radlab.cs.berkeley.eduradlab.cs.berkeley.edu/people/fox/static/pubs/pdf/n07a.pdf · windblown...

J U N E 20 03 $4. 95W W W. S CI A M. COM

COMPUTER,HEAL THYSELF

AN END TO DISASTROUS CRASHES?

Martian Mysteries

How Chain Letters Evolve

Mad Cow–TypePlague Strikes

Wild Deer

TEST TUBE BABIES AND CLONES • THE NEXT STEP FOR PHYSICS

COPYRIGHT 2003 SCIENTIFIC AMERICAN, INC.

E M E R G I N G D I S E A S E S

38 Shoot This DeerB Y P H I L I P Y A MWild deer in parts of the U.S. are dying of a contagious wasting illness similar to mad cow disease.Unchecked, it might endanger humans and livestock.

A S T R O N O M Y

44 The Unearthly Landscapes of MarsB Y A R D E N L . A L B E EThe weird dynamics shaping the surface of the Red Planet reveal that Mars is not just a colder, drier Earth.

C O M P U T I N G

54 Self-Repairing ComputersB Y A R M A N D O F O X A N D D A V I D P A T T E R S O NSystems inevitably fail. The key to reliable computing is building systems that crash gracefully and recover quickly.

B I O T E C H N O L O G Y

62 Pandora’s BabyB Y R O B I N M A R A N T Z H E N I GIf today’s social arguments against human cloning sound familiar, it’s because foes of in vitro fertilization raised them 20 years ago.

P H Y S I C S

68 The Dawn of Physics beyond the Standard ModelB Y G O R D O N K A N EAfter 30 years of triumphs, the Standard Model of particle physics is at the height of its success. Something even better is on the way.

I N F O R M A T I O N S C I E N C E

76 Chain Lettersand Evolutionary HistoriesB Y C H A R L E S H . B E N N E T T , M I N G L I A N D B I N M AStudies of chain letters show how to infer the family tree of anything that evolves, from genes to languages to plagiarized schoolwork.

contentsjune 2003

SCIENTIFIC AMERICAN Volume 288 Number 6 features

54 The key to crash recovery

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 5COPYRIGHT 2003 SCIENTIFIC AMERICAN, INC.

Cover image by Kenn Brown.

6 S C I E N T I F I C A M E R I C A N J U N E 2 0 0 3

departments8 SA Perspectives

A pound of flesh for transplants.

10 How to Contact Us10 On the Web13 Letters16 50, 100 & 150 Years Ago18 News Scan

■ SARS and bioterror preparedness.■ Grassroots efforts to meet Kyoto standards.■ Hybrid rockets finally blast off.■ Heightened U.S. security blocks foreign scientists.■ A quantum “violation” of the Second Law.■ By the Numbers: Globalization trends.■ Data Points: Future freshwater shortages.

32 InnovationsJames Cameron directs robots, not DiCaprio, in a return to the Titanic.

34 Staking ClaimsA court ruling could, some universities fear, harshly limit the freedom to experiment.

36 Insights U.N. inspector Rocco Casagrande reflects on the search for bioweapons in Iraq.

82 Working KnowledgeBypassing the ear with implants.

84 Technicalities Music lovers use cell phones to name that tune.

87 ReviewsIn Emotions Revealed, Paul Ekman decodes the vocabulary of facial expressions.

36

32

28

SCIENTIFIC AMERICAN Volume 288 Number 6

columns35 Skeptic B Y M I C H A E L S H E R M E R

The numerological nonsense of The Bible Code.

89 Puzzling Adventures B Y D E N N I S E . S H A S H APrime squares.

90 Anti Gravity B Y S T E V E M I R S K YWe are what we ate.

91 Ask the Experts Why do hangovers occur? When you shake a can of coffee, why do the larger grains end up at the top?

92 Fuzzy Logic B Y R O Z C H A S T

Scientific American (ISSN 0036-8733), published monthly by Scientific American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111. Copyright © 2003 by ScientificAmerican, Inc. All rights reserved. No part of this issue may be reproduced by any mechanical, photographic or electronic process, or in the form of a phonographic recording,nor may it be stored in a retrieval system, transmitted or otherwise copied for public or private use without written permission of the publisher. Periodicals postage paid at NewYork, N.Y., and at additional mailing offices. Canada Post International Publications Mail (Canadian Distribution) Sales Agreement No. 242764. Canadian BN No. 127387652RT;QST No. Q1015332537. Subscription rates: one year $34.97, Canada $49 USD, International $55 USD. Postmaster: Send address changes to Scientific American, Box 3187,Harlan, Iowa 51537. Reprints available: write Reprint Department, Scientific American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111; (212) 451-8877;fax: (212) 355-0408 or send e-mail to [email protected] Subscription inquiries: U.S. and Canada (800) 333-1199; other (515) 247-7631. Printed in U.S.A.


In 2001 more than 6,000 people in the U.S. diedwhile waiting for an organ transplant. The dire short-fall of organs compared with patient demand is grow-ing as the population ages and more people experienceorgan failure. Although new immunosuppressive drugshave helped bridge the gap by allowing surgeons totransplant an organ that is a less than perfect match,there just aren’t enough organs to go around.

The reasons vary. Some people have religious or cul-tural objections to organ donation; many families sim-

ply have a tough time making a de-cision at a time of personal tragedy.Living donors—those who volun-teer a kidney or parts of their liver orlungs—are understandably reluc-tant: they must undergo potentiallylife-threatening surgery and puttheir own future health at risk.

The organ shortage has led var-ious policymakers to propose radi-cal steps. Several programs underconsideration in the U.S. and else-

where provide financial incentives to living donors orto the families of deceased donors. One approach,which has been instituted in Pennsylvania and is sup-ported by the American Society of Transplant Sur-geons, offers families who donate a loved one’s organs$300 in food and lodging expenses. Editorials in med-ical journals advocating the program assert that theamount of money is intentionally small to “express ap-preciation” for the donation but not to serve as a pay-ment. It is akin to the token coffee mug or umbrella onereceives after donating to public radio or television.

Evidence that such programs will boost the organsupply is lacking, largely because of a paucity of stud-ies. More important, some worry that these programswould mark the first step in encouraging an inhu-

mane and subtly coercive market for spare body parts.Although the outright purchase of organs is illegal

in nearly every country in the world, a number haveblack markets for living-donor organs, and the resultshave been chilling. A study of 305 living kidney donorsin Chennai, India, found that 96 percent sold a kid-ney to pay off debts, receiving about $1,070 apiece.But three fourths of the respondents soon faced debtand penury once again, and 79 percent would not rec-ommend organ selling to others. Permitting trade in or-gans has already led to the exploitation of the poor.

This is an extreme example, but it illustrates thedanger of attaching monetary value to whole organs.Society should redouble its support of less drastic stepsto encourage families and to reduce the dangers to liv-ing donors. A host of bills now in Congress would cre-ate a “medal of honor” for donors, offer medical leavefor living donors, or establish life and disability in-surance for living donors in case they experienced neg-ative side effects.

These initiatives could be paired with expandedpublic education campaigns that would explain theneed for organ donation and demystify the process.Physicians and hospital personnel also require moretraining in encouraging organ donation. Many Euro-pean countries either have implemented or are exper-imenting with “opt-out” laws, whereby the deceasedis presumed to have consented to an organ donationunless he or she indicated otherwise. (Family membersstill have the final say.) These laws raise their own ques-tions, but they bear watching.

Studies have shown that more than 95 percent offamilies would consent to organ donation if theyknew it was the wish of their loved one. Appealing topeople’s better natures may not be the only way toraise the number of organs available for transplanta-tion, but it is the best place to start.


VIC

TOR

DE

SC

HW

ANB

ER

G S

cien

ce P

hot

o Li

bra

ry

SA Perspectives

THE EDITORS [email protected]

A Pound of Flesh



How to Contact UsE D I T O R I A L

For Letters to the Editors:Letters to the Editors

Scientific American415 Madison Ave.

New York, NY 10017-1111or

[email protected] include your name

and mailing address, and cite the article

and the issue in which it appeared.

Letters may be edited for length and clarity.

We regret that we cannot answer all correspondence.

For general inquiries:Scientific American415 Madison Ave.

New York, NY 10017-1111 212-754-0550

fax: 212-755-1976or

[email protected]

S U B S C R I P T I O N SFor new subscriptions,

renewals, gifts, payments, and changes of address:

U.S. and Canada 800-333-1199

Outside North America 515-247-7631

or www.sciam.com

or Scientific American

Box 3187Harlan, IA 51537

R E P R I N T S To order reprints of articles:

Reprint DepartmentScientific American415 Madison Ave.

New York, NY 10017-1111212-451-8877

fax: [email protected]

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 10017-1111Please allow three to six weeks

for processing.

A D V E R T I S I N Gwww.sciam.com has electroniccontact information for salesrepresentatives of ScientificAmerican in all regions of the 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: 415-403-9033MidwestDerr Media Group847-615-1921fax: 847-735-1457 Southeast/SouthwestMancheeMedia972-662-2503fax: 972-662-2577DetroitKaren Teegarden & Associates248-642-1773fax: 248-642-6138CanadaDerr Media Group847-615-1921fax: 847-735-1457 U.K.The Powers Turner Group+44-207-592-8331fax: +44-207-630-9922France and SwitzerlandPEM-PEMA+33-1-46-37-2117fax: +33-1-47-38-6329GermanyPublicitas Germany GmbH+49-211-862-092-0fax: +49-211-862-092-21SwedenPublicitas Nordic AB+46-8-442-7050fax: +49-8-442-7059BelgiumPublicitas Media S.A.+32-(0)2-639-8420fax: +32-(0)2-639-8430Middle East and IndiaPeter Smith Media &Marketing+44-140-484-1321fax: +44-140-484-1320JapanPacific Business, Inc.+813-3661-6138fax: +813-3661-6139KoreaBiscom, Inc.+822-739-7840fax: +822-732-3662Hong KongHutton Media Limited+852-2528-9135fax: +852-2528-9281

On the WebWWW.SCIENTIFICAMERICAN.COM

FEATURED THIS MONTHVisit www.sciam.com/ontheweb to find these recent additions to the site:

Safeguarding GPSThe space-based Global Positioning System (GPS) signalthat guides “smart” bombs and cruise missiles to their targetsis the underpinning of U.S. technological superiority on thebattlefield. Yet because it is relatively easy to jam, the systemis also the Achilles’ heel of U.S. military might. Although

the integrity of the signal wasmaintained in the war withIraq, attempts to corrupt itunderscored the need toprotect GPS-dependentweapons and navigationsystems. Against a morecapable enemy, GPS mightfind itself among the firstcasualties of any new conflict.

Strung Out on the Universe: Interview with Raphael BoussoThe Holy Grail for many of today’s theoretical physicists isa complete quantum-mechanical theory of gravity—usefulfor understanding the behavior of black holes, big bangsand entire universes. But bridging the gap between thesmallest and largest constituents of reality will probablyrequire a few totally new concepts (and shake our faith insome old ones). One researcher looking for these missingpieces is Raphael Bousso of Harvard University. The 31-year-old shared first prize in an international competitionfor young physicists last year for his work on the so-calledholographic principle, which aims to reconcile quantummechanics with black hole physics. His research has ledhim to think hard about string theory and cosmology, too.

Ask the ExpertsWhat causes insomnia?Henry Olders, who studies sleep patterns, depression andfatigue at McGill University, explains.

FREE Newsletters fromScientificAmerican.comDelivering information to your in-box, when you want it and how you want it.

Subscribe today:http://sciam.rsc03.net/servlet/campaignrespondent N

ASA


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 13

LettersE D I T O R S @ S C I A M . C O M

Established 1845

®


EVOLVING IDEASA conclusion one can draw from geneticprogramming is that this evolution is in itsessence teleological [“Evolving Inven-tions,” by John R. Koza, Martin A. Keaneand Matthew J. Streeter]: the entire pro-cess is organized to realize a set of goalsexpressed in “high-level statements.” Ifthe authors are correct in claiming thattheir successful genetic programming em-ulates evolution in our world, then thenatural processes must likewise be thoughtto operate successfully to achieve a con-ceptual goal. Perhaps modern biology isin need of a fundamental revision.

Ted KrasnickiRichford, Vt.

Regarding “Evolving Inventions”: Couldthere be any clearer evidence that intelli-gent designs can occur given raw mate-rials and a selection process—without theneed for an intelligent designer?

Wil StarkSanta Rosa, Calif.

The article raises interesting questionsabout patent law. Any of the designs cre-ated by genetic programming would, bythe standards applied by the U.S. patentoffice, be regarded as novel, and thereforepatentable, had they been conceived inthe ordinary way. But with this geneticprogramming machine, the obvious—andhence, by definition, nonpatentable—

thing to do is to input your wish list for awidget and wait for the design to comeout. Where is the inventive step? If suchmachines get common, patents could be-

come a thing of the past. I am pleased thatI have just retired as a patent attorney.

David L. McNeighCheshire, England

That some machine may one day cir-cumvent my livelihood as an inventor isdisturbing. Why are we so determined tomake ourselves obsolete? The only thingwe have left is creativity. I beg you, pleasestop this research. I do not wish to haveto make cheeseburgers to sustain myself.

Robert La Dellacruzvia e-mail

KOZA, KEANE AND STREETER REPLY: Genet-ic programming is patterned after naturalevolution, but it is definitely not the same. Ar-tificial evolution holds up an explicit goal inthe hope of solving a particular problem byharnessing the problem-solving abilities ofnatural evolution. In nature, self-replicatingentities evolve over time and acquire traitsthat enable them to survive and prosper intheir environment (which also contains com-peting organisms and predators), but there isno prespecified final goal.

DRINK UP?I enjoyed “Drink to Your Health?” byArthur L. Klatsky, but I believe some im-portant caveats are in order. First, obser-vational studies, such as those quoted insupport of the benefits to cardiovascularhealth of moderate alcohol drinking, are fraught with difficulties. Until recent-ly, physicians advised postmenopausalwomen—based on observational stud-ies—that hormone replacement therapy

WHEN SCIENTIFIC AMERICAN runs an article that addressesevolution in any fashion, we can count on receiving spiritedreplies from all areas of the opinion spectrum. This is no lesstrue when the subject is technological, rather than biological, innature. “Evolving Inventions,” by John R. Koza, Martin A. Keaneand Matthew J. Streeter [February], which discussed a way todevelop new devices with software, served as something of aRorschach test for people’s views. Some saw the authors’ work asstrongly supporting the Darwinian explanation, whereas othersthought that it did not support the idea of natural evolution. Theseand additional reactions to the February issue appear below.

E D I T O R I N C H I E F : John RennieE X E C U T I V E E D I T O R : Mariette DiChristinaM A N A G I N G E D I T O R : Ricki L. RustingN E W S E D I T O R : Philip M. YamS P E C I A L P R O J E C T S E D I T O R : Gary StixS E N I O R E D I T O R : Michelle PressS E N I O R W R I T E R : W. Wayt GibbsE D I T O R S : Mark Alpert, Steven Ashley, Graham P. Collins, Carol Ezzell, Steve Mirsky, George MusserC O N T R I B U T I N G E D I T O R S : Mark Fischetti, Marguerite Holloway, Michael Shermer, Sarah Simpson, Paul Wallich

E D I T O R I A L D I R E C T O R , O N L I N E : Kate WongA S S O C I A T E E D I T O R , O N L I N E : Sarah GrahamW E B D E S I G N M A N A G E R : Ryan Reid

A R T D I R E C T O R : Edward BellS 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 BrenningA S S I S T A N T A R T D I R E C T O R S :

Johnny Johnson, Mark ClemensP H O T O G R A P H Y E D I T O R : Bridget GeretyP 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 KellerC O P Y C H I E F : Molly K. FrancesC O P Y A N D R E S E A R C H : Daniel C. Schlenoff, Rina Bander, Emily Harrison

E D I T O R I A L A D M I N I S T R A T O R : Jacob LaskyS 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 ShermanM A N U F A C T U R I N G M A N A G E R : Janet CermakA 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 CherebinP 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 PlacidoP R I N T P R O D U C T I O N M A N A G E R : Georgina FrancoP R O D U C T I O N M A N A G E R : Christina HippeliC 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 TerleckiC I R C U L A T I O N D I R E C T O R : Katherine CorvinoC I R C U L A T I O N P R O M O T I O N M A N A G E R : Joanne GuralnickF 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 BrandfonA S S O C I A T E P U B L I S H E R : Gail DelottS A L E S D E V E L O P M E N T M A N A G E R : David TirpackSALES REPRESENTATIVES: Stephen Dudley, Hunter Millington, Stan Schmidt, Debra Silver

ASSOCIATE PUBLISHER, STRATEGIC PLANNING: Laura SalantP R O M O T I O N M A N A G E R : Diane SchubeR E S E A R C H M A N A G E R : Aida DadurianP 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 FlorekB U S I N E S S M A N A G E R : Marie MaherM 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 , S C I E N T I F I C A M E R I C A N . C O M :

Mina C. Lux

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 McGarveyP E R M I S S I O N S M A N A G E R : Linda HertzM 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. HanleyC H A I R M A N : Rolf GrisebachP 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. TeichgraeberV 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 : Charles McCullaghV I C E P R E S I D E N T : Frances Newburg


with estrogen would reduce their risks ofcardiovascular disease. Now randomized,controlled trials demonstrate that suchtherapy actually increases the risk of coro-nary heart disease and stroke. As Klatskynotes, the question of alcohol and coro-nary health could be answered only by arandomized, controlled trial, a lengthyand probably impractical undertaking.

Second, observational studies are ham-pered by the low proportion of NorthAmerican and European adults who donot drink—a proportion of these peoplehave quit drinking because of previous al-cohol-related problems, and their healthoutcomes cannot be extrapolated to thewider population. The reliability of ob-servational studies can thus be questioned.In Scandinavia (with its higher proportionof alcohol abstainers), health outcomecomparisons are less pronounced.

