystal display, but quartz oscilla-tors—essentially, vibrating tuning forks—provide thechronometers’ steady beat. “Even a cheap quartzwatch is accurate to one or two seconds a month,”says Lou M. Galie, vice president of research and de-velopment at Timex in Middlebury, Conn. “Far moreprecise than expensive mechanical watches.”
Since the Renaissance, interconnected gears andwheels—driven by pendulums, weights or springswound tight by human hands—have turned the armsof clocks. By the early 1800s Swiss craftsmen werefabricating intricate wristwatches and founding com-panies that would dominate the trade for more than acentury. Quartz clocks appeared for sale around 1940,and bulky watches tested the market in the 1960s, butmost watchmakers saw the technology as a curiosity.A few Swiss firms improved the designs. Yet Japan-ese companies miniaturized the oscillator, battery,motor and circuitry and stormed the market in the1970s. Traditional watchmakers took 20 years to re-cover and join the electronic movement.
Mechanical watches, finer than ever, are now lim-ited to the high-price luxury market. Quartz watch-es owe their low cost to integrated circuits and theirsuperior accuracy to the oscillator’s high frequencyof 32,768 vibrations a second. The spinning balancewheel that paces a mechanical watch typically rocksback and forth at about five beats a second.
Quartz oscillators were employed in the 1930s bymilitary scientists to provide accurate timing for nav-igation equipment. Today Swatch and several Japan-ese firms supply most of the world’s tuning fork os-cillators, says Anton Bally, Swatch Group manufac-turing president. They are mass-produced fromartificial quartz using a photolithographic process de-vised by East German defector Juergen Staudte in1968 at North American Aviation, now Rockwell.
—Mark Fischetti
Rock Clock
WORKINGKNOWLEDGE
DIGITAL quartz watch is paced by a crystal oscillator thatvibrates when a battery applies voltage. The pulses, in turn,create a voltage that is fed back to the fork so it resonatesat 32,768 beats a second. The beats time a microprocessor;it tells electrodes which shapes (numbers) to create on theliquid-crystal display. Electronics such as capacitorscompensate for circuit feedback error.
Quartzoscillator
Capacitor
Battery
Liquid-crystaldisplay
Controlbuttons
Displayelectrodes
Microprocessor
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 99
➤ NO BATTERY? Batteries power more than 95 percent of quartz
watches. Early batteries lasted 18 months at best; now the finest can
reach 10 years. Ironically, the current produced by the photovoltaic
cells in most solar-powered watches is used to refresh a recharge-
able battery hidden inside, although a few models store energy in
a supercapacitor.
➤ GET WITH THE PROGRAM: The microprocessor in a $10 digital watch
has more power than the processor in the Apple II that popularized
personal computers around 1980. In that year, Timex employed no
software engineers, says vice president Lou M. Galie, but today pro-
grammers make up more than half the engineering staff.
➤ COMPLICATED: Watchmakers call extra hardware functions such as
chimes or date indicators “complications.” In 1783 Marie Antoinette’s
lover commissioned watch pioneer Abraham Breguet to devise the most
complicated watch ever, but it wasn’t completed until after the queen
was beheaded. In 1927 auto magnate James Packard paid $2,500 for
a watch that could show star positions from his Ohio home.
➤ QUARTZ, QUARTZ EVERYWHERE: Tiny quartz tuning forks provide pre-
cise reference frequencies for millions of computer chips, cellular
phones, radio transmitters, satellite transceivers and music synthe-
sizers. The shorter the tines, the higher the frequency. Music teachers
might find them difficult to strike against a piano, however.
Topic suggested by reader Avraham Aharoni. Send ideas to [email protected]
WINDUP mechanical watch ispowered by turning the crown,which coils up the mainspring.Small quantities of the storedenergy are released through anescapement. A rotating balancewheel provides the timing.
AUTOMATIC (also called kinetic or self-winding) watch is powered by an uneven weight that swingsas the wearer’s arm moves,turning a generator in a quartzdesign or coiling the mainspring in a mechanical movement.
OSCILLATOR is manufactured in the shape of a tuning fork from man-made quartz usingphotolithography. Early oscillators were machine-cut from natural crystals. Because quartz is apiezoelectric material, it vibrates when voltage isapplied to its gold-plated electrodes.
ANALOG quartz watch uses the same oscillator andfeedback circuitry as a digital design, but integratedcircuitry pares down the vibrations to two reliable beats asecond. The beats instruct a stepper motor to turn on andoff, which tells the second hand to start and stop each tick.