Third, as Klatsky points out, alcoholwreaks serious damage onindividuals, communitiesand society. As a primarycare physician, I regularlysee patients whose liveshave been ruined by ex-cess alcohol. It behoovesus to be extremely cau-tious about alcohol con-sumption for perceivedcardiovascular benefits.

Steve CottamGreat Eccleston

Health CenterLancashire, England

KLATSKY REPLIES: Cottam is right that obser-vational data cannot completely rule out con-founders for associations. Undoubtedly, a con-founder of the observational association be-tween hormone replacement therapy andcardiovascular disease was that women whochose such therapy because they believed itto be beneficial also had a generally healthylifestyle. This situation was long suspected,and that fact influenced the decision to per-form clinical trials. It is unlikely, though, thatmoderate drinkers were similarly motivated,because most reports of the inverse alcohol-coronary relationship predated any wide-

spread knowledge of benefit, and drinking isnot typically a prescribed treatment.

I cannot agree, however, with the impli-cation that the alcohol-coronary data are in-consistent or unreliable. I’m not sure whichScandinavian studies are exceptions, but theCopenhagen Heart Study, for one, has shownstrong evidence for protection conferred bymoderate drinking. As Eric B. Rimm of the Har-vard School of Public Health recently wrote:“Few other associations are so uniformly re-ported in the literature despite diverse popu-lation samples, varying exposure, and incon-sistent control for confounding.”

Finally, I emphatically agree that all con-siderations of benefit by moderate drinkingneed to be considered in light of the terribletoll of heavy uncontrolled intake.

CAUTION ABOUT ANTIDEPRESSANTSI would like to mention a danger of anti-depressants such as lithium that wasn’t

covered in “Why? TheNeuroscience of Suicide,”by Carol Ezzell. Thesedrugs can cause a personwith bipolar disorder to“overshoot,” triggering amanic episode. It is sus-pected that a significantnumber of patients at-tempt suicide at the startof such an episode, as theycome out of their depres-sion. Among the newlyapproved mood stabiliz-ers that don’t have thisproblem are antiseizure

medications originally used for epilepsy,including Depakote, Tegretol, Neurontinand Lamictal.

R. Tim CosletSunnyvale, Calif.

MISSING THE TARGET?Michael Puttré, in “Satellite-Guided Mu-nitions,” missed a major class of guidedweapons: army missiles fired by the Mul-tiple Launch Rocket System (MLRS).MLRS fires the Army Tactical Missile Sys-tem, which has a range of up to 300 kilo-meters and was first used in Desert Storm.


MAT

T M

AHU

RIN

Letters

UNDERLYING neurobiologicalfactors may increase suicide risk.


The army missiles have a significant ad-vantage over air-delivered ones, becausepilots and expensive aircraft are not at risk,and reaction time to a call for support fromthe ground is considerably less. Last, thearticle should have mentioned dud rates,which are important for any weapon.

James B. LincolnColonel, U.S. Army (retired)

Annandale, Va.

For the sake of national security, it is un-wise to include detailed informationabout weapons, such as payload, rangeand accuracy. Even if this information isunclassified and readily available, I stillquestion the need for it to be published.

Jeff Korpavia e-mail

The term “smart” bomb is indeed an oxy-moron. The $20,000 for a kit to outfit a$3,000 “dumb” bomb could send a stu-dent to college for a year or support anunderprivileged American family. In thedeveloping world, that money couldbuild a clean-water well system for a vil-lage or provide vaccines for many people.

Nigel MackenzieVancouver, B.C.

SCIENTISTS AND PSYCHICSIn “Psychic Drift” [Skeptic], Michael Sher-mer asks why most scientists do not be-lieve in ESP and psi phenomena. An im-portant factor must be the way theirknowledge of the subject is in general lim-ited to unscientific articles in the media,plus the very limited number of researchpapers and articles, mainly hostile in char-acter, published in the major journals. Al-though the latter might appear to demon-strate that there is in essence no valid re-search in the area, in reality this situation ismuch more a reflection of negatively biasedpublication policies. Scientists are in a sit-uation similar to that of citizens of coun-tries where those in power have completecontrol over what they are allowed to read.

Brian D. JosephsonDepartment of Physics, Cavendish Laboratory

University of Cambridge

w w w . s c i a m . c o m

Letters


JUNE 1953TRUTH OR DAZE?—“Two lawyers andtwo psychiatrists on the Yale Universityfaculty recently issued a joint warningagainst the use of ‘truth serums’ in crim-inal investigations. The psychiatrists cit-ed clinical evidence to show that ‘normal’subjects readily hide what they wish tohide when under the influence of one ofthese drugs (sodium amytal), and that‘neurotic’ subjects frequently confess todeeds of which they are innocent. Thestatements elicited by drugs, they said,are more apt to be symbolically signifi-cant than objectively true.”

CHEMICAL SCRUBBER—“Chelation is nota brand-new discovery, but there is nowrising a flourishing industry which pro-duces made-to-order chelate compoundsfor many purposes, from softening waterto dissolving kidney stones. The varioususes of the chelate compounds all dependon one fascinating property: the ability ofthe crablike claw to seize and sequesteratoms of metal. Suppose that our water

supply contains dissolved salts of iron.The iron forms a sediment on standing;it discolors bathtubs and linens; it spoilsthe taste of tea. On the domestic scale itis very difficult to remove. We may, in-stead, add a chemical called EDTA to thewater. Now the iron will leave no stains.The iron is still there, yet it cannot be de-tected even by sensitive chemical tests. Itis tightly imprisoned—‘sequestered,’ inthe poetic language of chelation technol-ogy—by EDTA’s chelate rings. The soft-ening of water so far has been the largestuse of chelation.”

JUNE 1903THE DAWN HORSE—“The PaleontologicalDepartment of the American Museum ofNatural History, under the supervisionof Prof. Henry F. Osborn, the curator,has recently prepared a remarkable ex-hibit depicting the ancestry and evolutionof the horse. The blue-ribbon high-step-per of today is authentically traced backthree million years or more. At this re-mote time he was about the size of a fox,

only sixteen inches high, having four andfive toes, with which he scampered overthe marshes and shores of primevalearth. This noteworthy exhibit, the onlyone of its kind in America or elsewhere,has material from a special expedition forthe search of fossil horses that wasequipped and kept in the field for the pasttwo seasons. A series of fine water-colorpaintings by Charles R. Knight [see illus-tration] complete the display.”

FROM MARVEL TO JUNK—“The Ferriswheel, one of the attractions of the Chi-cago Exposition of 1893, was recentlysold at public auction for $1,800, en-gines, boilers, and all. Originally the con-trivance cost $362,000. It is said there areabout $300,000 worth of bonds out-standing against the owners of the wheel,as well as another $100,000 of debt.”

JUNE 1853FISH CORNUCOPIA—“The ‘SacramentoUnion’ says of the Sacramento river: ‘thewater of the river must be alive withsalmon, or such quantities caught dailywould sensibly reduce their numbers. Butexperienced fishermen inform us, whilethe run lasts, that no matter how manyare employed in the business, or howmany are taken daily, no diminution canbe perceived. Estimates give the numberof men employed at about 600; the num-ber of fish taken daily on average, 2,000,which would give as each man’s catch afraction over three a day.”

THE NEWS ON TOFU—“The Chinese pre-pare an actual cheese—legumin cheese—

from peas, called ‘tao-foo,’ which theysell in the streets of Canton. In preparingthis cheese, the paste from steeped groundpeas is boiled, which causes the starch todissolve with the casein. After straining,the liquid is coagulated by a solution ofgypsum. This coagulum is worked uplike sour milk, salted, and pressed intomoulds to make cheese.”


Chemical Claw ■ Horse Ancestor ■ Chinese Cheese

THE EXTINCT EOHIPPUS, as depicted by Charles R. Knight, 1903

50, 100 & 150 Years AgoFROM SCIENTIFIC AMERICAN



SUP

RI

Reu

ters

“The hospitals have been closed andpeople are dying.” A brief but chill-ing dispatch from the city of Guang-

zhou provided one of the outside world’sfirst hints of the chaos in southern China’sGuangdong Province as the mysterious dis-ease now known as SARS spread unchecked.“When I got [the message], the province wasalready in disarray, with wholesale demon-strations in the streets,” says retired U.S.Navy infectious disease investigator StephenCunnion, of his friend’s report that he post-

ed to ProMED-mail, an international infec-tious disease listserv.

Chinese officials have issued an extraor-dinary apology, effectively admitting thatmonths of secrecy and denial after the newillness appeared last November created acase study in how not to handle an infectiousdisease outbreak. But in the end, Chinamight have done the world a favor of sortsby providing a test of global readiness for aneven more devastating future epidemic,whether naturally occurring or unleashed in

an act of terrorism. WithSARS (severe acute respira-tory syndrome) having hit22 countries by mid-April,world preparedness looksdecidedly mixed.

“This was not the bigone,” says David Heymann,executive director for com-municable diseases for theWorld Health Organization.His global alerts helpedmost countries to gird forSARS. But Heymann, whosegroup keeps a lookout forkiller influenza strains thatmight emerge from thesame region, admits that heis worried. “We’ve alwayshad confidence in Hong

EM

ER

GIN

G D

ISE

ASE

S

Caught Off GuardSARS REVEALS GAPS IN GLOBAL DISEASE DEFENSE BY CHRISTINE SOARES

■ Severe acute respiratorysyndrome, or SARS, kills about

5 percent of its victims; another 10 to 15 percent survive

only because of modern intensive-care practices.(Influenza typically has a

1 to 2 percent mortality rate.)

■ The disease is caused by a newcoronavirus—one of a family of

large RNA viruses withglycoprotein “crowns”—that

invades immune cells. In SARS, theresulting inflammation of lung

tissue can lead to severepneumonia and even hemorrhage.

Two other coronaviruses causeabout one third of common colds.

■ As of April 15, the number ofprobable SARS cases had reached3,235 worldwide—2,650 of them

in Hong Kong and mainland China,162 in Singapore, 100 in Canada

and 35 in the U.S. Deaths worldwide: 154.

Up-to-date figures are atwww.who.int/csr/sarscountry/en/

FAST FACTS:VIRAL DEATH

SCANnews

WORLDWIDE SPECTER of SARS leads flight attendants from Qatar Airways to don masks on arrival at Jakarta, Indonesia.



LUIS

ASC

UI

Get

ty

newsSCAN

The Internet helped the world to learn quickly about SARS:

February 9:

The World Health Organization firstgets word of pneumonia cases inChina’s Guangdong Provincethrough an e-mail from a formerstaffer’s son living there.

A mention was picked up by theWeb-crawling program of the GlobalPublic Health Intelligence Network,a joint WHO/Canadian surveillanceproject created in 1998.

February 10:

A message by Stephen Cunnion, a retired U.S. Navy medicalinvestigator, showed up onProMED-mail, an internationallistserv launched in 1994. A tersenotice from Hong Kong healthofficials stating their awareness ofa pneumonia outbreak in mainlandChina accompanied it.

SPREADINGTHE WORD

Kong,” with its more modern and better-financed facilities, Heymann states, “whereaswe didn’t have confidence that China was pre-pared. Now I think we’ll have to reassessHong Kong.”

The spread of SARS may have been nosurprise to the Institute of Medicine (IOM),which in March quietly released a report,“Microbial Threats to Health: Emergence,Detection, and Response,” an update to itsstartling 1992 analysis of gaping holes in U.S.defenses against a natural microbial assault.The IOM’s latest analysis takes a global viewand adds the risk of bioterrorism. Althoughthe new study finds a few areas of improve-ment, it concludes that “the outlook is bleakon a number of fronts.”

The report notes that global surveillancehas improved—certainly none of the elec-tronic systems that alerted the world to SARSexisted in 1992—but experts judge it inade-quate. Stephen Morse of Columbia Universi-ty, an author of the original IOM report, be-lieves that health surveillance is not compre-hensive enough. “There’s still a lot of frag-mentation of knowledge. Nature isn’t stand-ing still; neither are potential terrorists.” Forinstance, even with its knowledge of the se-vere pneumonia outbreak in Guangdong asearly as February 9, WHO could do nothingwithout Chinese cooperation. “What we nev-er have is the teeth to go in if a country refusesinformation,” Heymann laments. A fullmonth passed until WHO learned aboutpneumonia patients who had infected an un-usually high number of hospital staff inHanoi, Hong Kong and Singapore. By then itwas too late for Hong Kong, where authori-ties were caught off guard by the ability ofSARS to spread not only through close con-tact but also through contaminated surfacesand possibly sewage. With dozens of newcases daily—many of them hospital work-ers—city health officials admitted defeat incontainment efforts by early April.

Preventing such a meltdown in medical in-frastructure is the rationale behind the currentU.S. program to vaccinate health workersagainst smallpox. But there are no vaccinesagainst some of the other “Category A” bio-agents, considered the most dangerous poten-tial weapons, such as the plague, tularemiaand Ebola pathogens. And no vaccine or drugis effective against the suspected SARS agent,a coronavirus so different from others in that

diverse viral family, it earned its own groupdesignation.

SARS is unlikely to have been a product ofbioengineering. Terrorists today are more aptto use known pathogens than to invent one,Morse says, but “a few years down the road,those who are technicallysophisticated will be able todo things that are moreimaginative.”

The U.S. research enter-prise must become equallyimaginative, the IOM re-port warns. Because thethreat of bioterrorism is partof a continuum with natu-rally occurring disease, theauthors urge a national“comprehensive infectiousdisease research agenda.”As if to make their point, avirus-gene microarray orig-inally conceived for infec-tious disease research provided the Centersfor Disease Control and Prevention’s firstclue that a coronavirus could be the SARSculprit. (Animal studies, too, have now im-plicated the coronavirus.)

Armed with a recently sextupled biode-fense budget, the National Institute of Aller-gy and Infectious Disease (NIAID) is aggres-sively recruiting researchers to develop nov-el antimicrobials and vaccines. The proposedProject BioShield, targeted to Category Apathogens, would entice biotechnology firmswith a $6-billion pool and a guaranteed cus-tomer in the form of the federal government.NIAID is also inviting biotech companies towork on a SARS vaccine, for which the po-tential worldwide market is probably incen-tive enough. Like the international laborato-ry network that identified the SARS agentwith unprecedented speed and cooperation,the endeavor could demonstrate what mod-ern science can accomplish when seriouslyapplied to combating microbial threats.

Heymann hopes the world does take alesson from the SARS experience: “It’s excel-lent practice for what might be coming.When you think of other diseases that havespread, like AIDS, it’s going to be very im-portant that when the next one comes, we doit even better—much better.”

Christine Soares is based in New York City.

SARS INFECTION is suspected because of the presence offluid congestion, which appears on x-rays as diffuse whiteareas in the lungs, most noticeable between the ribs.



KR

T

newsSCAN

F rustrated by federal inaction on pre-venting climate change, states and mu-nicipalities have begun reducing green-

house gas emissions on their own. In fact,their influence could be greater than that of

many countries that haveratified the Kyoto Pro-tocol, the internationalagreement that set reduc-tions of carbon emissionsbut that the U.S. has re-fused to ratify. In the pro-cess, the local-area poli-cies are serving as incu-bators for new proce-dures and technologiesthat will be important toa coordinated nationaleffort.

“There’s been a re-markable turn of eventsin the past two to four

years,” observes Susan Tierney of Lexecon,an economics consulting firm in Cambridge,Mass., and past assistant secretary for poli-cy in the U.S. Department of Energy. Tradi-tional first actors on air-quality issues, suchas California, New Jersey and the New En-gland states, have initiated programs to re-duce emissions. States are motivated not onlyby the danger of climate change but by thehope of cleaner air, cost savings from energyefficiency, and marketing opportunities forrenewable energy.

Such a “bottom-up” approach has a largeglobal potential: “If they were considered asindependent nations, U.S. states would com-prise about 25 of the top 60 countries thatemit greenhouse gases,” remarks Barry Rabeof the University of Michigan at Ann Arbor,whose “Greenhouse and Statehouse,” a PewCenter report, presents case studies of initia-tives in nine states. Texas alone exceedsFrance in emissions.

Raab reveals a surprising range of situa-tions among those states working to cut emis-sions. States moving ahead have been suc-cessful, he says, in couching the climatechange as a more immediate problem, such

as New Hampshire’s concern over the possi-ble loss of maple trees and the concomitantloss of tourism dollars from autumn’s leafpeepers. Many states have a champion push-ing the issue, such as Robert Shinn, formeradministrator of the Department of Environ-mental Protection in New Jersey. California’shistoric Pavley Bill of 2002, requiring strictlimits on vehicle emissions in 2009, couldserve to force redesigns of entire automobilefleets. Sixteen states now require utilities topurchase “green power.” Texas, for instance,sells renewable-energy credits and has seen asixfold increase in wind power generation be-tween 1999 and 2002.

The six New England states (Connecticut,Maine, Massachusetts, New Hampshire,Rhode Island and Vermont) have banded to-gether with five Canadian provinces (NewBrunswick, Newfoundland and Labrador,Nova Scotia, Prince Edward Island and Que-bec) to enact a Climate Change Action Plan.Written in 2001, the scheme aims to curbgreenhouse emissions to 1990 levels by 2010and then by an additional 10 percent by 2020.(Under the Kyoto Protocol, the U.S. wouldhave had to reduce average emissions in 2008through 2012 to 7 percent below 1990 levels.)

The first step calls for states to assess theamount of their greenhouse gas emissions;only 38 states have completed these invento-ries, which account for 87 percent of U.S.emissions. Then, to reduce emissions, plan-ners are focusing initially on “low-hangingfruit,” including replacing sport utility vehi-cles in state government fleets, acquiring moreenergy-efficient office equipment, and usinglight-emitting diodes for traffic lights. Sevenactivities in the region reported emissions re-ductions or sequestrations totaling 1.2 millionmetric tons of CO2-equivalent (MMTCE).

Cities, too, are acting on their own. Thir-ty-one specific plans have been filed by 141U.S. members of the International Councilfor Local Environmental Initiatives, repre-senting 16 percent of U.S. emissions. TenMMTCE of emissions have been eliminated,according to the council’s Susan Ode, inwhich western cities such as San Diego, Port-

Acting LocallyIN CURBING GREENHOUSE GAS EMISSIONS, STATES GO IT ALONE BY DAVID APPELL

GLO

BAL

WAR

MIN

G

Greenhouse gas emissions arecalculated in millions of metric

tons of CO2-equivalent (MMTCE), a measure that adds together the

climate warming potential of thedifferent atmospheric greenhouse

gases in units relative to that of carbon dioxide.

■ Estimated U.S. greenhouse gasemissions, 2001: 1,883.3 MMTCE

■ Emissions in 1990: 1,683 MMTCE

■ Completed state action plans: 20

■ Annual greenhouse gasreductions, 2000: 3.2 MMTCE

■ Potential reductions by 2010: 71 MMTCE

■ Potential by 2020: 96 MMTCE

■ Estimated cost savings by 2010:$8 billion

FAST FACTS:HOT AIR

REPLACING SPORT UTILITY VEHICLES with fuel-efficient autos isone strategy of states trying to reduce carbon emissions.



NAS

A

newsSCAN

W ith little fanfare last December,Lockheed Martin Space Systemslaunched a suborbital sounding rock-

et from a NASA pad in Virginia. Forty-fourmiles over the Atlantic, the five-story-tall,two-foot-diameter craft released an 800-pound payload. The package, containingaerodynamic reentry experiments, was noth-ing particularly special. The booster itself,however, was rather exceptional—it was thefirst launch of a rocket powered by a large-scale hybrid rocket motor.

Such rockets attempt to combine the bestof solid and liquid propulsion, the tradition-al engine types. In a liquid-fuel rocket, thefuel and oxidizer, often liquid hydrogen andoxygen, are stored separately and thenmixed to create combustion. Liquid-fuelrocket motors burn efficiently, provide high

thrust and, critically, can be throttled andeven stopped and restarted. Such controlpermits planners to tailor the rocket’s tra-jectory. Complexity, though, is high, and sotends to be the price tag.

Simpler and cheaper are solid-fuel en-gines; their fiery impetus comes from burningpremixed fuel and oxidizer grains that arepacked like coffee grounds into a cylindricalcasing. Unfortunately, the solid propellants—

usually aluminum fuel and ammonium per-chlorate oxidizer—burn fairly inefficiently,are toxic to the environment, and are difficultto fabricate and handle safely. A solid rock-et cannot be throttled, either—once lit, it runsuntil the fuel is expended.

Hybrid propulsion offers significant ad-vantages, claims Randy Tassin, a vice presi-dent at Lockheed Martin’s Michoud Opera-tions in New Orleans. “Hybrids are nonex-plosive, can be throttled, are low cost andenvironmentally benign,” he says. In addi-tion, the compact power plant can producenearly as much thrust as liquid-fuel motors.In a typical hybrid rocket motor, a rubberyfuel—a synthetic polymer called hydroxyl-terminated polybutadiene—cast into thetubular hull combusts fiercely when ignitedin the presence of oxygen, pumped in from aseparate tank as a liquid or a gas.

“The main difficulty in hybrid rockettechnology is controlling the way the propel-lant burns,” Tassin explains. The perfor-mance of hybrid fuels is not well understood,

land, Ore., and Salt Lake City are prominent.Although individual states cannot replace

a federal initiative, their patchwork regulato-ry approach could compel businesses to seekmore consistent, predictable nationwide stan-dards. States, however, often encounter thesame reluctance that has dominated the na-tional climate change scene. “We think,whether it’s federal, state or local, they’re ill-advised policies that are not going to helpstate or national economies and only succeedin putting more Americans out of work,” saysDarren McKinney of the National Associa-

tion of Manufacturers, an industrial trade or-ganization opposed to the Kyoto Protocol.

Still, the collective effort of the states is al-ready beginning to compensate for the lackof reductions by the Bush administration.“You may have some American states thatare better prepared, from a policy standpoint,to reduce greenhouse gases than a number ofnations that have ratified Kyoto,” Raab com-ments. The earth’s atmosphere will takewhatever help it can get.

David Appell lives in Ogunquit, Me.

Hybrids Take OffENGINEERS RECONSIDER CROSS-BRED PROPULSION BY STEVEN ASHLEY

RO

CK

ETS

POWERFUL PLUME blasts out of a hybrid rocket motor during a ground test conducted in 1999 at NASA’s Stennis Space Center in Mississippi. An aerospace industry consortium developed the250,000-pound thrust engine prototype as part of a $20-million program.



newsSCAN

A fter the September 11 attacks, theclampdown on those from overseaswishing to study in the U.S. was inevi-

table. The Patriot Act of 2001 quickly imple-mented an electronic system for tracking for-eign students, and officials are now exten-sively reviewing visa applications of scientists,engineers and students in technical fields.These and other ongoing efforts are creatinga “viscous” visa system, notes William F.Brinkman, president of the American Physi-cal Society (APS). Although such a systemmakes it harder for would-be terrorists to slipthrough, Brinkman maintains that it couldhobble the U.S. economy and actually com-promise national security.

The most visible effect of the visa restric-tions may be on the highly international en-deavors of physics. At Fermi National Accel-erator Laboratory in Batavia, Ill., scientistsfrom Vietnam, China, India and Russia, whoall had supplied equipment for one of the de-tectors, were unable to arrive and operate it.A dozen scientists missed a September 2002

conference at Brookhaven National Labora-tory. Vladimir B. Braginsky, a Russian whoheads a research group at the California In-stitute of Technology, could not return in timefor a meeting; Rashid A. Sunyaev, anotherRussian and director of the Max Planck In-stitute for Astrophysics in Garching, Ger-many, had to abandon his fall visit to Caltech.Many institutions are advising their foreignscientists to avoid leaving the U.S. (Similar de-lays are plaguing the biomedical field.)

American colleagues are frustrated thatthe U.S. Department of State, which in thepast was responsive to their concerns, nowseems to be turning a deaf ear. Kip S. Thorneof Caltech, who vainly sought updates on theRussian scientists’ applications, likens the“visa bureaucracy” to a black hole: “You can’tget any information out.” The proposed Pa-triot Act II would reportedly classify back-ground material on visas as confidential,which could make it impossible to fix aflawed application.

Senior scholars face delays, but students

because the specific interaction of the semi-solid fuel and oxygen is complex. “To get ahigh burn rate,” he continues, “you need toadd more surface area to the propellant,which results in relatively complicated fuelgeometries.” Some hybrid-fuel structureslook like wagon wheels in cross section, withmultiple oxidizer-injection ports set between“spokes” of fuel. During burning, the fuelsegments get thinner and thinner, sometimesbreaking off, which makes the motor runrough or even become unstable.

Researchers at the U.S. Air Force, aero-space firms and universities have worked foryears to address the unresolved technical is-sues. From 1999 to 2002, NASA and an in-dustry group spent about $20 million devel-oping and ground-testing a hybrid rocket en-gine that generated 250,000 pounds ofthrust. For the 60,000-pound-thrust hybridmotor that powered the Lockheed Martinsounding rocket, engineers configured the

fuel to burn in a staged, and hence more sta-ble, fashion than previous designs.

As a result of this slow but steady pro-gress, hybrid motor technology is now com-ing under greater consideration for variousmissions, including targets for “Star Wars”antimissile systems and upper stages of larg-er boosters. NASA is contemplating using hy-brids to propel the crew-escape capsule on anext-generation Orbital Space Plane. The in-ert fuel could be safely stored until an emer-gency, when onboard liquid oxygen could bepumped in, making the escape module readyto blast away to safety.

In the meantime, Lockheed Martin man-agers are focusing on suborbital soundingrockets for hybrid motors. NASA launches 25to 30 sounding rockets annually; the U.S. mil-itary uses them as well. Says Tassin: “It’s ourbelief that hybrids will eventually supplantsolid rockets and even some liquid types inmany future applications.”

Boxed OutSCIENCE LOSES AS THE U.S. TIGHTENS VISA RULES BY MADHUSREE MUKERJEE

SEC

UR

ITY

Scientists at Stanford Universityare investigating paraffin wax as a

potential fuel for hybrid rockets.Its high burn rate could produce

thrust equivalent to that generatedby the best liquid-fuel rockets.

When researchers ignite theirsmall demonstration motors, the

resulting heat causes nearby waxto melt. As the liquefied paraffinmixes with injected oxygen, the

surface area for combustionincreases, thus yielding a burn

rate triple that of other hybrid propellants.

CANDLESTICKROCKETS

The academic endeavors offoreign students who enter the U.S.

to study may be limited because of Patriot Act II. The law would allowagencies to engage in surveillance

without court consent to ensurethat universities comply

with the electronic registrationsystem, which also controls a

student’s field of study. That coulddeter foreign scholars from

working on any projects outsidetheir stated field.

PATRIOTICRESTRICTIONS



C.

SHE

RB

UR

NE

Ph

otoL

ink/

Get

ty

newsSCAN meet with denials. In 2002 the number of stu-

dent visas granted was 234,322, down 20 per-cent from 2001. Stuart Patt, a spokespersonfor the State Department, contends that thisdrop reflects an overall downturn of visa ap-plications since 9/11. At the same time, an

APS survey of physics departments learnedthat 13 percent of foreign students admitted(including 20 percent of those from China)were denied visas, with almost half of thosefrom China facing some kind of difficulty.These problems left science faculty scramblingto fill teaching and research positions andthreaten the viability of some small programs.

Part of the reason for the increase in visadenials is that consular officials are being heldpersonally responsible—and possibly crimi-nally liable—if they grant a visa to someonewho goes on to commit a terrorist act. IrvingA. Lerch of the APS points out that consularofficials may not be able to distinguish a be-nign field of study from a related but danger-ous one and would deem it safer to deny visasto most applicants in certain broad categories,such as condensed-matter physics or biotech-nology. (In practice, most students are reject-ed because they cannot prove an intent to re-turn home after they complete their pro-grams.) Even the secretary of state is notempowered to overturn a consular officer’sdenial of a visa. Yet sadly, the State Depart-ment’s inspector general discovered in De-cember 2002 that inadequately trained juniorofficials made most visa decisions and thatthey were too inconsistent in their back-ground checks to foil a determined terrorist.

What may not make it through is U.S. pre-eminence in the physical sciences. In scienceand engineering fields, between 35 to 50 per-cent of doctoral degrees go to foreigners,

many of whom stay: in physics, a third of thefaculty is foreign born. U.S. science gatheredmomentum during World War II, thanks tothe influx of trained Europeans. Several esti-mates attribute fully half the growth in theAmerican economy since then to innovationin science and technology, with “aliens” hav-ing played no small part. Lerch fears that asignificant, permanent reduction in the num-bers of visas for scientists and engineers couldcause a long-term downturn in the economy.

Outsiders have also contributed to de-fense: Albert Einstein and Enrico Fermi,whose ideas lay behind the atom bomb, wereoriginally citizens of then enemy countries.But new regulations prevent foreigners frombeing employed on a host of “unclassifiedbut sensitive” projects in academia and in-dustry. “There are categories of people whocan’t work in certain categories of knowl-edge,” notes the CEO of a security-relatedsoftware firm, who requested anonymity.“There is difficulty getting talented people,across the board.” Concern about staff qual-ifications at the national weapons labs wasalready running high after the Wen Ho Leeaffair at Los Alamos, which promulgatedperceptions of racial profiling that madeeven some U.S. citizens reluctant to apply forpositions at the labs. The new restrictions areexacerbating the problem and, according tothe trade group Information Technology As-sociation of America, could undermine long-term security.

The most immediate concern, however, isthe insensitive implementation of existing reg-ulations. In January the U.S. Immigration andNaturalization Service arrested and detainedPakistani journalist Ejaz Haider for failing toreport for fingerprinting. (All men from cer-tain countries must register for backgroundchecks.) Haider, who had issued warningsabout Islamic holy warriors long before 9/11,was a visiting scholar at the Brookings Insti-tution, a Washington, D.C., think tank, andhad apparently been assured by consular offi-cials that he need not register. The affair madeheadlines in Pakistan. “Everyone here is sur-prised that the INS is not able to distinguishbetween friend and foe,” comments A. H.Nayyar, a physicist at Quaid-e-Azam Univer-sity in Pakistan. “This is very scary for friends.”

Madhusree Mukerjee, who holds a Ph.D. in physics, lives in Montclair, N.J.

According to Stuart Patt, aspokesperson for the U.S.

Department of State, visas arebeing delayed by “interagencyreview”: the department seeks

information on the individual thatother federal agencies might have.

All visa applicants from so-calledstate sponsors of terrorism—Cuba,

Libya, Iran, Iraq, North Korea,Sudan and Syria—and some

applicants from nations that posea risk of nuclear proliferation—

China, India, Israel, Pakistan andRussia—face such scrutiny. Also,consular officials have to be alert

to 16 kinds of potentiallydangerous technologies (specific

aspects of nuclear science,biotechnology, propulsionsystems, lasers, robotics,

materials science, advancedcomputation and others) that

a scholar from any nation mightseek to acquire.

BARRIERSTO STUDY

VISA RESTRICTIONS may also hinder American science.



SAM

UE

L VE

LASC

O

newsSCAN

Some researchers have focused onquantum heat engines thatharness the random twitchings ofparticles—that is, their Brownianmotion. When two reservoirs ofelectrons at different energiesmake contact, electrons flowtoward the colder bath. But randomback-and-forth electron motionsspoil the efficiency of most devicesdesigned to turn that flow intowork. In 2002 Heiner Linke of theUniversity of Oregon and hiscolleagues reasoned that asemiconductor filter for electronsof specific wavelengths couldprevent wasteful energy flow andallow a Brownian heat engine tooperate at or near the Carnot limit.Martin A. Green of the Center forThird Generation Photovoltaics inSydney, Australia, says that hisgroup is currently exploring whethera similar effect could enhance theefficiency of solar cells.

NEED TO KNOW:RANDOM WORK

The fathers of thermodynamics got a lotof mileage from thought experimentsabout gas-filled engines conjoined to

reservoirs of hot and cold. Today a few physi-cists are playing with quantum mechanics inhopes of finding new methods to control andcreate energy flow in quantum versions of thesteam engine. Their research suggests that itis possible to “beat” the inviolable secondlaw of thermodynamics with some quantumsleight-of-hand.

The second law limits the efficiency of anyphysical process. In essence, it states that, toperform work, energy must flow betweentwo reservoirs set at different temperatures.The flow introduces disorder into the system.The temperature difference between the twobaths determines the engine’s maximum, orCarnot, efficiency, named after 19th-centu-ry French physicist Sadi Carnot.

Marlan O. Scully’s quantum optics groupat Texas A&M University has calculated away to extract work from a single heat bath,thereby surmounting the Carnot limit andgiving the appearance of breaking the secondlaw. This setup would rely on photons re-bounding in a small cavity between two mir-rors, one of which would act as a piston. Thebath is a circulating gas of atoms that emitsheat in the form of photons as it passesthrough the mirrored cavity. The atoms areprepared in a special fashion. Each atom hasthree electron states: an excited state and twonearly identical relaxed states that are quan-tum-mechanically mixed. This so-called co-herence interferes with the absorption of pho-tons but permits the emission of photons toproceed unfettered, causing an excess of heatbeyond what the temperature of the atomsalone would dictate.

Coherence is a unique property of quan-tum mechanics that allows laser photons tomarch in lockstep and atoms to be in twostates at once. Thermodynamically speaking,coherence is an extra dose of order, whichpushes the engine out of the uniform state ofequilibrium, where the second law applies.An incoherent photon gas, in contrast,pumps a piston with the usual Carnot effi-ciency, as shown in a study by physicist M.

Howard Lee of the University of Georgia. Scully states that his team’s analysis

demonstrates effects that classical heat en-gines cannot produce—a tiny bit of coherencecan cause a significant boost in work output.Coherence is as fragile as a house of cards,however, and building it up in this case coststhree or four times as much energy, in theform of microwaves, as the engine puts out.These handicaps may mean that the effectprobably will not readily find applications.Nevertheless, Seth Lloyd of the Massachu-setts Institute of Technology comments thatScully’s investigation may help point the waytoward using coherence to bring lasers orthermoelectric refrigerators closer to theirideal efficiency limits. “If quantum coherencecould do that for you, that would be great,”Lloyd notes.

Scully is also attempting to construct andtest a laser-engine hybrid. In 2002 he proposedthat a quantum “afterburner” could squeezeextra work from some ideal engines that op-erate below the Carnot limit if the exhaustatoms were stimulated to produce laser light.

Most physicists see no reason why quan-tum mechanics should damage Carnot’s re-sult. Quantum changes preserve disorder, sothe second law is built in from the beginning,Lloyd observes. Looking for ways to improveheat engines is “a praiseworthy branch ofquantum engineering,” he remarks, “but don’texpect violations of the second law—it’s notgoing to happen.”

JR Minkel is based in NewYork City.

Law and DisorderA QUANTUM STEAM ENGINE GETS AROUND THE SECOND LAW BY JR MINKEL

PH

YSIC

S

QUANTUM POWER: A microwave bath puts atoms in a special state, one that enables some atomsto emit photons (as heat) into the mirrored cavity but not to absorb them. The photons push thepiston to do work; some heat escapes to enable the piston to recompress.