Quartz
Electrode
Quartz oscillator(sealed)
Capacitor
Battery
Swingingweight
Mainspring
Escapement
Mainspringor generator
Integratedcircuit
Steppermotor
Gears andwheels
Crown (to set hands)
Balancewheel
DID
YOU
KNOW
...
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
We stand 15 feet apart from one an-other until we form a long line thatstretches through dry grass and aroundmesquite shrubs, and then we start walk-ing through the scrubby Arizona desertunder a midday sun, our eyes scanningthe ground. Chris Reed, a Phoenix nativewho has done this before, calls out whenhe finds a circle of stones embedded inthe earth, and the line breaks as all 14 ofus cluster around for the official word.
“Very odd,” says J. Scott Wood, chiefarchaeologist for the Tonto National For-est. “Record it as a feature.” After scrap-ing in the dirt for a while and lighting acigar, Wood leans against his cane andbegins to muse: it’s a little big to be a stor-age pit, about right for a granary, butthere is no compressed dirt floor. TheHohokam, Salado and other peoples wholived here between 850 and the late 1200sor so didn’t use stone slabs as floors, hesays, but still, the ground should be com-pacted. In this way, on an October daythat reaches upward of 90 degreesFahrenheit, an unusual weeklong fieldseason begins.
Unusual, because those accompany-ing Wood and three other U.S. ForestService archaeologists are not profes-sionals—although some have had ar-chaeological training—but rather volun-teers, many of whom spend their vaca-tions or retirement working alongsideresearchers in the field, surveying sites,making discoveries. They are all partici-pants in Passport in Time (PIT), a U.S.Department of Agriculture program thatbegan regionally in 1989 and went na-tional in 1991. An average of 2,500 peo-
ple a year join projects that include doc-umenting petroglyphs on Kosciusko Is-land in Alaska, restoring old forest-firelookouts in Washington State and exca-vating a sauropod in Colorado. (Visitwww.passportintime.com for more in-
formation about PIT and how to apply.) Since the program’s inception, more
than 12,800 people between the ages of12 and 80 have worked on ventures in 38states. Unlike similar projects in whichindividuals pay to do fieldwork—such as
VOYAGES
100 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
DAN
NY
LEH
MAN
Cor
bis
Passport in TimeVOLUNTEERS JOIN ARCHAEOLOGICAL AND HISTORICAL FOREST SERVICE PROJECTS AROUND THE COUNTRY,LEARNING FIELD TECHNIQUES BY MARGUERITE HOLLOWAY
ROOSEVELT LAKE in Tonto National Forest in Arizona stretches for about 25 miles. The U.S. Forest Service runs one or two Passport in Time projects in the forest every year.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
U.S
. FO
RE
ST S
ER
VIC
E A
P P
hot
o
Earthwatch Institute, based in Maynard,Mass., or Crow Canyon ArchaeologicalCenter in Colorado—Passport in Timerequires only that participants pay theirway to the site and for their upkeep oncethere. Depending on the location of thestudy area, volunteers stay in motels,Forest Service quarters, campgrounds ortheir own RVs.
“Some people go from PIT to PIT,”says Stephen Germick, an archaeologist atTonto National Forest and its Passport inTime coordinator. “They leave in Mayand return home in September.” Germickhas come to rely on the expertise of theseso-called PIT-heads, and he selects a mixof novices and returnees, many of whomhave come to possess professional-levelknowledge. One man in his 70s, for in-stance, has been working on another Ton-to forest project—surveying the 1875 Sil-ver King Mine—since 1996. “He is essen-tial. He is our mapper,” Germick says.
Mapping is an essential part of thisOctober project, too, and those volun-teers who have never done it before areinitiated quickly. The site we walk dur-ing our first pedestrian survey is called
the Armer Complex, after a ranchingfamily that settled here in the 1870s. Itwas also the home of several NativeAmerican tribes for hundreds of years.As the largest site in the Tonto Basin andone that clearly experienced waves of set-tlement and trade, as well as a mysteriousand sudden demise, it is important to un-derstanding the prehistory of the area,Germick and Wood explain.