Escapingheat

Piston

AtomsMirror

Mirror

Microwaveheat bath

Path of atoms



ILLU

STR

ATIO

N B

Y M

ATT

CO

LLIN

S; S

OH

O/N

ASA

(top

); J

ER

EM

Y W

OO

DH

OU

SE P

hot

odis

c/G

etty

newsSCAN C L I M A T E

Rising SunHumans may be shoulderingtoo much of the blame forglobal warming, according toa new look at data from sixsun-gazing satellites. Theysuggest that Planet Earth hasbeen drenched in a bath of so-lar radiation that has been intensifying overthe past 24 years—an increase of about 0.05percent each decade. If that trend began ear-ly last century, it could account for a signifi-cant component of the climatic warm-up thatis typically attributed to human-made green-

house gases, says RichardC. Willson of ColumbiaUniversity’s Center forClimate Systems Researchin Coronado, Calif. Will-son concedes that the cli-mate’s sensitivity to suchsubtle solar changes is stillpoorly understood, butthe evidence merits keep-

ing a close eye on both the sun and humansto better gauge their relative influences onglobal climate. “In 100 years I think we’llfind the sun is in control,” he says. His team’sreport appears in the March 4 GeophysicalResearch Letters. —Sarah Simpson

Protons and neutrons would be like twins if not for charge symmetry breaking (CSB), a sub-tle effect that causes neutrons to be 0.1 percent heavier than protons. Had the imbalance gonethe other way, hydrogen would not have survived to form stars. Physicists believe that CSBhinges on the repulsion strength and mass difference between the up and down quarks in-side nuclear particles, but they haven’t been able to pin down the exact values. Now all thepuzzle pieces are in place. Researchers working at the Indiana University Cyclotron Facilityin Bloomington have made the first observation of a long-sought, rare reaction in which twoheavy hydrogen nuclei produce a helium nucleus and a neutral pion, which partially medi-ates the force holding nuclei together. The reaction rate depends on the mass difference andrepulsion interaction. An experiment at the Tri-University Meson Facility in Vancouver hasmeasured another key sign of the cracked symmetry: a slight preference for pions and heavyhydrogen nuclei to fly off in one direction when formed from proton-neutron collisions. The-orists have already begun the arduous calculations required to extract the quark propertiesfrom the results. Both groups announced their findings at the April meeting of the AmericanPhysical Society. —JR Minkel

P H Y S I C S

Why Neutrons Outweigh ProtonsIn March the United Nationsreported on the state of the world’s

freshwater. Population growthcould mean that by the middle of

this century, seven billion peoplein 60 countries could be affected

by a lack of clean water. Yet little isbeing done to confront the

impending crisis.

Percent of the world’s accessiblefreshwater used by humans: 54

Percent estimated to be used by 2025: 70

Percent used for agriculture: 69

For industry (average): 22

For industry, high-incomecountries: 59

For industry, low-incomecountries: 8

Annual number of deaths fromwater-related diseases: 5 million

Annual number sickened by poorwater: 2.3 billion

Available water per person, in liters per day:

Countries with the least:Bahamas: 181

United Arab Emirates: 159Gaza Strip: 142

Kuwait: 27

Country with the most:Greenland: 29.5 million

U.S. (contiguous): 20,300SOURCE: World Water Assessment Program;

s e e w w w . w a t e r y e a r 2 0 0 3 . o r g

DATA POINTS:NOT ALL WET

Having extra color may help male birdswoo mates, but it also attracts predators.Scientists analyzed 21 years of data fromthe North American Breeding Bird Survey,which thousands of volunteers across thecontinent gathered by counting all birds seenor heard during breeding-season mornings.On average, wildlife ecologists find that“dichromatic” bird species die out nearly 25percent more often than their monochromat-ic relatives. Two-toned birds in general don’tgo completely extinct, because as soon as onespecies vanishes from a neighborhood, other

colorful ones take itsplace. The report, ap-

pearing in the April 17 Pro-ceedings of the National Academy of SciencesUSA, supports the theory that species understrong sexual-selection pressure face greaterrisks of becoming locally extinct and suggeststhat human activities that block migrationscould jeopardize the survival of dichromaticspecies. —Charles Choi

GLOBAL WARMER? The sun as seen in extreme ultraviolet.

PAINTED BUNTING andother colorful birds go locally extinct 25 percent more oftenthan their drab cousins.

E C O L O G Y

Sexy and Delicious



CO

UR

TESY

OF

CAL

VIN

SIM

ER

LY (

top

); U

ND

ER

WO

OD

& U

ND

ER

WO

OD

Cor

bis

(b

otto

m)

newsSCAN

■ The oldest verifiably original DNAhas been found in Siberiansediment. The plant geneticmaterial may be 400,000 yearsold and harbor clues about thepaleoenvironment.

Science, April 18, 2003

■ A survey has found catastrophicdecreases in gorillas andchimpanzees in westernequatorial Africa, the laststronghold of these apes.Hunting and, more recently, anEbola outbreak appear to be theprimary causes.

Nature, April 10, 2003

■ In a mouse study, the drugAccutane prevented theaccumulation of lipofuscin, atoxin that causes the maculardegeneration of Stargardt’sdisease. The drug, which canproduce night blindness as aside effect in acne treatment,apparently mimicked the effectsof light deprivation.

Proceedings of the National Academyof Sciences USA (online), March 17, 2003

■ Fertilizing the ocean with iron toencourage the growth ofplankton, which absorb carbondioxide from the air, may not bevery effective in sequesteringthe greenhouse gas, because thecarbon may not sink deepenough to remain locked away.

Science, April 4, 2003

BRIEFPOINTS

B I O T E C H

Corrupted ClonesDespite claims by a UFO cult and mav-erick physician Severino Antinori, mostscientists think today’s cloning methodscannot make a viable baby. Now newresearch suggests that cloning of pri-mates could very well be impossible. Sci-entists have attempted to clone rhesusmacaques, but none of the resulting em-bryos survived implantation in a surro-gate mother. The researchers found thatalthough cell division appeared superfi-cially normal, chromosomes were div-vied up unevenly: some cells ended upwith too many and others too few. Usu-ally spindles of protein tubes help to pullopposite ends of a dividing cell apartand ensure that chromosomes split upequally. In normal rhesus egg cells, vi-tal spindle proteins are concentrated near the eggs’ chromosomes, which are inadvertently re-moved during the first steps of each of the four different nuclear-transfer techniques the in-vestigators tried. The location of the spindle proteins could make cloning embryonic stemcells difficult “and reproductive cloning unachievable,” the researchers say in the April 11issue of Science. —Charles Choi

P S Y C H O L O G Y

The Unusual SuspectsThe more confidence an eyewitness has when identifying a suspect, the stronger that evidencetypically becomes in court. But revelations about the malleability of people’s memories areupsetting this conventional wisdom. In a recent experiment conducted at Iowa State Univer-sity, all 253 participants who watched a staged crime video chose a suspect from a six-manlineup—even though the true culprit was not among them. Unaware that they were mistak-

en, witnesses who were then told,“Good, you identified the sus-pect,” tended to further overstatetheir confidence and recollection ofdetails, including the criminal’s fa-cial features. The false certaintyprevailed whether they heard theaffirmation immediately followingthe lineup or a full 48 hours later.The Iowa researchers concludethat law enforcers must curb on-the-spot comments about suspectsand secure statements about a wit-ness’s confidence right away toavoid tainting future testimony.The report appears in the MarchJournal of Experimental Psychol-ogy: Applied. —Sarah Simpson

CRIMINAL LINEUPS have been around for decades—this one is from Chicago in 1927—but police comments to witnessesafterward may produce false certainty in the testimony.

BAD SPLITTING: Mitotic spindles (red) separate chromosomes(blue) in cloned macaque eggs undergoing cell division. The chromosomes do not divide properly in primates.



RO

DG

ER

DO

YLE

newsSCAN

Many think that globalization is a recentdevelopment, but its origins go backto the early 19th century. This fact is

apparent from a new study by sociologistChristopher Chase-Dunn of the University ofCalifornia at Riverside and his colleagues.Their data, which are based on the relationbetween imports and gross domestic product,show that the initial wave of globalization be-gan about 1830 and peaked about 1880. Dur-ing this time, international commerce, withthe abandonment of mercantilism, first be-came a force in the lives of ordinary people.Before the 19th century, international tradewas a paltry affair mostly confined to luxuries,such as spices and tobacco. This early wave isassociated with the growth of railroads, moreefficient ocean transport,and the political victory ofmanufacturing and tradinginterests over those of thelandowners, signaled by the1846 repeal of the Britishcorn laws. (Those laws im-posed duties on importedcorn and thereby kept priceshigh.) The second wave co-incided with the rise of elec-tricity and steel around1900 and peaked in the1920s. The current wavebegan after World War II asa result of the creation of in-ternational institutions such as the GeneralAgreement on Tariffs and Trade, the prede-cessor of the World Trade Organization.

Decreasing costs of transport and com-munication underlie the long-range increasein world trade, but no satisfactory reason ex-plains the wave pattern. Chase-Dunn cites“hegemonic stability,” in which a great powerprovides stable conditions. The first and thirdcoincide with, respectively, the eras of Britishand U.S. hegemony, but, as he notes, the theo-ry does not account for the second wave, whichoccurred when Britain was in relative declineand the U.S. had not yet asserted its power.

During most of the 19th and early 20thcenturies, America did not follow Britain’s

free-trade policy but instead imposed high tar-iffs to protect manufacturing. The U.S. be-came more open to imports only after WorldWar II. But it still lagged behind other majorcountries in trade participation—not surpris-ing considering its vast domestic market,which could supply a larger variety of de-mands than smaller economies could. Never-theless, the greater involvement of the U.S. hasbeen the primary factor in world trade ex-pansion since World War II.

In the long run it is very likely that inter-national trade will continue to expand as thecosts of transport and communication con-tinue to decline. Perhaps the most formidableobstacle to trade growth in the near future isfailure to reform government practices that

foster doubt and mistrust. Transparency In-ternational, an organization funded by sever-al European governments, polls well-informedindividuals in more than 100 countries regard-ing the extent of misuse of public power forprivate benefit. Its 2003 report shows that trustin the institutions of industrial nations averages7.3 out of a perfect score of 10; developingcountries average only 2.3. How governmentsdeal with this issue of integrity could largelydetermine the next phase of worldwide trade.

Next: Globalization’s winners and losers.

Rodger Doyle can be reached [email protected]

Trade GlobalizationIT IS NEARLY TWO CENTURIES OLD AND LIKELY TO CONTINUE BY RODGER DOYLE

BY

THE

NU

MB

ER

S

The 12 Percent of Largest World Goods Trading and ServicesCountries in 2000

Imports Exports

U.S. 18.7 14.1

Germany 8.1 8.3

Japan 6.3 7.2

U.K. 5.5 5.2

France 4.6 5.0

Canada 3.6 4.1

Italy 3.7 3.9

China* 2.6 3.3

Netherlands 3.2 3.4

Belgium 2.7 3.0

Korea 2.5 2.7

Spain 2.3 2.2

* E x c l u d e s s e r v i c e s , w h i c h a c c o u n tf o r a b o u t 2 0 p e r c e n t o f

i n t e r n a t i o n a l t r a d e w o r l d w i d e .

S O U R C E S : I n t e r n a t i o n a l T r a d eS t a t i s t i c s Y e a r b o o k , 2 0 0 0 , U n i t e d

N a t i o n s ; O E C D S t a t i s t i c s o nI n t e r n a t i o n a l T r a d e i n S e r v i c e s ,

1 9 9 9 – 2 0 0 0 .

BIG-TIMEPLAYERS

WORLD

SOURCE: Christopher Chase-Dunn et al., 2000. Data are shown as five-year moving averages.

Impo

rts

as P

erce

nt o

f To

tal D

omes

tic P

rodu

ct

U.S.

1800 1840

22

20

18

16

14

12

10

8

6

4

2

01880

Year1920 1960 2000

Third wave

Second wave

First wave of globalization

Trade Globalization since 1795:Waves of Integration in the World-System. Christopher Chase-Dunn,

Yukio Kawano and Benjamin D.Brewer in American Sociological

Review, Vol. 65, No. 1, pages 77–95; February 2000.

Globalization, Trade, andDevelopment: Some Lessons from History. Alan M. Taylor.

National Bureau of EconomicResearch working paper 9326;

November 2002.

FURTHERREADING


At the beginning of the movie that made Leonardo Di-Caprio a megastar, a camera-toting unmanned robotventured into a cavernous hole in the wreck that sits onthe bottom of the Atlantic, 12,640 feet from the surface.The 500-pound vehicle, christened Snoop Dog, couldmove only about 30 feet along a lower deck, hamperedby its bulky two-inch-diameter tether hitched to a sub-marine that waited above. The amount of thrust neededto move its chunky frame stirred up a thick cloud. “Thevehicle very quickly silted out the entire place and madeimaging impossible,” director James Cameron recalls.

But the eerie vista revealed by Snoop Dog on that1995 expedition made Cameron hunger for more. Hevowed to return one day with technology that could ne-gotiate anyplace within the Titanic’s interior.

In the past six months two documentaries—one forIMAX movie theaters called Ghosts of the Abyss, theother, Expedition: Bismarck, for the Discovery Chan-

nel—demonstrated the fruits of a three-year effort thatCameron financed with $1.8 million of his own mon-ey to make this vision materialize. The payoff was two70-pound robots, named after Blues Brothers Jake andElwood, that had the full run of two of the world’s mostfamous wrecks, the Titanic and the Bismarck, whichthey visited on separate expeditions.

The person who took Jake and Elwood from dreamto robot is Mike Cameron, James’s brother, an aero-space engineer who once designed missiles and who alsopossesses a diverse background as a helicopter pilot,stunt photographer and stuntman. (Remember thecorpse in the movie The Abyss, from whose mouth acrab emerges?) Giving the remotely operated vehiclesfreedom of movement required that they be much small-er than Snoop Dog and that the tether’s width be tapereddramatically so as not to catch on vertical ship beams.

Mike Cameron took inspiration from the wire-guid-ed torpedoes used by the military that can travel for manymiles. His team created vehicles operable to more than20,000 feet (enough to reach as much as 85 percent of theocean floor). The dimensions of the front of the robotare 16 inches high by 17 inches across, small enough tofit in a B deck window of the Titanic. The bots have aninternal battery so that they do not need to be poweredthrough a tether. Instead the tether—fifty-thousandthsof an inch in diameter—contains optical fibers that relaycontrol signals from a manned submersible vehicle hov-ering outside and that also send video images in the oth-er direction. The tether pays out from the robot, a designthat prevents it from snagging on objects in the wreck.

James Cameron thought the project would be astraightforward engineering task, not much harder thandesigning a new camera system. “This turned out to bea whole different order of magnitude,” he says. “Therewas no commercial off-the-shelf hardware that wouldwork in the vehicles. Everything had to be built fromscratch.” If the team had known this early on, he added,“we wouldn’t have bothered.” Water pressure on the


PH

OTO

GR

APH

S C

OU

RTE

SY O

F W

ALT

DIS

NE

Y P

ICTU

RE

S

Innovations

The Abyss Transit SystemJames Cameron commissions the making of robots for a return to the Titanic By GARY STIX

LITTLE EYES: Remotely controlled robots use onboard cameras to explore the deepest innards of sunken vessels.


cable that carried the optical fibers could create micro-scopic bends in the data pipe, completely cutting off thecontrol signals from the submersibles. Dark Matter inValencia, Calif. (Mike Cameron’s company), had to de-vise a fluid-filled sheath around the fiber to displace theminuscule air pockets in the cable that could lead to themicrobending.

To save weight, the frame—similar to a monocoquebody of a race car—was made up of small glass hollowspheres contained in an epoxy matrix. The thruster con-tained a large-diameter, slowly rotating blade with noz-zles that diffused the propulsive flow, minimizing thechurning that would otherwise disturb the caked silt. Ahigh-resolution video camera, along with an infraredcamera for navigation, was placed in the front of thecraft along with three light-emitting-diode arrays for filllighting and two quartz halogen lamps for spotlighting.

The winter of 2001 marked a critical juncture. Itwas six months before dives to the Titanic could be safe-ly attempted, and James had to determine whether toproceed or wait another year. “Mike was really, reallynegative on the idea, but I decided to go for it,” the di-rector says. He felt he couldn’t afford to wait longer andthought that a fixed deadline would focus the engi-neering staff at Dark Matter. For his part, Mike wascontending with an unending series of design chal-lenges. “It was such an overwhelming set of problemsthat I had very little confidence that certain parts wouldbe solvable in the time we had,” Mike says.

A few weeks before the dives commenced in the sum-mer of 2001, the robots’ lithium sulfur dioxode–basedbatteries caught fire while being tested in a pressure tank,destroying what was to have been a third robot. Mikewanted to delay the dives, but James found a supplierof another type of lithium battery and pressed ahead.

At the dive site, Jake and Elwood took starring roleswith their 2,000-foot tethers, exploring for the first timein about 90 years remote parts of the ships, including theengine room, the firemen’s mess hall and the cabins offirst-class passengers—even focusing in on a bowler hat,a brass headboard and an intact, upright glass decanter.The images lack the resolution and novel quality of thehigh-definition, three-dimensional IMAX images, theother major technological innovation of Ghosts of theAbyss. Jake and Elwood’s discoveries, however, drawthe viewers’ interest because of what they convey of theTitanic’s mystique. “You actually feel like you’re outthere in the wreck,” Mike says. He remembers his broth-er piloting the robots with the helicopter stick that hadbeen installed in the Russian submersible from which therobots were launched. “Jim ended up being a cowboy

pilot,” Mike says. “He was far more aggressive with thesystem than I was.”

One scene in Ghosts of the Abyss reveals the tensionthat sometimes erupted between the brothers. Jamescontemplates moving one of the robots through a cab-in window that is still partially occluded by a shard ofglass that could damage the vehicle or cut the data teth-er. When James declares that he is going to take Jake in,moviegoers can hear Mike pleading with his brother notto do it, ultimately relenting once the bot has negotiat-ed the opening.

The decision to install a new type of battery at the lastminute came to haunt the expedition; Elwood’s lithium-

polymer battery ignited while in the bowels of the ship.James manipulated the remaining robot into the Titan-ic to perform a rescue operation by hooking a cord to thegrill of the dead bot and towing it out. At the surface—

on the deck of the Russian scientific vessel the Keldysh,from which the two submarines carrying Jake and El-wood to the Titanic were launched—Mike rebuilt El-wood with a backup battery. During the next dive, therobot caught fire again while it was still mounted on thesubmarine, endangering the crew. Finally, Mike workedfor an 18-hour stretch to adapt a lead-acid gel batteryused for devices onboard the mother ship into a powersource for Elwood, enabling the expedition to continue.