The site is often impossible to study.If Roosevelt Lake—a 25-mile-long reser-voir formed in 1911 when the RooseveltDam was completed—is full, the ruins aresubmerged. But because of the severedrought the Southwest has been experi-
encing, parts of the ArmerComplex have been ex-posed, and Germick hasbeen using PIT volunteersto map remnant storagerooms, granaries, burialsites, housing assemblages,trash mounds and agricul-tural plots—some of thembarely discernible circles orovals or corners of embed-ded stones.
Participants also look for the manyother records of various cultures. On ourfirst day in the field, volunteers find allmanner of pottery shards, and Wood an-alyzes the composition and pigment ofseveral of them to determine their age andorigin. They come from northeastern,east-central and southern Arizona, fromthe Anasazi, the Mogollon and the Ho-hokam. William Ramsey, a retired firecaptain from California and frequentPIT-goer, discovers a lovely fragment: redwith a series of black squares and lines—
a music-bar design. Wood examines it,pointing out the difference between salt-glazed paint on one part and vegetable
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 101
DROUGHT exposed ancient walls in September 2002. These ruinswere covered again when rains replenished Roosevelt Lake.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
paint on another. He identifies it asPinedale Black-on-red from between1275 and 1350.
The expert eyes of others locate ametate (grinding slab) and hammerstones. Frances Mayse—on her 11th PIT
and a steward for an archaeological sitenear her Tucson home—finds a beautifulmano (grinding stone). She then picks upwhat looks like a simple rock to me. “Seethe platform and the bulb of percussion?This would have been a scraper,” she ex-
plains. “I would have walked by this be-fore. But now I know. You can never goout for a hike again without walking intothe bushes.” It is true. We don’t look upat the mountains—the Sierra Ancha,Mazatzal and Apache—that frame thissite. Our eyes only cast about at our feetfor the past.
On the second day, after we have onceagain puttered and spluttered across thelake in a finicky boat provided by the near-by ranger station, Germick and JenniferBerke, a Tonto forest ethnohistorian, setme to mapping a large compound. Ger-mick puts pink flags in some of the corners
and along walls and points out the drag-on teeth—upright stones that were used asbase supports for mud and poles. Westand in one corner and stretch measuringtape to the next. Once we have recordedthe distance, we use the compass to figureout the direction of the far corner and thensketch the wall on graph paper, noting theright length in the right orientation. For anovice, it is not intuitive. But I have nochoice but to get it and get it right: Ger-mick and Berke are off measuring and dis-cussing, setting flags, bantering and argu-ing about what this extensive, virtuallyerased compound could have been.
Soon I, too, can begin to see some-thing of what they see. I see pit houses,then the later adobe compounds. I seeplots of agave and traders coming in fromover the mountains bringing differentpottery and textiles. And I can see howPIT-heads are born.
102 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
MAR
GU
ER
ITE
HO
LLO
WAY
VOYAGES
NORTH SHORE of the lake is where the ArmerComplex is found. Last October, PIT volunteersmapped part of the site.
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
REVIEWS
When he was still a student, RichardFeynman hinted at a career to come as ascientific wonderer when he wrote: “Iwonder why. I wonder why. / I wonderwhy I wonder / I wonder why I wonderwhy / I wonder why I wonder!”
Such wondering, and meta-wonder-ing, takes us to the heart of what geneti-cist-cum-neuroscientist Francis Crick(who would know) calls “the major un-solved problem in biology”—explaininghow billions of neurons swapping chemi-cals give rise to such subjective experiencesas consciousness, self-awareness, andawareness that others are conscious andself-aware.
The body of literature attempting tosolve this problem is extensive, and get-ting one’s mind around the field is a her-culean task successfully executed by psy-chologist Susan Blackmore in her de-lightful introduction, Consciousness.Presented as a textbook, it is so highly en-gaging that I recommend it for generalreaders, too. In many ways, the book isstructured like a brain, with loads of in-dependent modules (boxes and sidebarsfeaturing profiles, concepts and activi-ties) tied together by a flowing narrative
and integrated into a conceptual whole.The easy problem, Blackmore says, is
explaining each of the functional parts ofthe brain, such as “the discrimination ofstimuli, focusing of attention, accessingand reporting mental states, deliberatecontrol of behavior, or differences be-tween waking and sleep.” In contrast, thehard problem in consciousness studies “isexperience: what it is like to be an organ-ism, or to be in a given mental state.”