The bots, now fitted with a new, nonflammable bat-tery that Mike designed, may find service beyond mo-tion pictures. The U.S. Navy has funded Dark Matter tohelp it assess the technology for underwater recovery op-erations of ships or aircraft. The bots also have potentialfor scientific exploration of deep-sea trenches. After trav-eling to the Titanic and the Bismarck, the team went onto probe mid-Atlantic hydrothermal vents, discoveringmollusks in a place where scientists had never encoun-tered them before. As adventure aficionados, the broth-ers speculate that a descendant of Jake and Elwoodmight even be toted on a mission to Europa, one of Ju-piter’s moons, to investigate the waters that are suspect-ed to exist below its icy shell. The Cameron siblings, whotinkered with home-built rafts and rockets as children inOntario near Niagara Falls, hope to be around longenough to witness their robotic twins go from the bot-tom of the ocean to the depths of space.


Robots Jake and Elwood, with their 2,000-foot

tethers, took on starringroles, exploring the remotest

reaches of the Titanic.


Academic biologists routinely work with genes with-out so much as a second thought. They focus their at-tention on determining the function of the gene and theprotein it produces, not on whether the relevant DNAis patented or not. According to some universities andscholarly associations, a recent federal appeals courtdecision means that the entire scholarly research com-

munity, not just biologists, willbe spending a lot more timewith lawyers to determinewhether their investigations vi-olate someone’s patent rights.

Universities have often la-bored under the assumptionthat using research tools andmaterials is a permissible prac-tice: noncommercial uses fallunder a research exemptionthat precludes liability for pa-tent infringement. The Court ofAppeals for the Federal Circuit(CAFC), which hears appealsof patent cases, issued a rulinglast year that defines this safe

haven for researchers so narrowly that it becomes vir-tually useless. It reiterated that the exemption appliesonly “for amusement to satisfy idle curiosity, or forstrictly philosophical inquiry.” But noncommercial,academic research, the court decided, serves to furtherthe “legitimate business objectives” of the university,so patented equipment and materials do not warrantan exemption.

Some universities fear that researchers will nowhave to devote time and grant money to conductingpatent searches and arranging licensing agreements be-fore proceeding with their experiments. The rationalefor a research exemption is based on legal opinions is-sued by judges that date back to 1813. But many mem-bers of the legal community argue that despite these

precedents, the exception for academics was alwaysvery narrow. “It’s a widespread urban legend that thisresearch exemption would protect university re-search,” notes Lynn H. Pasahow, an intellectual-prop-erty attorney with the law firm Fenwick & West inMountain View, Calif. The case for the exemption hasalso been weakened because universities are now in-volved more than ever in obtaining licensing revenuesfor discoveries made by their researchers.

What spurred the CAFC’s ruling was a suit broughtby John Madey, the inventor of the free-electron laser.He brought a claim against Duke University after thatinstitution removed him as head of a laboratory in1997 and he had moved to another university. In thecomplaint, he charged that Duke had, among otherthings, violated his patent rights by continuing to usethe laser equipment. On appeal, the CAFC sent the caseback to a lower court, saying that it had erred by us-ing an overly broad interpretation of the research ex-emption to decide that Duke had not infringed Ma-dey’s patents. Duke has asked the U.S. Supreme Courtto review the CAFC’s decision, and a number of otheruniversities and associations have supported Duke byfiling friend of court briefs.

Not everyone is worried. Lita Nelsen, director of thetechnology licensing office at the Massachusetts Instituteof Technology, expects the fallout from the case to beminimal—most individuals and companies are not in-terested in bringing high-profile suits against academicresearchers. “Most of the time the researcher doesn’tknow a patent exists and is not doing any harm,” Nelsensays. “If somebody comes to us and says, ‘I hold thepatent, and I want you to stop doing something,’ the firstthing I’d say is ‘Do you care?’ And if the person did, we’drespect the patent.” Even if the Supreme Court does nottake up the case, Madey v. Duke may trigger useful de-bate—and may lead to a push for legislation that clari-fies when universities and nonprofits may conduct re-search without first making a call to their attorney.


JEN

NIF

ER

KAN

E

Staking Claims

Sign HereWill a scientist need a legal opinion before starting the next experiment? By GARY STIX



BR

AD H

INE

S

Skeptic

In the epilogue of In Memoriam A.H.H., Alfred, Lord Tennysoncaptured the essence of the quest for a single unifying principleand purpose in nature: “One God, one law, one element,/Andone far-off divine event,/To which the whole creation moves.”

The noble dream of finding teleological succor in the marchof time has become big business, as demonstrated by worksfrom Hal Lindsey’s 1970s blockbuster The Late Great PlanetEarth to today’s Left Behind series, by Tim LaHaye and Jerry B.Jenkins. (Both are said to have sold in the tens of millions.) Andif you can sprinkle your homiletics with scientistic jargon,

so much the better. The latestand most egregious example ofthe (mis)use of science in the(dis)service of religion is Mi-chael Drosnin’s Bible Code II,enjoying a lucrative ride on theNew York Times best-seller list,as did the 1997 original.

According to proponents ofthe Bible Code—itself a subset

of the genre of biblical numerology and Kabbalistic mysticismpopular since the Middle Ages—the Hebrew Pentateuch can bedecoded through an equidistant-letter-sequencing software pro-gram. The idea is to take every nth letter, where n equals what-ever number you wish: 7, 19, 3,027. Print out that string of let-ters in a block of type, then search left to right, right to left, topto bottom, bottom to top, and diagonally in any direction forany interesting patterns. Seek and ye shall find.

Predictably, in 1997 Drosnin “discovered” such currentevents as Yitzhak Rabin’s assassination, Benjamin Netanyahu’selection, Comet Shoemaker-Levy 9’s collision with Jupiter,Timothy McVeigh and the Oklahoma City bombing, and, ofcourse, the end of the world in 2000. Because the world did notend and current events dated his first book, Drosnin continuedthe search and learned—lo and behold—that the Bible predict-ed the Bill and Monica tryst, the Bush-Gore election debacleand, of course, the World Trade Center cataclysm.

Just like the prophecies of soothsayers past and present, allsuch predictions are actually postdictions (note that not one psy-

chic or astrologer forewarned us about 9/11). To be tested sci-entifically, Bible codes would need to predict events before theyhappen. They won’t, because they can’t—as Danish physicistNiels Bohr averred, predictions are difficult, especially about thefuture. Instead, in 1997 Drosnin proposed this test of his the-sis: “When my critics find a message about the assassination ofa prime minister encrypted in Moby Dick, I’ll believe them.”

Australian mathematician Brendan McKay did just that, lo-cating no fewer than nine political assassinations secreted in thegreat novel, along with additional discoveries in War and Peaceand other tomes (see cs.anu.edu.au/~bdm/dilugim/moby.html).American physicist David E. Thomas predicted the ChicagoBulls’s NBA championship in 1998 from his code search of LeoTolstoy’s novel. He also recently unearthed “the Bible code isa silly, dumb, fake, false, evil, nasty, dismal fraud and snake-oilhoax” from Bible Code II (see www.nmsr.org/biblecod.htm).

If there is an encrypted message in all this numerologicalpoppycock it is this: there is a deep connection between howthe mind works and how we perceive the world works. We arepattern-seeking animals, the descendants of hominids who wereespecially dexterous at making causal links between events innature. The associations were real often enough that the abili-ty became engrained in our neural architecture. Unfortunately,the belief engine sputters occasionally, identifying false patternsas real. The habit of faltering may not be enough to prevent youfrom passing on your genes for detecting false positives to thenext generation, but it does create superstitious and magicalthinking. This process is coupled to the law of large numbersthat accompanies our complex world, where, as it is said, mil-lion-to-one odds happen eight times a day in New York City.

Given our propensity to look for patterns in a superfluity ofdata, is it any wonder that so many are taken in by such codi-fied claptrap? The problem is pervasive and a permanent part ofour cognitive machinery. The solution is science, our preeminentpattern-discriminating method and our best hope for detectinga genuine signal within the noise of nature’s cacophony.

Michael Shermer is publisher of Skeptic (www.skeptic.com)and author of Why People Believe Weird Things.

Codified ClaptrapThe Bible Code is numerological nonsense masquerading as science By MICHAEL SHERMER

Just like theprophecies of

soothsayers pastand present, all

such predictionsare actually

postdictions.


The Saddam Center for Biotechnology on the campus ofBaghdad University boasted a state-of-the-art facility,replete with surreptitiously imported equipment for am-plifying tiny amounts of DNA and running tests withgels to determine protein sizes. “It looked like you werewalking into a laboratory in one of the better-equippedU.S. institutions,” remembers Rocco Casagrande, whobegan his trips as a United Nations inspector to variousIraqi facilities in mid-December 2002.

The lab was ideal for performing DNA amplifica-

tion using the polymerase chain reaction (PCR) to makecountless copies of genes. Oddly, though, the only thingthese expensive machines were being used for was ge-netic fingerprinting of goats involved in what the Iraqissaid were in vitro fertilization experiments. Iraq doesnot suffer from problems with goat fertility. An infer-tile goat is eaten for dinner, not sent to an IVF clinic.Casagrande and the others took samples from the laband combed through records on a computer hard disk,to no avail. No evidence of cloning genes for makingbioweapons was found. They speculated that the facil-ity could be used for human cloning, but in the end theynever figured out its real purpose.

This experience was not the only time during histhree-month stay that Casagrande encountered projectsthat did not quite make sense. But neither did the bio-logical weapons inspection team come across the an-thrax, botulinum or any pathogen that had been part ofthe notorious program that the Iraqi governmentclaimed was now defunct.

Casagrande was one of about 10 U.S. representa-tives on the roughly 100-member team of nuclear,chemical, biological and missile inspectors, a contrastto the investigations in the 1990s, when many more ofthe officials were American. Every day the team receivedlists of sites to visit from U.N. headquarters in NewYork City. Some destinations were obvious, such as thebiotechnology center; others were gleaned from intelli-gence reports. Once they arrived, a few inspectors con-ducted interviews while the rest looked for suspiciousactivity. Casagrande and his colleagues became famil-iar faces to Iraqis in the months immediately precedingthe war. At night Iraqi television broadcast extensivecoverage, identifying the inspectors by name and coun-try of origin. Casagrande couldn’t go into a restaurantor shop without being recognized.

This 29-year-old—only a few years beyond a doc-torate in biology from the Massachusetts Institute ofTechnology—had the job of refitting the biological


JAM

ES

SALZ

ANO

Profile

One Last LookAlthough United Nations weapons inspector Rocco Casagrande and his colleagues found nobioweapons in Iraq, they could sense that the government had not come clean By GARY STIX

Insights

■ Member of the 20-person U.N. bioweapons inspection team that visitedabout 150 Iraqi sites from December 2002 through early March of this year.

■ Sites visited included: breweries, dairies, hospitals, airfields, ammunitiondumps, pharmaceutical manufacturers and a tomato cannery.

■ “I’d never be comfortable leaving the country and saying Iraq doesn’t have biological agents. It wasn’t behaving like a country that doesn’t havebiological weapons.”

ROCCO CASAGRANDE: WITNESS TO HISTORY



HU

SSE

IN M

ALLA

AP

Ph

oto

analysis laboratory used to test samples taken during its dailytours. U.N. inspectors had last operated the lab in 1998, beforethey were expelled from Iraq; in the interim, it had become anesting place for pigeons.

As a child growing up in a Philadelphia suburb, Casagrandewas fascinated with both science and history. While doing hisdoctoral work, he became involved with the Harvard SussexProgram’s Chemical and Biological Warfare Colloquium, ledby Harvard University microbiologist Matthew Meselson, andrealized that cultivating an expertise in biowarfare would serveas a means to combine his two interests. Before he left for Iraq,Casagrande was developing weapons biodetectors for SurfaceLogix, in Cambridge, Mass.

The U.S. State Department tooknotice of Casagrande after he wrote ar-ticles for Nonproliferation Review andBioscience chronicling the potentialthreat of biowarfare against crops andlivestock. The government later rec-ommended him for the U.N. post. (Heis now employed by Abt Associates inCambridge, setting up a homeland se-curity consulting practice.)

In Iraq, Casagrande was alwaysaware that a positive result on any ofthe countless assays performed in the re-furbished lab could reverberate aroundthe world. “We tried not to think aboutwhat the implications were for what wemight find,” he says. “But we couldn’thelp but realize that this could be a turning point in history.”The 20 or so members of the bioweapons team, one fifth of thetotal complement of inspectors, which also included chemical,missile and nuclear contingents, visited sites ranging from brew-eries to munitions facilities. The inspectors had been taught dur-ing training sessions in Geneva how mundane equipment forproducing such routine items as beer or tomato sauce could alsobe employed for culturing anthrax or another bioweapons agent.

Interviews often took on a surreal quality that reflected thedeep-seated fear that gripped the populace. A simple question—

“How long have you been head of this facility?”—could elicit afive-minute answer that never addressed the original query. Re-actions could turn hostile. Casagrande recalls speaking with thehead of an agricultural research center whose director accusedhim of being personally responsible for sanctions against thecountry. Technicians sometimes flatly disavowed the presenceof certain microbes. In one instance, Casagrande had to repri-mand a worker in a university lab who had refused to ac-knowledge possession of a strain of anthrax that was found bythe inspectors (the strain could be used only for making a vac-cine, not a bioweapon).

Iraqi “minders” constantly tracked the inspectors and fol-

lowed them everywhere. A certain warmth developed betweenthe two groups. On a trip south of Baghdad the inspectors hadto wait endlessly as the Iraqis fished in two bags of unmarkedkeys to enter 150 triple-padlocked cinderblock buildings thatturned out to house just conventional ammunition. One min-der told his charges that if an invasion ever came, it would beimportant to give the Iraqis three months’ notice so that theywould have enough time to open the bunkers. “It was actuallya friendly relationship, but we understood they weren’t ourfriends,” Casagrande comments. Chumminess had its risks. Anescort, whom the inspectors knew as Mr. Wa’ad, remarked tothe team that he envied his relatives who had emigrated to the

U.S. “He soon disappeared,” Casa-grande says.

Some of the Iraqi scientists were ea-ger to exchange information with out-siders and might launch into a discus-sion with Casagrande about bacterialindicators of soil health. He even metthe infamous Rihab Taha, the formerdirector of the country’s biologicalweapons program, who related to theinspectors that she now spends her timecaring for her children as a Baghdadhomemaker. (Women headed about athird of the civilian laboratories visited.)

During their stay, the team mem-bers never uncovered what they wereseeking. Still, Casagrande came awaywith a distinct uneasiness. It seems

unimaginable to him that a government so obsessed with doc-umentation—the moving of a centrifuge from one room to an-other required extensive paperwork—would be unable to ac-count for how it disposed of pathogens from its previousbiowarfare program and to reveal what it did with large quan-tities of growth media used to culture pathogenic agents.

There were places the inspectors did not look. Bioweaponscould have been secreted in off-limits religious sites. Iraq was, infact, in the midst of a mosque-building boom, including the re-cently completed Mother of All Battles mosque, with minaretsshaped to resemble Scud missiles. Also questionable was the dis-covery of a possible smallpox vaccination program. “It makesyou wonder why someone in Iraq thought they needed to be vac-cinated against smallpox,” says Casagrande, who as a U.N. in-spector chose not to offer an opinion about the U.S.-Iraq war.

Despite the frustrations, Casagrande feels that the work wasnot wasted. The inspectors had their stay cut short. But infor-mation that they gathered might help in conducting follow-upinvestigations to unravel the extent of the regime’s conjecturedclandestine programs to cultivate anthrax, botulinum and oth-er mass killers. After all, those supposed weapons stocks wouldbe the after-the-fact basis for waging a war.


JUST OIL? U.N. weapons inspector Rocco Casagrandeexamines dilapidated oil barrels on a farm in Juwesma,Iraq, 26 miles southwest of Baghdad, in January.


SHOOTTHIS DEER

WHITE-TAILED DEER were huntedextensively in Wisconsin’s eradicationzone; this deer, spared because it liveson a game farm, nonetheless remainsquarantined indefinitely.



DAV

ID N

EVA

LA

Last year the Wisconsin Department of Natural Resources in-stituted special hunting periods to try to wipe out upward of18,000 deer. During the fall, dead deer were taken to registra-tion areas, where state employees in protective suits and glovesdragged carcasses from pickup trucks and lifted them onto plas-tic-covered picnic tables. With hacksaws, they severed the heads,double-bagged them and sent them for testing; the bodies them-selves were incinerated.

The Dairy State’s massacre is an attempt to keep a fatal ail-ment known as chronic wasting disease (CWD) from infectingits other 1.6 million deer. The testing enables wildlife officials toascertain the scope of the epidemic—running at nearly 1.6 per-cent—and determine whether the culling can slow the spread.Currently no practical live test exists to check whether an ap-parently healthy, wild animal is actually incubating the sickness;only a brain sample will do.

The disease occurs because a pathogen peppers neural tis-sue full of microscopic holes and gums up the brain with toxicclumps of protein called amyloid plaques. Long confined to apatch of land near the Rocky Mountains, the disease has shownup in 12 states and two Canadian provinces. The sickness pass-es readily from one deer to another—no deer seem to have a nat-ural resistance. “From everything we’ve seen,” commentsMichael W. Miller, a CWD expert with the Colorado Divisionof Wildlife, “it would persist. It would not go away on its own.”

The urgency also reflects concern about the nature of CWD,which belongs to the same family as a better-known scourge:bovine spongiform encephalopathy (BSE), or mad cow disease.Spread by animal-based feed inadvertently containing tissuefrom sick cows and sheep, BSE emerged in the U.K. in the 1980s

and continues to plague that country at a low level. (Nearly twodozen other countries have now also reported cases.) In 1996scientists realized that BSE can pass to humans who eat infect-ed meat, leading to a fatal condition: variant Creutzfeldt-Jakobdisease, or vCJD (distinct from the more common sporadic CJD,which arises spontaneously in one in a million people). Re-searchers are now trying to figure out whether CWD could in-fect humans and livestock and thereby create an American ver-sion of the U.K.’s mad cow disaster.