Adding up all of the solved easy prob-lems does not equal a solution to the hardproblem. Something else is going on inprivate subjective experiences—calledqualia—and there is no consensus onwhat it is. Dualists hold that qualia areseparate from physical objects in theworld and that mind is more than brain.Materialists contend that qualia are ulti-mately explicable through the activities ofneurons and that mind and brain are one.Blackmore, uniquely qualified to assess allcomers (she sports multihued hair, is adevotee of meditation, and studies alteredstates of consciousness), allows the myri-ad theorists to make their case (includingher own meme-centered theory) so thatyou can be the judge.
Making a strong case for the materi-alist position is Gerald M. Edelman’s lat-est contribution, Wider Than the Sky, of-fered as a “concise and understandable”explanation of consciousness “to the gen-eral reader.” Concise it is, but as for un-derstandable, Edelman understates: “Itwill certainly require a concentrated efforton the part of the reader.”
As director of the Neurosciences In-stitute in La Jolla, Calif., a Nobel laureateand author of several books on conscious-ness (Neural Darwinism, The Remem-bered Present and Bright Air, BrilliantFire), Edelman has impeccable credentials.But science writing for a general audienceinvolves more than expunging scholarlyreferences and providing a glossary of tech-nical terms as a substitute for clear expo-sition. To wit, on memory Edelman writesthat “it is more fruitfully looked on as aproperty of degenerate nonlinear interac-tions in a multidimensional network ofneuronal groups.” Such prose is commonthroughout the book, which is a shame be-cause Edelman is a luminously entertain-ing conversationalist, and his theory thatthe brain develops in a Darwinian fashionof neuronal variation and selection, and
The Major Unsolved Problem in BiologyTHREE BOOKS TRY TO EXPLAIN CONSCIOUSNESS BY MICHAEL SHERMER
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 103
THE QUEST FOR CONSCIOUSNESS: A NEUROBIOLOGICAL APPROACHby Christof KochRoberts & Company Publishers, Englewood, Colo., 2004 ($45)
WIDER THAN THE SKY: THE PHENOMENAL GIFT OF CONSCIOUSNESSby Gerald M. EdelmanYale University Press, New Haven, Conn., 2004 ($24)
CONSCIOUSNESS: AN INTRODUCTIONby Susan BlackmoreOxford University Press, New York, 2004 (paperbound, $39.95)
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
that consciousness is an emergent prop-erty of increasingly complex and inte-grated neuronal groups, has considerablesupport from neuroscience research.
An ideal combination of exquisiteprose and rigorous science can be foundin California Institute of Technology neu-roscientist Christof Koch’s The Quest forConsciousness. A rock climber adornedwith a tattoo of the Apple Computer logoon his arm, Koch takes an unabashedneurobiological approach, the natural ex-tension of what his longtime collaboratorFrancis Crick started in 1994 when he
wrote in The Astonishing Hypothesis“that ‘you,’ your joys and your sorrows,your memories and your ambitions, yoursense of personal identity and free will,are in fact no more than the behavior ofa vast assembly of nerve cells and their as-sociated molecules.” To me, the most as-tonishing aspect of this theory is that it isastonishing to anyone. Where else couldthe mind be but in the brain?
Nevertheless, finding the neuronalcorrelates of consciousness (NCC) hasproved elusive, so instead of concocting agrand unified theory, Koch and Crick un-
104 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
REVIEWS
LONELY PLANETS: THE NATURAL PHILOSOPHY OF ALIEN LIFEby David Grinspoon. Ecco, New York, 2003 ($39.95)As he tells engagingly the story of humankind’s long fascinationwith the possibility of extraterrestrial life, Grinspoon ponders theimpact of a first contact in the form of a radio message from anintelligent civilization. “It might be frightening, liberating, uplifting,disturbing, or all of the above, but I say, ‘Bring it on.’ ” And what ifthe first form of extraterrestrial life to be discovered turns out to bemicrobes? It “would enlarge our kingdom.” Grinspoon, principalscientist in the department of space studies at the Southwest Research Institute andadjunct professor of astrophysical and planetary sciences at the University of Coloradoat Boulder, concludes with his own belief: “I think our galaxy is full of species who havecrawled up from the slime of their home worlds, evolved self-awareness and started totinker, passed beyond the threat of technological self-extermination, and transcendedtheir animal origins to move out into the cosmos.”
A BRAND-NEW BIRD: HOW TWO AMATEUR SCIENTISTS CREATED THE FIRST GENETICALLY ENGINEERED ANIMALby Tim Birkhead. Basic Books, New York, 2003 ($26)The brand-new bird is the red canary. It was the object of a quest thattwo Germans—Hans Duncker, a high school teacher interested ingenetics, and Karl Reich, a bird keeper—carried on in Bremen formany years, beginning in 1921. Duncker’s idea was to pluck thegenes from a red siskin (a relative of the canary) and insert them intothe yellow canary. His method was cross-breeding. The effort fellshort of the goal, producing canaries of a reddish coppery hue. But itled to success years later by others who recognized the subtle connection betweengenes and the environment, in this case a diet containing carotenoids. Birkhead, professorof evolutionary biology at the University of Sheffield in England, makes a grand story byweaving in lore about genetics, bird keeping and the people involved in the quest.
All the books reviewed are available for purchase through www.sciam.com
THE EDITORS RECOMMEND
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
dertook a very specific research programfocusing on the visual system, to under-stand precisely how photons of light strik-ing your retina become fully integrated vi-sual experiences. Koch and his colleagues,for example, discovered a single neuronthat fires only when the subject sees animage of President Bill Clinton. If thisneuron died, would Clinton be im-peached from the brain? No, because thevisual representation of Clinton is dis-tributed throughout several areas of thebrain, in a hierarchical fashion, eventual-ly branching down to this single neuron.The visual coding of any face involves sev-eral groups of neurons—one to identifythe face, another to read its expression, athird to track its motion, and so on.
This hierarchy of data processing al-lows the brain to economize neural activ-ity through the use of combinatorics: “As-sume that two face neurons responded ei-ther not at all or by firing vigorously.Between them, they could represent fourfaces (one face is encoded by both cells notfiring, the second one by firing activity inone and silence in the other, and so on).Ten neurons could encode 210, or abouta thousand faces. . . . It has been calculat-ed that less than one hundred neurons aresufficient to distinguish one out of thou-sands of faces in a robust manner. Con-sidering that there are around 100,000cells below a square millimeter of cortex,the potential representational capacity ofany one cortical region is enormous.”
Given that the brain has about 100 bil-lion neurons, consciousness is most likelyan emergent property of these hierarchicaland combinatoric neuronal connections.How, precisely, the NCC produce qualiaremains to be explained, but Koch’s sci-entific approach, in my opinion, is the onlyone that will solve the hard problem.
Michael Shermer writes the Skepticcolumn for Scientific American and ispublisher of Skeptic (www.skeptic.com)and author of The Science of Good andEvil (Henry Holt and Company, 2004).
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 105w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 105COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
PUZZLINGADVENTURES
108 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
CO
RN
ELI
A B
LIK
Consider a square grid in which six north-southstreets, separated by gaps of 10 miles each, are ele-vated above six east-west streets laid out in a simi-lar fashion [see illustration below]. Entrance andexit ramps connect the streets at every intersection.Because there are no traffic lights, switching froma north-south street to an east-west street (and viceversa) takes essentially zero time. The grid has verylittle traffic, but the local police patrol very carefullyfor speeders.
The speed limits follow an unusual pattern. Thelimit is 10 miles per hour for the southernmost east-west street, 20 miles per hour for the east-weststreet immediately to the north, and so on. (There-fore, the limit for the northernmost east-west streetis 60 miles per hour.) Similarly, the speed limits forthe north-south streets range from 10 miles perhour for the westernmost to 60 miles per hour forthe easternmost. Let’s label the intersections usingtheir column and row numbers: the southwesterncorner of the grid is (1,1), the southeastern corner
is (6,1), the northwestern corner is (1,6), and so on.As a warm-up problem, can you determine thefastest legal way to get from (1,1) to (6,3)?
As it turns out, there are several optimally quickroutes. One of them goes from (1,1) to (2,1), whichtakes one hour, then to (2,3), which takes anotherhour, and finally to (6,3), which takes an addition-al hour and 20 minutes. The slowest direct route(that is, covering the same distance as the quickestroutes) goes from (1,1) to (6,1) in five hours andthen to (6,3) in another 20 minutes.
Your challenge is to visit every intersection in asshort a time as possible, starting from (1,1). Howdo you do it? And is there a better place to start inorder to visit every intersection in a shorter time?My guess is no, but I would very much like to seea clever proof. (Conjectures can so easily turn outto be wrong.)