Pathological ProteinTHE DISEASE AGENT COMMON to all these maladies is theprion (“PREE-on”), a term coined in 1982 by Stanley B. Prusin-er of the University of California at San Francisco. The prionis a protein that exists in all animals, although the exact aminoacid sequence depends on the species. It takes one of twoshapes. Folded correctly, it is the normal prion protein (PrP),which is especially abundant in brain cells and may help pro-cess copper. Folded incorrectly, the prion protein becomes apathogenic entity that kills. The malformed protein has the abil-ity to refold copies of normal PrP in its own image, therebymaking more of itself.

Prusiner’s conception of prions initially met with great skep-ticism. That a pathogen could replicate and pass on its traitswithout assistance from nucleic acids (DNA or RNA) violatedthe orthodoxy of molecular biology. But enough evidence hasaccumulated to prove that some proteins can in fact copy them-selves and that variants of PrP are essential players in spongi-form encephalopathies.

That prions lack any DNA or RNA is also the prime reason

A place called the eradication zone, lying about 40 miles west of Madison, Wis., coverssome 411 square miles. There thousands of white-tailed deer live—or rather, used to live.

Chronic wasting disease, a cousin of mad cow disease, isspreading among wild deer in parts of the U.S. Left unchecked,

the fatal sickness could threaten North American deerpopulations—and maybe livestock and humans

By Philip Yam


why they are so tough. Germicidal light,formaldehyde baths and boiling water allpromptly disrupt bacterial and viral nu-cleic acids, yet such treatments have littleeffect on malformed prions. Researchershave exposed prion-contaminated tissueto a dry heat of 600 degrees Celsius andleft it buried for three years, only to findthat the material, though greatly weak-ened, was still infectious. Indeed, physi-cians have unwittingly passed prion dis-eases on to patients via surgical instru-

ments and transplanted organs that hadundergone standard sterilization proce-dures. (Prion disinfection requires extend-ed heating or corrosive chemicals such assodium hydroxide.)

Foothold in the FoothillsTHE RESILIENCE OF misfolded prionsappears to be a key reason why chronicwasting disease has persisted and spreadfrom its presumed starting point near FortCollins, Colo. There, in 1967, at thestate’s Foothills Wildlife Research Facili-ty, CWD made its first recorded appear-ance, in captive mule deer that were beingmaintained for nutritional studies (muledeer are the most common type in theWest). As the name of the disease sug-gests, affected deer lose weight over thecourse of weeks or months. They oftenbecome thirsty, which drives them todrink large amounts of water and, conse-quently, to urinate a great deal; they alsostart slobbering and drooling. They maystop socializing with fellow deer, becomelistless or hang their heads. Death typi-cally ensues three to four months after

symptoms start, although some victimsexpire within days and others in about ayear. The incubation period, duringwhich the animals show no symptoms,ranges from about 20 to 30 months.

The Fort Collins facility became aCWD death trap. Between 1970 and1981, 90 percent of the deer that stayedmore than two years died from the diseaseor had to be euthanized. In 1980 thescourge emerged outside Colorado, at theSybille Research Unit in southeastern

Wyoming, 120 miles northwest of FortCollins. The two facilities had exchangeddeer for breeding purposes, thus indicat-ing that the disease was infectious—evento a different species: soon the elk at thefacilities contracted the disease. (Deer andelk both belong to the cervid family.)

For years, researchers thought CWDresulted from nutritional deficiencies, poi-soning, or stress from confinement. But in1977 Elizabeth S. Williams, studying forher doctorate at Colorado State Universi-ty, discovered that this view was mistak-en. When Williams looked at brain slicesfrom infected animals, she saw that thetissue was full of microscopic holes. “Ihappened to be taking a course in neu-ropathology and had studied a lot of brainlesions,” she recalls. The holes were un-mistakably like scrapie, the sheep sicknessthat was the first documented spongiformencephalopathy.

In fact, CWD appears to have origi-nated from scrapie. Richard E. Race ofthe National Institutes of Health RockyMountain Laboratories in Hamilton,Mont., conducted test tube studies that

revealed no distinction between the mal-formed PrP of scrapie sheep and CWDcervids. Consistent with this discovery,Amir Hamir of the U.S. Department ofAgriculture’s National Animal DiseaseCenter in Ames, Iowa, found no differ-ence in the appearance of brain samplesfrom elk with CWD and elk experimen-tally infected with scrapie. (BSE alsoprobably arose from scrapie, after cowsate feed derived from infected sheep.)

But unlike BSE in cows (or vCJD in

humans), the cervids were not getting illfrom their food. CWD behaves more likescrapie, in that the sickness spreadsamong individuals, although no one real-ly knows how it does. The prions couldlurk in the urine. During rutting season,deer bucks lap up the urine of perhapsdozens of does to find out which are inheat. Elk females lick males that havesprayed themselves with urine. Salivacould be a vector, too; in both deer andelk, individuals meet and greet by lickingeach other’s mouths and noses, thus ex-changing drool. Ranched elk may swapsaliva when they feed in close quarters. Itis also possible that animals take in thepathogen while grazing in areas wheresick animals have shed prions on theground in their feces, urine and saliva.

By 1985 veterinarians discoveredCWD in free-ranging deer and elk, gener-ally within about 30 miles of the twowildlife facilities. Whether the diseaseoriginated in the wild and spread to thecaptives, or vice versa, is not known. Thetwo populations had plenty of time tomingle. Especially during mating season,wild cervids nosed up to captives throughthe chain-link fences. Incubating deercould also have escaped or been released.

Both facilities tried hard to eradicateCWD. The Sybille center killed all thedeer and elk in the affected area and wait-ed a year to introduce new animals; fouryears later deer and elk started comingdown with CWD. The Fort Collins facil-ity acted more aggressively. Officials first


■ Chronic wasting disease (CWD) is a fatal condition spreading among wild deer in some parts of North America. It kills in part by making holes in the brain.

■ Malformed proteins called prions trigger the disease. The extreme durability of prions and CWD’s long incubation times make controlling the spread of the sickness difficult.

■ Studies are under way to see whether CWD can infect humans and livestock.

Overview/Chronic Wasting Disease

Geography provided natural barriers, but humans provided a way past them: along the roads, IN A TRUCK.


killed off all the resident deer and elk;then they turned several inches of soil andrepeatedly sprayed structures and pas-tures with swimming-pool chlorine, whichreadily wipes out bacteria and viruses. Af-ter waiting a year, they brought in 12 elkcalves, but a few years afterward two ofthose elk contracted CWD.

The disease’s persistence has perma-nently contaminated an area of about15,000 square miles in northeastern Col-orado, southeastern Wyoming and (be-ginning in 2001) southwestern Nebraska.The incidence of CWD among the cervidsin this so-called endemic area averagesabout 4 to 5 percent but has reached 18percent in some places. To help keep thedisease confined here, the research facili-ties stopped trading captive animals witheach other. In fact, no captive cervids nowleave the endemic area alive: “They’reonly allowed out to come to my necropsyroom,” wryly remarks Williams, now atthe University of Wyoming. More impor-tant were the mountains and other natur-al barriers, which scientists expected

would keep CWD from spreading rapid-ly out of the endemic area. There was,however, an easy way past those naturalbarriers: along the roads, in a truck.

Out and AboutSOME 11,000 GAME FARMS andranches holding hundreds of thousands ofdeer and elk dot the U.S. and Canada. Be-sides harvesting the meat, ranchers cansell the antlers—those from elk are mar-keted as a supplement in vitamin stores(“velvet antler”) and as an aphrodisiac inAsia (“velvet Viagra”). To start suchfarms, ranchers must buy breeding cervids.Somewhere along the line, businesses musthave picked up incubating animals fromthe endemic area. And the interstate tradeof cervids continued the spread, westacross the Continental Divide and eastacross the Mississippi River. (These daysmost states regulate such trade.)

The first farmed cervid to display signsof CWD was an elk that fell ill in 1996 ona ranch in Saskatchewan. By 2001 some20 ranches reported cases across six states

(Colorado, Kansas, Montana, Nebraska,Oklahoma and South Dakota) and oneother Canadian province (Alberta). Quick,aggressive measures—namely, killing offthe herds—appear to have eliminated theproblems on the ranches.

Nevertheless, the transport of incu-bating cervids may have carried CWD towild populations in those states and be-yond—such as to white-tailed deer in Wis-consin’s eradication zone. But preciselywhen and how mule deer gave it to white-tailed deer, the most common type in theeastern U.S., is unknown and may neverbe clear. “By the time these problems arediscovered,” Miller says, “they have prob-ably been sitting there for decades, whichmakes it difficult to go back and retracehow things came about.” Based on epi-demiological models and on Wisconsin’sroughly 1.6 percent incidence in the erad-ication zone, Miller thinks CWD hadprobably been lurking there since the ear-ly 1990s.

Wisconsin’s approach makes sense toscientists studying prion diseases. “The


CO

UR

TESY

OF

WYO

MIN

G G

AME

AN

D F

ISH

DE

PAR

TME

NT

(elk

); D

AVID

NE

VALA

(sa

mp

lin

g s

tati

on);

E

LIZA

BE

TH W

ILLI

AMS

Un

iver

sity

of

Wy

omin

g (

mic

rog

rap

h);

CO

RN

ELI

A B

LIK

(m

ap

)

CHRONIC WASTING DISEASE roughens thecoat of an infected elk, which also showsearly signs of emaciation (top left). Theillness, which has appeared in 14 statesand provinces (map), fills brain tissuewith holes (white areas in micrograph). In a temporary lab near Black Earth, Wis.(above), a researcher extracts tissuesamples from the head of a white-taileddeer killed in the state’s eradication zone.

Sybille Research Unit

Endemic area

Wisconsin eradication zoneFoothills Wildlife Research Facility(Fort Collins)

States/provinces whereCWD has been found

Wild outbreaks


idea is to find a fairly small focus and getrid of all the animals in the area,” in thehopes of preventing CWD from attaininga permanent hold in the region, Williamssays. A rapid spread is possible in Wis-consin because the deer population in thestate’s southwestern corner is dense:Thomas Givnish, an expert on the ecolo-gy of diseases at the University of Wis-consin–Madison, notes that it runs about50 to 100 deer per square mile, or 10times that of the endemic area aroundFort Collins. “The alternative is to do noth-ing,” Williams observes, and then “you

know it’s going to be established.” By theend of March, Wisconsin hunters hadbagged 9,287 deer—which will cut the fallpopulation by 25 percent but will noteliminate CWD, notes state wildlife biol-ogist Tom Howard. A few more seasonsof liberal hunting may be needed.

Considering the persistence of prions,Wisconsin may have to live with CWD, asColorado does. “The disease has been herea long time,” Miller comments of CWDaround northeastern Colorado. “We can’tget rid of it here. We try to get infectionrates down so that it can’t spread.” Millersays that Colorado had hoped to purgeCWD through culling. But “we discov-ered we were 10 to 20 years too late. Itwas already out there; we didn’t realize

it.” That statement may apply to otherstates that have found CWD among wilddeer, including Illinois and New Mexico.

Venison and BeyondNO ONE KNOWS whether CWD canpass to humans. A test tube study mixedCWD prions with normal prion proteinsfrom cervids, humans, sheep and cows.The CWD prions had a hard time con-verting normal human PrP—less than 7 percent of the protein was changed.The downside is that CWD prions con-verted human PrP about as efficiently asBSE prions do. And because BSE has infected humans, CWD might pose asimilar risk. But because beef is far morepopular than venison, CWD doesn’t pre- D

AVID

NE

VALA


DEER CARCASSES from March 2003 hunts in Wisconsin’s eradication zone were loaded intoa refrigerated storage unit. The heads were sentfor CWD testing; the bodies were incinerated.


sent quite the same public health threat.To see if CWD has already infected

people, the Centers for Disease Controland Prevention investigated the deaths ofthe three young venison eaters who suc-cumbed to sporadic Creutzfeldt-Jakobdisease. All were younger than 30 years,which is exceedingly rare in CJD. In fact,

through May 31, 2000, just one other U.S.case of sporadic CJD occurred in this agegroup since surveillance began in 1979.

The first was a 28-year-old cashier,who died in 1997; she had eaten deer andelk as a child, from her father’s hunts inMaine. The second was a 30-year-oldsalesman from Utah who had been hunt-ing regularly since 1985 and who died in1999. The third was a 27-year-old truckdriver from Oklahoma who died in 2000;he had harvested deer at least once a year.Tests of the 1,037 deer and elk taken dur-ing the 1999 hunting season from the re-gions where the victims’ meat originatedall turned up CWD negative (none of themeat came from the endemic area). Thevictims’ brains showed no unique damageor distinct biochemical signs, as is the casewith other prion diseases in humans.

Six other patients (all at least middle-aged) raised suspicions about the CWDrisk to humans. Three were outdoorsmenfrom the Midwest who had participatedin wild deer and elk feasts and died in the1990s. The other cases were reported inApril and include two from WashingtonState who hunted together. Researchers,however, could not find any connectionwith CWD. And states with CWD havenot discovered a higher incidence ofCreutzfeldt-Jakob disease.

These observations may seem reas-suring, but it is too early to conclude thatCWD does not pose a human health haz-ard. The incubation period of prion dis-eases may span upward of 40 years, andCWD has been spreading noticeably inonly the past 10. The rarity of prion dis-eases and the low national consumptionof deer and elk (compared with beef) make

it hard to draw any firm conclusions. Be-cause of the uncertainties with CWD andthe fact that animal prion diseases havejumped to humans, the CDC warns againsteating food derived from any animal withevidence of a spongiform encephalopathy.

Scientists are still trying to determineif CWD poses a threat to livestock. In an

ongoing experiment begun in 1997,Hamir and his colleagues injected brainsuspensions from CWD mule deer intothe brains of 13 Angus beef calves. Twobecame ill about two years after inocula-tion, three others nearly five years after.Hamir began repeating the experiment inNovember 2002, this time with the brainsof CWD white-tailed deer.

Under more natural conditions, bo-vines have not contracted CWD. Williamshas kept cows with infected cervids, andmore than five years on, the cows are stillhealthy. Bovines kept with decomposingCWD carcasses or isolated in pens thatonce housed CWD animals have also re-mained free of prion disease. (These re-ports are good news for pasture-grazingcows, which might find themselves in thecompany of wild deer.) To see whetherCWD might pose a danger when eaten,Williams has begun feeding CWD brainmatter to calves. The long incubation ofthese illnesses, however—BSE incubatesfor up to eight years—means these exper-iments must continue for several years.

If U.S. livestock so far seem to be safefrom CWD, the same cannot be said ofother animals. If an infected deer dies inthe forest and nobody is there to see it,plenty of coyotes, bobcats and other car-nivores will, and they will gladly scavengewhat remains of the wasted carcass.

Moreover, during the clinical phase, CWDanimals undoubtedly make easier prey.The canine family is evidently immune toprion diseases, but felines can contractthem. Transmission studies with moun-tain lions have begun, and local lions thatdie for unknown reasons end up on thepathology table, Williams says.

Although many states have uncoveredCWD, other states “are looking darnhard,” Miller says, but have not found it—among them, Arizona, Kansas, Michigan,Montana, Nevada and New Jersey. Ap-parently, only pockets of outbreaks exist.Wildlife managers therefore have a fight-ing chance to keep CWD from gaining apermanent grip throughout the country,so long as control efforts begin promptly.Unfortunately, not all states with CWDare as aggressive as Wisconsin when itcomes to surveillance and eradication.

To stop or at least slow the spread ofthe fatal sickness, extensive culling ap-pears to be the best strategy. One couldhope that CWD occurs naturally in deerand that the epidemics will run theircourse and leave behind CWD-resistantcervids. Some lines of sheep, for instance,are immune to scrapie. But so far allwhite-tailed and mule deer appear to beuniformly susceptible. “I don’t think ge-netics is going to save us on this,” remarksthe NIH’s Race. Sadly, the only way to savethe deer, it seems, is to shoot them.

Philip Yam is Scientific American’snews editor. This article is adaptedfrom his book, The PathologicalProtein: Mad Cow, Chronic Wasting,and Other Deadly Prion Diseases,published in June.


Risk Analysis of Prion Diseases in Animals. Edited by Corinne I. Lasmézas and David B. Adams.Scientific and Technical Review, Vol. 22, No. 1; April 2003.

The Pathological Protein: Mad Cow, Chronic Wasting, and Other Deadly Prion Diseases. Philip Yam. Copernicus Books, 2003. (www.thepathologicalprotein.com)

Chronic Wasting Disease Alliance Web site is at www.cwd-info.org/

U.S. Department of Agriculture Web site on CWD is at www.aphis.usda.gov/lpa/issues/cwd/cwd.html

M O R E T O E X P L O R E

It is too early to conclude that chronic wasting disease does not pose a HUMAN HEALTH HAZARD.


Captain John Carter, the hero of the adventure

novels of Edgar Rice Burroughs, was a gentle-

man of Virginia and an officer of the Confed-

eracy. Impoverished after the Civil War, he went looking

for gold in Arizona and, while being chased by Apache war-

riors, fell and struck his

head. He returned to

consciousness on an arid

planet with twin moons,

populated by six-legged creatures and beautiful princesses

who knew the place as “Barsoom.” The landscape bore an

uncanny resemblance to southern Arizona. It was not en-

tirely dissimilar to Earth, only older and decayed. “Theirs

is a hard and pitiless struggle for existence upon a dying

planet,” Burroughs wrote in the first novel.

IT WOULD TAKE YOU ABOUT FIVE MINUTES to hike across the area shown in this image, on the north side of Newton Crater in the southernhemisphere of Mars. You would leave your footprints on lightly frosted soil (bright areas), clamber over windblown features such as sanddunes and jump across possibly water-carved features such as gullies. These landforms probably continue to form even today. Like otherMars Global Surveyor images, this one is a composite of high-resolution grayscale and low-resolution color; the colors are onlyapproximate. It is a far cry from the vague (and often fanciful) view of Mars a century ago (above).