Grid Speed BY DENNIS E. SHASHA
Finish (6,3)
Start (1,1)
10 miles
10 20 30 40 50 60Speed Limit (miles per hour)
10
20
30
40
50
60
Spee
d Limit
(mile
s per
hour)
Slowest routeAn optimallyfast route
Dennis E. Shasha is professor of computer scienceat the Courant Institute of New York University.
Answer to LastMonth’s PuzzleEncode your messageas a number that will become the x coordinate of a point P in three-dimensional space.Choose two othercoordinatesrandomly. Now selectfive planes that allintersect at point Pand assign a differentplane to each courier.Define each planeusing the coordinatesof three other (non-P)points in the plane.
Two nonparallelplanes meet at a line,and any plane notcontaining that linebut intersecting it willhit the line at a singlepoint. Knowing theplanes of any twocouriers would givethe enemy no usefulinformation aboutpoint P, but any threecouriers togethercould find the criticalpoint readily and thus determine the x coordinate to decipher the message.
Web SolutionFor a peek at theanswer to thismonth’s problem,visit www.sciam.com
CROSSTOWN TRAFFIC: The red route shows one answer to the warm-up problem. The yellow route shows the slowest paththat covers the same distance. But what is the fastest way to visit every intersection?
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
ANTIGRAVITY
When I was a small boy, I was stronglyin favor of reduced regulatory intrusioninto private activity. I remember riding inthe 1953 Dodge with my mother andasking her, “Wouldn’t it be great if therewere no stop signs?” She patiently ex-plained to me that, no, it would be verybad because of all the carnage and chaos.
I recalled that incident recently whenI received an e-mail entitled “If you areover 35, you should be dead.” As I am infact over 35, I decided to read it to clearup the mystery of why I was not dead.On opening the message, I thought (anact that provided axiomatic evidence ofmy being) that I recognized the contentsas an antiregulatory diatribe I had al-ready received a few times. And if I’vegotten this note a few times, perhapssome Scientific American readers have re-ceived it as well. Therefore, I decided torespond here, because I don’t have allyour e-mail addresses.
First, some highlights of the e-mail:“According to today’s regulators and bu-reaucrats, those of us who were kids inthe 40’s, 50’s, 60’s, or even maybe theearly 70’s probably shouldn’t have sur-vived. Our baby cribs were covered withbright colored lead-based paint. Whenwe rode our bikes, we had no helmets. Aschildren, we would ride in cars with noseatbelts or air bags. We fell out of trees,got cut and broke bones and teeth, andthere were no lawsuits from these acci-dents. Please pass this on to others whohave had the luck to grow up as kids, be-fore lawyers and government regulatedour lives, for our own good! People un-der 30 are WIMPS!”
I’m going to go out on a limb and as-sume that the target audience for thismessage is people who are alive. In dataanalysis, this is what’s known as selectionbias. Indeed, many kids didn’t wear seatbelts way back when. Some of them arenow, in technical medical terminology,dead. The dead ones don’t write suchrants. Kids brain-damaged by lead orpreventable blunt trauma may write, butthey are probably not responsible for theabove e-mail. Probably.
Still, life unfettered by bureaucratic in-terference remains tempting. And so I findmyself musing on the good old days. Imean the really old days—30,000 yearsago. Bureaucrats and lawyers didn’t evenexist yet. We were on our own and tookresponsibility for our actions. When wewere attacked by a lion, we bled until wepassed out and died. When we lost our
teeth, we stopped eating and died. If wewere painting a cave wall and scratched afinger that then got infected, we didn’tcomplain to the Occupational Safety &Health Administration. And we didn’tsue Og, who made the paint, or Oog, whochose the wall. We just waited for the in-fection to spread, and we keeled over. Ifwe wanted to go somewhere, driven bymind-numbing desperation to find somescrap of decaying antelope before westarved to death, we went on our owntwo bare feet. If we lived to 35, we gotstared at for being so elderly. And wetalked about how kids were wimps be-cause they wore animal skins on their feet.
Back in the present, all this talk aboutstarvation has made me hungry. Ordinar-ily I might have a treat for which I first de-veloped a taste when I was too young toappreciate stop signs: a burger. But on thisearly January day I have just learned thatthe U.S. Department of Agriculture has fi-nally decided to ban the sale of “downer”cattle, animals too sick to walk, in thewake of the country’s first case of madcow disease. (Nearly 200,000 downerswere slaughtered for sale last year.) Theban comes “only a few weeks after the de-partment and allies in the powerful meatlobby blocked an identical measure inCongress,” the Washington Post pointsout. The same article also quotes Repre-sentative Gary L. Ackerman of NewYork, who notes that the USDA “has seenthe light, but that’s only because they’vebeen struck by lightning.” Well, until I’msure that the regulation banning downersis having the desired effect, perhaps thischicken will have turkey instead.