10 METERS

MARSTheUnearthly

Landscapesof

TheUnearthlyLandscapesof

The Red Planet is no dead planet By Arden L. Albee


In science as well as science fiction,Mars is usually depicted as a version ofEarth in its extreme—smaller, colder, dri-er, but sculpted by basically the same pro-cesses. Even well into the 20th century,many thought the planet had flowing wa-ter and proliferating plants. The resem-blance to Earth fell apart when spacecraftin the late 1960s revealed a barren, cra-tered world, more like the moon. But itquickly returned with the subsequent dis-coveries of giant mountains, deep can-yons and complex weather patterns. TheViking and Mars Pathfinder images fromthe surface look eerily Earth-like. LikeBurroughs, researchers compare the equa-torial regions of Mars to the AmericanSouthwest. For the polar regions, themodel is the Dry Valleys of Antarctica, afrozen desert in a landscape of endless ice.

But if there is one thing researchershave learned from recent Mars explo-ration, it is to be careful about drawingsuch comparisons. In the past five years,spacecraft have collected more informa-tion about the Red Planet than all previ-ous missions combined. Mars has provedto be a very different and more compli-cated planet than scientists thought be-forehand. Even the single biggest ques-tion—Was Mars once warm and wet,possibly hospitable to the evolution oflife?—is more nuanced than people havetended to assume. To make sense ofMars, investigators cannot be blinded bytheir experience of Earth. The Red Plan-et is a unique place.

Mars as the Abode of Dust MARS EXPLORATION has certainlyhad its up and downs. In the past decadeNASA has lost three spacecraft at Mars:

Mars Observer, Mars Climate Orbiter(intended as a partial replacement forMars Observer) and Mars Polar Lander.Lately, though, the program has had arun of successes. Mars Global Surveyorhas been taking pictures and collecting in-frared spectra and other data continu-ously since 1997. It is now the matriarchof a veritable family of Mars spacecraft.Another, Mars Odyssey, has been orbit-ing the planet for more than a year, map-ping the water content of the subsurfaceand making infrared images of the sur-face. As this article goes to press, NASA

plans to launch the Mars ExplorationRovers, successors to the famous So-journer rover of Mars Pathfinder [see boxon page 50]. Around the same time, the European Space Agency expects to sendoff the Mars Express orbiter, with itsBeagle 2 lander. The Nozomi orbiter,sent by the Institute of Space and Astro-nautical Science in Japan, should arriveat Mars in December.

Never before have scientists had sucha comprehensive record of the processesthat operate on the surface and in the at-mosphere [see box on page 52]. Theyhave also studied the craters, canyons andvolcanoes that are dramatic relics of thedistant past. But there is a huge gap in ourknowledge. Between ancient Mars andmodern Mars are billions of missingyears. No one is sure of the conditionsand the processes that sculpted Mars dur-ing most of its history. Even less is knownabout the subsurface geology, which willhave to be the subject of a future article.

Present-day Mars differs from Earthin a number of broad respects. First, it isenveloped in dust. Much of Earth’s sur-face consists of soil derived by chemical

weathering of the underlying bedrockand, in some regions, glacial debris. Butmuch of Mars’s surface consists of dust—very fine grained material that has settledout of the atmosphere. It drapes over allbut the steepest features, smothering theancient landscape. It is thick even on thehighest volcanoes. The dustiest areas cor-respond to the bright areas of Mars longknown to telescope observers.

Dust produces otherworldly land-scapes, such as distinctively pitted terrain.As dust settles through the atmosphere, ittraps volatile material, forming a mantleof icy dust. Later on, the volatile ices turnto gas, leaving pits. Intriguingly, the thick-ness of the icy, dusty mantle on Marsvaries with latitude; near the poles, MarsOdyssey has shown, as much as 50 per-cent of the upper meter of soil may be ice.On slopes, the icy mantle shows signs ofhaving flowed like a viscous fluid, muchin the manner of a terrestrial glacier. Thismantle is becoming the focus of intensescientific scrutiny.

Second, Mars is extremely windy. It isdominated by aeolian activity in much theway that Earth is dominated by the actionof liquid water. Spacecraft have seenglobe-encircling dust storms, huge dustdevils and dust avalanches—all wroughtby the wind. Dust streaks behind obsta-cles change with the seasons, presumablybecause of varying wind conditions.

Where not dust-covered, the surfacecommonly shows aeolian erosion or de-position. Evidence for erosion shows upin craters, from which material appears tohave been removed by wind, and in yar-dangs, bedrock features that clearly havebeen carved by windblown sand. Evi-dence for deposition includes sand sheetsand moving sand dunes. The latter arecomposed of sand-size grains, which movearound by saltation—multiple bounces of


CO

UR

TESY

OF

ww

w.n

irga

l.n

et/s

chem

as/

sch

iap

are

lli –

ma

p.h

tml

(pa

ge

44

); N

ASA/

JPL/

MAL

IN S

PAC

E S

CIE

NC

E S

YSTE

MS

MO

C 2

–3

20

(p

ag

e 4

5);

NAS

A/JP

L/M

ALIN

SP

ACE

SC

IEN

CE

SYS

TEM

S M

OC

2–

25

9 (

opp

osit

e p

ag

e)

■ Two ongoing missions to Mars, Mars Global Surveyor and Mars Odyssey, areraising difficult questions about the Red Planet. Flowing water, ice and wind haveall helped carve the landscape over the past several billion years. The processesare both similar and dissimilar to those acting on Earth’s surface. Scientists’experience of Earth have sometimes led them astray.

■ The question of whether Mars was once hospitable is more confusing than ever.Spacecraft have gathered evidence both for and against the possibility. Threeupcoming landers, two American and one European, could prove crucial toresolving the matter.

Overview/The Martian Surface

LAYERED TERRAIN looks surreal, almost like atopographic map, but is quite real. It covers thefloor of western Candor Chasma, a ravine that ispart of the Valles Marineris canyon system.Scientists have identified 100 distinct layers,each about 10 meters thick. They could besedimentary rock originally laid down by water,presumably before the canyon cut through theterrain. Alternatively, the layers could be dustdeposited by cyclic atmospheric processes. Thisimage was taken by Mars Global Surveyor.


100 METERS


MAR

S O

RB

ITE

R L

ASE

R A

LTIM

ETE

R S

CIE

NC

E T

EAM

, M

ALIN

SP

ACE

SC

IEN

CE

SYS

TEM

S (t

rue

colo

r m

osa

ic);

OD

ED

AH

ARO

NSO

N C

ali

forn

ia I

nst

itu

te o

f Te

chn

olog

y (

all

oth

er m

osa

ics)

; SO

UR

CE

: M

ARS

OD

YSSE

Y M

ISSI

ON

(w

ate

r);

PH

ILIP

CH

RIS

TEN

SEN

Ari

zon

a S

tate

Un

iver

sity

(g

eolo

gy

); M

ARS

GLO

BAL

SU

RVE

YOR

MIS

SIO

N (

all

oth

er m

osa

ics)

0°90°

0°

30°

60°

–90°

–60°

–30°

180° 180°90°270°

TRUE COLOR Mars is four worlds in one: the heavily crateredsouthern hemisphere (with riverlike valley networks), the smoothernorthern hemisphere (with hints of an ancient shoreline), the equatorialregion (with giant volcanoes and canyons), and the polar caps (withbizarre, protean terrain). This map combines wide-angle cameraimages with altimetry, which brings out details. The color is realistic.

TOPOGRAPHY The elevation spans 30 kilometers from the lowestbasins (dark blue) to the highest volcanoes (white). For comparison,the range of elevation on Earth is only 20 kilometers. The large bluecircle in the southern hemisphere represents the Hellas impact basin,one of the biggest craters in the solar system. Girdling Hellas is a vastring of highlands about two kilometers in elevation.

CRUSTAL THICKNESS Combining the topographic map withmeasurements of Mars’s gravity, researchers have deduced thethickness of the Martian crust: roughly 40 kilometers under thenorthern plains and 70 kilometers under the far southern highlands.The crust is especially thick (red) under the giant Tharsis volcanoesand thin ( purple) under the Hellas impact basin.

WATER Neutrons reveal the presence of water in the top meter ofsoil. The energy of these particles, which are produced when cosmicradiation bombards the soil, is sapped by the hydrogen within watermolecules. A dearth of medium-energy (“epithermal”) neutrons meanswater-rich soil (blue). The implied amount of water, most of it in the farsouth, would fill two Lake Michigans. More may lie deeper underground.

OLYMPUS MONS

THARSIS

NEWTONCRATER

VALLESMARINERIS

PATHFINDER SITE

WHITE ROCK

HELLAS BASIN

Landing Sites: Gusev Crater (1), Meridiani Planum (2), Isidis Planitia (3)

CANDORCHASMA

ARGYRE BASIN

GLOBAL VIEWS of MARS

MAGNETISM Mars lacks a global magnetic field, yet areas of itscrust are magnetized up to 10 times as strongly as Earth’s crust. In these areas, iron-rich rocks have become bar magnets, suggestingthat Mars had a global field at the time the rocks solidified from a molten state. The east-west banding resembles patterns producedby plate tectonics on Earth, but its origin is unknown.

GEOLOGY Infrared spectral measurements reveal rock types.Basalt (green), a primitive volcanic rock, dominates the southernhemisphere. Andesite (blue), a more complex volcanic rock, seems tobe common in the north. Near the equator is an outcrop of hematite(red), a mineral typically produced in the presence of water. In largeregions, dust (tan) or clouds (white) hide the underlying rock types.

–9 –5 0 5 10 15Elevation (kilometers)

0 2 4 6 8 10Epithermal Neutron Flux (counts per second)

0 20 40 60 80 100Thickness (kilometers)

–500 –100 –10 25 100 1,000Vertical Magnetic Field (nanoteslas)

2

1

3

HEMATITE


windblown grains along the ground. Ittakes a stronger wind to loft dust directlythan to initiate saltation, so this phenom-enon accounts for most of the dust kickedinto the atmosphere.

Aeolian activity seems to have per-sisted since the time of heavy cratering,back when the solar system was stillyoung. Many images show craters withvarying degrees of erosion: some are shal-low and partially filled with deposits andsand dunes, whereas others are pristine—

deeper and bowl-shaped. Michael Malinand Kenneth Edgett of Malin Space Sci-ence Systems in San Diego, the researchfirm that operates the Mars Global Sur-veyor camera, have inferred a sequence ofprocesses: Sand was blown through theregion, and some of it got trapped incraters; other craters formed later. Whereand how such a volume of sand was pro-duced and how it was blown around re-main a mystery, however.

The Angry Skies of MarsA THIRD WAY in which Mars differsfrom Earth is in its amazing variety ofweather and climate cycles, many ofwhich are similar to those on Earth, manylike nothing on Earth. The Martian day isalmost the same as an Earth day, but theMartian year is 687 Earth days. The tiltof Mars’s rotation axis, which producesseasons, is very close to that of Earth’s.Mars lacks the precipitation and oceansthat are so crucial to weather on Earth.But the atmospheric pressure (less than 1percent of that on Earth) varies seasonal-ly by about 25 percent, driven by the con-densation and sublimation of carbondioxide frost at the poles. The thin atmo-sphere has a very low heat capacity, so thesurface temperature swings by more than100 degrees Celsius from day to night.The thermal properties of the thin atmo-sphere are dramatically affected by dustand ice particles suspended in the air. Theupshot is that, despite being so thin, theatmosphere has complex circulation pat-terns and dynamics. A daily weather re-port might talk of strong winds, high-lev-el ice clouds, low-level fogs, seasonalfrost, dust devils and massive dust storms.

As on Earth, storm systems often spi-ral southward from the northern polar

regions. But the largest dust storms typi-cally start during the southern spring asthe planet rapidly heats up. Periodicallythey coalesce and come to encircle the en-tire planet. Mars Global Surveyor close-ly followed the evolution of a four-monthglobal dust storm that started in June2001. Contrary to scientists’ expectations,it was not, in fact, one single global storm,but the confluence of several regionalstorms. Malin has compared the climaticeffect of the dust raised by this storm withthe aftermath of Mount Pinatubo’s erup-tion on Earth in 1991—namely, a briefbut widespread cooling.

The polar ice caps play a key role inthe atmospheric cycles. Their size andshape, as shown by topographical mea-surements, indicate that the caps are pre-dominantly water ice, as opposed to so-called dry ice, made of carbon dioxide:dry ice is not as rigid as water ice, and itcould not support the observed domelikeshape. A major new discovery has beenthat the layer of dry ice that covers muchof the south polar cap is being erodedaway at a high rate. Clearly, the erosioncannot go on forever. Nor can the currentdust sinks and sources remain in their cur-rent states indefinitely. To replenish the iceand dust, other cycles must be occurring,perhaps tied to orbital variations. Malinand Edgett have suggested that wind con-ditions may be less intense now than in thefairly recent past, another hint that theMartian climate changes with time.

A fourth major difference betweenEarth and Mars is the behavior of liquidwater. Liquid water is unstable at the sur-face under present pressure and temper-ature conditions. It does not rain. Still,water ice can—and does—persist at somedepth within the Martian soil during allor much of the year. On Mars, as onEarth, several types of patterned groundmark the presence of ice-rich soil. MarsOdyssey has detected ground ice over

most of the planet outside the equatorialregions, and models predict that the iceextends to considerable depths.

Liquid water can sometimes leak ontothe surface. In 2000 Malin and Edgett de-scribed fresh gullies that look like water-carved features on Earth [see “Gully GeeWhiz,” by George Musser; ScientificAmerican, September 2000]. In the en-suing excitement, researchers advancedmany theories to explain them: leakingaquifers (which would be inexplicablyperched high on crater rims); pressurizedgeysers of water; high-pressure outburstsof carbon dioxide gas; volcanic heatsources at depth. Finally, earlier this yearPhilip Christensen of Arizona State Uni-versity discovered gullies that clearlyemerge from underneath a bank of snowand ice. He concluded that they are relat-ed to Martian climate cycles. In colder pe-riods, slopes become blanketed with amixture of snow and dust. Sunlight pen-etrates this insulating blanket, heating itenough for water to melt under the snowand to run down the slope, creating gul-lies. In warmer periods, the blanket meltsor evaporates entirely, exposing the gullies.

Layer upon LayerDESPITE THE ABUNDANCE of water,however, Mars is arid. It has the miner-alogy of a nearly waterless surface. OnEarth, the action of warm liquid waterproduces weathered, quartz-rich soils, hy-drated clays, and salts such as calciumcarbonate and sulfate. Beach sand andsand dunes are largely quartz. On Mars,spacecraft have yet to find any deposits ofthese minerals. The darker Martian dunesare basaltic, consisting mainly of mineralssuch as pyroxene and plagioclase, whichon Earth would readily weather away. Itfollows that the present cold and dry at-mospheric conditions have persisted sincea time far back into the planet’s history.

Has Mars always differed so much


ARDEN L. ALBEE is the project scientist—that is, overall leader of the science team—for theMars Global Surveyor mission. He had the same role for the ill-fated Mars Observer mission,for which Surveyor is a partial replacement. Albee is emeritus professor of geology and plan-etary science at the California Institute of Technology and served as chief scientist of NASA’sJet Propulsion Laboratory from 1978 until 1984. His research interests run from field ge-ology to compositional analysis of rocks, meteorites, comets and lunar samples. He stillmakes time for his eight children and 11 grandchildren.

THE

AU

THO

R


from Earth? Below the mantles of dustand sand are numerous signs that the RedPlanet has transformed over time. To be-gin with, the planet has a striking di-chotomy in landscape between its north-ern and southern hemispheres. The south-ern hemisphere is high in altitude andheavily cratered (indicating an ancientsurface). The northern one is a vast, low-lying plain with fewer craters (indicatinga younger age). In between is the immenseTharsis Plateau, intermediate in age andcapped by volcanoes that dwarf any onEarth. Using the new high-resolution dataon these volcanoes, James W. Head III ofBrown University has found flow patternsthat look strikingly like mountain glaci-ers—and that may suggest the presence ofice under a blanket of rock and dust.

The northern lowlands are exceed-ingly level, leading to speculation that

they were lake beds during a significantchunk of Martian history. They appear tobe covered with multiple layers of vol-canic flows and sedimentary debris thatoriginated in the south. Detailed newtopographic maps have unveiled “stealthcraters”—faint circular expressions, evi-dently part of an ancient cratered surfacethat has been buried by a thin layer ofyounger deposits.

Along the edge of the southern high-lands are features that could only havebeen carved by liquid water. These featuresare tremendously larger than their coun-terparts on Earth. The famous canyonValles Marineris would run from Los An-geles to New York with a width extend-ing from New York to Boston and adepth similar to the elevation of MountMcKinley. No terrestrial canyon comesclose. At its head is a jumbled terrain, in-

timating that water flowed not in a steadytrickle but in concentrated, catastrophicoutflows, scouring the surface along itspath. Other Martian outflow channelshave similar features. Because these fea-tures are carved into the Tharsis Plateau,they must have an intermediate age.

Streamlined islands and other featuresin these channels look much like the scab-lands in the northwestern U.S., whichwere gouged by the Spokane Flood to-ward the end of the last ice age, about10,000 years ago. During the massive del-uge, a lake roughly the size of one of theGreat Lakes burst its ice dam and rushedout within just a few days. On Mars, suchcalamities were 10 to 100 times as devas-tating. They may have been triggered byvolcanic heat sources or by the generalheat flow from the interior of the planet.Heat would have melted ice underneaththe thick permafrost layer, building uptremendous pressures until the water fi-nally burst out.

The most contentious water-relatedfeatures of all are the valley networks. Lo-cated throughout the southern highlands,they have a branching, dendritic patternreminiscent of rivers on Earth—suggestingthat they were formed by surface runofffrom rainfall or snowfall. They are thestrongest hint that Mars was once aswarm as Earth. But these networks lookrather different from rain-fed rivers onEarth. They more closely resemble rivernetworks in desert areas, which are fed bywater that slowly seeps from subterraneansources. Such streams typically originatein steep-walled amphitheaters rather thanin ever smaller tributaries. Heated debateshave been taking place at scientific meet-ings over the crucial question: Did it rainon early Mars?