110 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
ILLU
STR
ATIO
N B
Y M
ATT
CO
LLIN
S; P
HO
TOG
RAP
H B
Y FR
ANK
VE
RO
NSK
Y
Regulation RagRED TAPE CAN BE A PAIN, UNTIL YOU ACTUALLY NEED TAPE BY STEVE MIRSKY
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
Q
Q
w w w . s c i a m . c o m S C I E N T I F I C A M E R I C A N 111
SAR
A C
HE
N
Robertson D. Davenport, associate professor of pathology andmedical director of the Blood Bank and Transfusion Service atthe University of Michigan at Ann Arbor, explains:
Blood transfusions from strangers are less likely to be re-jected than organ transplants, for three key reasons. First, trans-fusions are given intravascularly, or into the circulatory system.Immune responses to antigens (foreign substances) received in-travascularly tend to be less pronounced than responses fromother routes. Second, transplanted organs contain immune cellsthat can stimulate the recipient’s defenses, whereas most suchcells in the blood are filtered out before transfusion. Third, thebody replaces transfused red blood cells within three months,reducing the chances that they willbe recognized as “alien,” whereastransplants may remain in place formany years.
Rejection of blood is relativelyrare. The risk of an acute hemolyticreaction is about one in 80,000 unitstransfused. To understand this oc-currence, it helps to review some ba-sic immunology. The two main types of immune responses arehumoral and cellular. Humoral responses produce antibodiesspecific to a foreign antigen. These antibodies may attach to theantigen, forming immune complexes; the liver and spleen de-stroy the complexes. The complexes can also activate the com-plement pathway. Complement activation can punch holes inthe membranes of bacteria or cells coated with antibodies. Im-mune responses elicited by blood transfusions are usually hu-moral. Organ transplants, in contrast, usually evoke cellularimmune responses, which lead to the creation of cell-killingagents called cytotoxic lymphocytes.
The key antigens in blood transfusions are in the ABO sys-tem. Blood types can be A (having A but not B antigens on redcells), B (having B but not A), AB (both A and B), or O (neither).Virtually everyone over the age of six months has antibodies tothe antigens they don’t produce. A patient who receives bloodwith the wrong antigens can have a serious reaction, includingbreakdown of red cells and a strong inflammatory response thatcould lead to kidney failure or even death.
More common is a delayed hemolytic reaction, occurringin one in 5,000 units transfused. The antigens involved are usu-ally not in the ABO system but in one of the other 25 knownblood group systems, such as Rh (Rhesus). The antibodies pro-duced tend not to activate complement, so the transfused cellsare not usually broken down. Instead the spleen removes thecells, and a milder inflammatory response may occur days toweeks after the transfusion; sometimes this reaction can lead torenal failure.
How can deleted computer filesbe retrieved at a later date?Clay Shields, professor of computer science at Georgetown Uni-versity, offers this answer:
“Deleted” files can be restored because they aren’t reallygone—at least not right away. This is because it is faster andmore efficient for computers to overwrite data only when nec-essary, when no other space is available to write new data.
A computer stores information in chunks called sectors. Afile may be written across several sectors and might be scatteredaround the disk. The operating system keeps an index of whichsectors belong to which files and a directory that maps the filenames to the index entries.
When a user deletes a file, its directory entry is either re-moved or labeled as deleted. A deleted file can thus be salvagedif the index information and sectors have not yet been reused.
Such recovery is easy in operating systems that simply markdirectory entries as deleted. A program scans the directory fordeleted entries and presents a menu of files to recover. In othertypes of systems, recovery is more complicated. The directoryentries may be lost, making it harder to find the file. The recov-ery program must look through all the index information andpiece together files from various sectors. Because sectors mayhave been reused, only parts of the file may be accessible.
Why are blood transfusions notrejected, as can happen with organs?
—K. Dahlke, Troy, Mich.
ASK THE EXPERTS
For a complete text of these and other answers from scientists in diverse fields, visit www.sciam.com/askexpert
COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.
112 S C I E N T I F I C A M E R I C A N M A R C H 2 0 0 4