The timing of the water networkscould be the key to making sense of them.Recent detailed studies of the northernedge of the highlands show that immenseamounts of material eroded during—

rather than after—the intense meteorbombardment that took place early inMartian history. These analyses implythat the distribution of water kept chang-ing as impacts reworked the landscape.Craters filled with water and debris, andchannels began to link them together into N

ASA

ARTI

ST C

ON

CE

PT/

MAR

S E

XPLO

RAT

ION

PR

OG

RAM

THIS MONTH NASA PLANS to launch twin rovers to the Red Planet, and the EuropeanSpace Agency will send off a lander, too. The three of them, scheduled to arrive nextJanuary, will be robotic geologists—studying the geologic history of landing sites,investigating what role water played there and determining how suitable pastconditions would have been for life.

The rovers deserve particular mention, because they will give scientistsunprecedented mobility. Each rover can travel 100 meters a day. For comparison, theSojourner rover on the 1997 Mars Pathfinder lander traversed 100 meters over itsentire mission [see “The Mars Pathfinder Mission,” by Matthew Golombek; SCIENTIFIC

AMERICAN, July 1998]. A mast, about 1.5 meters high, supports a binocular camera anda thermal-emission spectrometer, one of many instruments that can analyze thecomposition of materials. A robot arm holds the other instruments: a Mossbauerspectrometer, an alpha-particle detector, an x-ray spectrometer and a microscope.The arm also carries a scraper to clean off rock surfaces for study. A dish antennabeams signals directly to Earth, and a black rod antenna relays data through theMars Global Surveyor or Mars Odyssey orbiter.

Deciding on their landing sites has been one of the most exciting uses of datafrom the orbiters. Unlike the selection of sites for the Viking landers in 1976—whichwas a matter of a few people’s gut instinct—the choice for the rovers involved a longdeliberation among dozens of scientists and engineers. Weighing tantalizing geology

(such as suspected water-related features) againstpotential dangers (such as steep slopes and highwinds), they winnowed an initial list of more than 150possible sites down to seven, then four and finally, onApril 11, two: Gusev Crater, whose layered depositsmight be lake-bed sediments, and Meridiani Planum,which is rich in coarse-grained hematite, a mineraltypically formed in association with liquid water. (TheEuropean Beagle 2 lander will touch down on IsidisPlanitia, a possible sedimentary basin.) —A.L.A.

Going for a Drive


NASA’s Mars Exploration Rover


IN THE PLAINS northwest of Olympus Mons, dustdevils are sweeping over the land and leavingstreaks in their wake (right). A similar scene hasunfolded in the Argyre Basin (below) and east ofValles Marineris (bottom), where a devil wascaught in the act. These tornadolike vortices—thought to be created as warm air rises off thesurface—clear away light-colored dust andexpose comparatively dark soil. They are one ofthe many wind-related processes that arecontinuously reshaping the Martian surface. Thepicture at the right is from Mars Odyssey; thetwo below are from Mars Global Surveyor.

NAS

A/AR

IZO

NA

STAT

E U

NIV

ER

SITY

TH

EM

IS S

CIE

NC

E T

EAM

, FE

BR

UAR

Y 6

, 2

00

3 (

rig

ht)

; N

ASA/

JPL/

MAL

IN S

PAC

E S

CIE

NC

E S

YSTE

MS

MO

C 2

–2

20

(le

ft t

op)

AND

MO

C 2

–3

18

(le

ft b

otto

m)

500 METERS

250 METERS1 KILOMETER


a network, but impacts continually dis-rupted this process. For instance, the Ar-gyre Basin, 1,000 kilometers in diameter,may once have been filled to its brim withwater. It is part of a valley system thatbrought water from near the South Pole,through the basin, into channels thatcrossed the equator. The roles of waterand ice in these systems, both above-ground and underground, remain unclear.In any case, these networks are very dif-ferent from hydrologic systems on Earth.

A final clue to Martian history comesfrom one of the biggest surprises deliveredby Mars Global Surveyor, the extent towhich the uppermost crust consists of lay-ered deposits. Almost everywhere that thesubsurface is exposed—on walls ofcanyons, craters, mesas and valleys—it islayered. The layers differ from one an-other in thickness, color and strength.They show that the Martian surface hasundergone complex sequences of deposi-tion, crater formation and erosion. Theoldest layers are the most extensive. The

higher layers have been partially strippedaway, apparently blown by the wind.

Where did the layers come from? Thelack of boulderlike blocks argues againsttheir being volcanic flows, although theycould be volcanic ash. Ultimately, how-ever, most of the layers probably origi-nated in impact debris. On the moon, sci-entists observe overlapping rings of im-pact debris, which mark craters of differingages. Similarly, Mars is so heavily crateredthat the upper crust has been stirred uplike soil tilled by a gardener. Water andwind then scattered this material.

Blue Mars?IN A SENSE, scientists’ ideas about earlyMars are more uncertain than they haveever been. This doubt comes to the forewhen researchers address the question ofliquid water. The presence or absence ofliquid water is fundamental to geologicprocesses, climate change and the originof life. The early valley networks and thelater flood channels attest to an abun-

dance of water. The evidence for earlyrainfall suggests that the atmosphere wasonce much denser. But spacecraft havefound no evidence for deposits of car-bonate minerals, which would be the ves-tiges expected from an early dense carbondioxide atmosphere [see “The Climate ofMars,” by Robert M. Haberle; Scientif-ic American, May 1986].

At this point, scientists have threemain hypotheses. Perhaps the early at-mosphere was indeed thick. The planetmight have had lakes, even oceans, free ofice. Robert A. Craddock of the NationalAir and Space Museum and Alan D.Howard of the University of Virginia re-cently suggested that the carbon dioxidewas lost to space or locked up in carbon-ate minerals that have so far escaped de-tection. Intriguingly, Mars Odyssey spec-tra have revealed small amounts of car-bonate in the dust.

Alternatively, perhaps Mars had a fair-ly thin atmosphere. It was a wintry world.Any standing bodies of water were covered


NAS

A AR

TIST

CO

NC

EP

T/G

LOB

AL S

UR

VEYO

R P

RO

GR

AM

BOTH MARS GLOBAL SURVEYOR and Mars Odyssey circle theplanet in paths that take them over both poles. Their orbitsremain fixed as Mars spins below them, allowing theinstruments to observe day and night swaths over the entireplanet. This continuous coverage can track changes in thesurface, atmosphere, gravity and magnetic field.

Global Surveyor has five main instruments. Its laseraltimeter has measured the overall shape and the topography ofMars with an altitude precision of aboutfive meters, which means that Mars isnow better mapped than most of Earth.The camera takes red and blue medium-resolution images of the entire surface,as well as high-resolution images—1.4meters per pixel, as good as the picturestaken by the spy satellites of the1960s—of limited areas. A Michelsoninterferometer measures the emittedthermal infrared spectrum with highspectral resolution but low spatialresolution, suitable for mapping themineral composition and thermalproperties of the surface. Amagnetometer determines the magneticfield. Finally, the spacecraft itself counts

as an instrument, because its motion is sensitive to variations inMartian gravity. The gravitational field reveals the thickness ofthe crust and changes in the size of the polar ice caps.

Odyssey complements Global Surveyor. Its camera lacks ahigh-resolution mode but takes images in five selected colorbands. Its infrared imager has low spectral resolution but highspatial resolution. Another instrument measures gamma-rayand neutron fluxes, which are sensitive to hydrogen just below

the planet’s surface; Odyssey istherefore the first spacecraft capable ofpeeking under the surface of Mars, to adepth of about one meter.

These spacecraft also monitor theatmosphere. Cameras scan the entireplanet daily, much like Earth-orbitingweather satellites. Twelve times a day,the thermal-emission spectrometertakes readings of temperature,pressure, cloud cover and dustabundance. Additionally, as radiotransmissions pass through the Marsatmosphere, they are diffracted; signalprocessing can infer the variation oftemperature and pressure with altitude.

—A.L.A.MARS GLOBAL SURVEYOR (artist’s conception)

Monitoring Mars, 24/7/687


in ice. Snow might have fallen, rechargingthe groundwater and leading to tempo-rary trickles of water across the surface.Steven M. Clifford of the Lunar and Plan-etary Science Institute in Houston, amongothers, has conjectured that melting un-der a glacier or a thick layer of permafrostcould also have recharged subterraneanwater sources. Although Mars was bit-terly cold, periodic bursts of relativelywarmer temperatures could have reinvig-orated the planet. Orbital shifts, similar tothose that trigger ice ages on Earth, drovethese climate cycles. Head, John F. Mus-tard of Brown and others have pointed tothe latitude dependence of the ice and dustcover as evidence for climate change.

Finally, perhaps the climate cycleswere insufficient to make Mars warmenough to sustain liquid waters. The plan-et had clement conditions for only briefperiods after major impacts. Each suchimpact deposited water-rich material andpumped enough heat and water into theatmosphere to permit rain. Soon, though,the planet returned to its usual frozenstate. Victor Baker of the University ofArizona has argued that the intensive vol-canism in the Tharsis region periodicallymade early Mars quite a temperate place.

It is also very possible that none ofthese options is correct. We simply do notyet know enough about early Mars tohave any real understanding of its cli-mate. We must wait for future explo-ration. Unlike Earth, Mars has preservedmuch of its ancient landscape, which mayyield clues to the conditions under whichit formed. Indeed, understanding howMars became so different from Earth willhelp geologists grapple with Earth’s ownhistory. The new lander missions willsoon provide some of these clues.

A broadcast version of this article will air May 29 on National

Geographic Today, a program on the National Geographic Channel. Please check your local listings.


NAS

A/JP

L/M

ALIN

SP

ACE

SC

IEN

CE

SYS

TEM

S M

OC

2–

26

4;

CAL

VIN

J.

HAM

ILTO

N (

inse

t)

Mars 2000. Arden L. Albee in Annual Review of Earth and Planetary Sciences, Vol. 28, pages 281–304; 2000.

Mars: The Lure of the Red Planet. William Sheehan and Stephen James O’Meara. Prometheus Books, 2001.

The Mars Global Surveyor Mars Orbiter Camera: Interplanetary Cruise through Primary Mission.Michael C. Malin and Kenneth S. Edgett in Journal of Geophysical Research, Vol. 106, No. E10, pages 23429–23570; October 25, 2001.

Nature Insight: Mars. Special section in Nature, Vol. 412, pages 207–253; July 12, 2001.

For information on NASA’s Mars missions, visit www.jpl.nasa.govFor information on the European Space Agency’s Mars Express, visit sci.esa.int/marsexpressFor information on the Japanese Nozomi mission, visit www.isas.ac.jp/e/enterp/missions/index.html


“WHITE ROCK,” seen by the Viking spacecraft in the 1970s (inset), is a prime example of how Mars’sEarth-like appearance can deceive. The feature looks tantalizingly like a pile of salt deposited byliquid water. But spectral measurements now show it to be generic dust that somehow got compactedor cemented together. The reddish dust buried existing features, such as the crater at the top right,and was in turn buried by black sand. The image, taken by Mars Global Surveyor, attests to a bafflingly complex sequence of geologic events.

100 METERS

CRATER

LAYERS


D igital computing performance has improved 10,000-fold in the past two

decades: what took a year of number crunching in 1983 takes less than an hour

nowadays, and a desktop computer from that era can’t match the processing power

of one of today’s handheld organizers.

We pay a price for these enhancements, though. As digital systems have grown in

complexity, their operation has become brittle and unreliable. Computer-related fail-

ures have become all too common. Personal computers crash or freeze up regularly;

Internet sites go offline often. New software upgrades, designed to augment perfor-

mance, may leave things worse than they were before. Inconvenience aside, the situa-

tion is also an expensive one: annual outlays for maintenance, repairs and operations

far exceed total hardware and software costs, for both individuals and corporations.

By embracing the inevitability of system failures, RECOVERY–

BY Armando Fox AND David Patterson

54 S C I E N T I F I C A M E R I C A NCOPYRIGHT 2003 SCIENTIFIC AMERICAN, INC.

– ORIENTED computing returns service faster

S C I E N T I F I C A M E R I C A N 55COPYRIGHT 2003 SCIENTIFIC AMERICAN, INC.

Our group of research collaboratorsat Stanford University and the Universityof California at Berkeley has taken a newtack, by accepting that computer failureand human operator error are facts of life.Rather than trying to eliminate computercrashes—probably an impossible task—

our team concentrates on designing sys-tems that recover rapidly when mishapsdo occur. We call our approach recovery-oriented computing (ROC).

We decided to focus our efforts on im-proving Internet site software. This kindof highly dynamic computing system mustevolve and expand quickly in response toconsumer demands and market pres-sures—while also serving users who expectinstant access at any time. Consider the ex-ample of the Google search engine, whichin just a few years has gone from locatinghundreds of millions of Web pages of Eng-lish text to three billion pages in more than20 languages in a dozen formats, plus im-ages. Meanwhile the number of dailyGoogle searches has grown from 150,000to 150 million—the site is now 1,000 timesbusier than it was at the outset.

Because of the constant need to up-grade the hardware and software of In-ternet sites, many of the engineering tech-niques used previously to help maintainsystem dependability are too expensive tobe deployed. Hence, we expect Internetsoftware to be a good proving ground forour ideas and perhaps a model for othercomputing systems, including desktopand laptop machines. If ROC principlescan help the big animals in the computa-tional jungle, they might do the same forthe smaller species.

Following a proven engineering strat-

egy first adopted during the era of cast-iron-truss railroad bridges in the 19thcentury, our initial step was to see whatwe could learn from previous failures.Specifically, we asked: Why do Internetsystems go down, and what can be doneabout it? We were a bit surprised to findout that operator error was a leadingcause of system problems. Traditional ef-forts to boost the dependability of soft-

ware and hardware have for the mostpart overlooked the possibility of humanmistakes, yet in many cases operators’miscues accounted for more downtimethan any other cause [see box above].

Operators may face such difficultiesbecause computer designers and pro-grammers have frequently sacrificed easeof use in the quest for better performance.Database software, for example, can re-quire a full-time staff of trained adminis-trators to manage it. Ironically, becausehardware and software have growncheaper over time, operator salaries arenow often the biggest expense in runningcomplex Internet sites.

With these issues in mind, our team isexploring four principles to guide the con-struction of “ROC-solid” computing sys-tems. The first is speedy recovery: prob-lems are going to happen, so engineersshould design systems that recover quick-ly. Second, suppliers should give opera-tors better tools with which to pinpoint


SLIM

FIL

MS

(pre

ced

ing

pa

ges

); J

OH

NN

Y JO

HN

SON

(a

bov

e),

SOU

RC

E F

OR

PIE

CH

ART:

DAV

ID O

PP

EN

HE

IME

R

■ Despite the undoubted power of today’s computers, users continue to betormented by their systems’ stubborn unreliability. Recovery-orientedcomputing (ROC) design practices could do much to solve this predicament.

■ ROC principles—which comprise efforts to engineer rapid-recovery capabilities,software tools to locate faults quickly, “undo” functions to reverse humanoperators’ mistakes, and the means to inject errors to test systems’ ability toreturn to service—may eventually take much of the frustration out of computing.

■ Benchmarking programs that evaluate the speed with which computer systems return to full service would encourage industry efforts at improving dependability.

Overview/High-Dependability Computing

FOURDAYS

DOWNTIME BY “NINES”

NINE HOURS

Hardware 15%Software 34%

Operator error 51%

Annu

al D

ownt

ime

(min

utes

)ONE HOUR FIVE

MINUTES

99.99999.9999.9

Availability Metric (percent) 99.0

REASONS FOR WEB SITE FAILURES

600

500

400

300

200

100

0

5,000

6,000

WHOSE FAULT WAS THAT?TRADITIONAL ENGINEERING approaches to raisingthe reliability of computer systems largely ignorethe possibility of operator error. But in manycases, human mistakes account for moredowntime (time during which the system is notfunctioning) than hardware problems or softwarebugs. The pie chart (right) depicts a breakdown oftypical failure causes for three Internet sites.

For many industries, computer systemdowntime can be costly or even life-threatening.Engineers call the proportion of time a computingsystem functions correctly itsavailability, which is measured by“nines” ( graph). A system that runswithout crashing 99.999 percent of thetime, for example, has an availability of“five nines,” which corresponds toabout two hours of downtime over 25years of operation. Rather than seekingto reduce the number of failures,proponents of recovery-orientedcomputing advocate methods toshorten the time needed to bringsystems back online. Boostingavailability from two nines to five nines,for instance, shrinks total recovery timefrom 90 hours to five minutes a year.


the sources of faults in multicomponentsystems. Third, programmers ought tobuild systems that support an “undo”function (similar to those in word-pro-cessing programs), so operators can cor-rect their mistakes. Last, computer scien-tists should develop the ability to inject testerrors; these would permit the evaluationof system behavior and assist in operatortraining. We plan to release all the ROC-inspired software applications we write tothe computing community at no cost.

To foster the adoption of our ap-proach, we also advocate the develop-ment and distribution of benchmark pro-grams that would test the speed of a com-puting system’s recovery. This softwarewould measure the computer industry’sprogress in raising dependability and en-courage companies to work towardachieving this end.

Quick ComebackMANY A USER REBOOTS his or herpersonal computer routinely—either pre-emptively, because the machine is behav-ing strangely, or reactively, because it hascrashed or seized up. Rebooting worksfor large computers, too, because it wipesthe slate clean and fixes a whole class ofso-called transient failures—that is, prob-lems that appear intermittently.

Unfortunately, most systems take along time to reboot and, worse, may losedata in the process. Instead we believe thatengineers should design systems so thatthey reboot gracefully. If one were to lookinside a computer, one would see that it isrunning numerous different software com-ponents that work together. During on-line shopping, for instance, some softwaremodules let customers search through theavailable merchandise; others permititems to be added to a “shopping cart.”Still other software enables the comple-tion of the purchase. Yet another layer ofprogramming choreographs all thesefunctions to produce the overall experience

of using the Web site by ensuring that eachportion of code does its job when required.

Frequently, only one of these modulesmay be encountering trouble, but whena user reboots a computer, all the soft-ware it is running stops immediately. Ifeach of its separate subcomponents couldbe restarted independently, however, onemight never need to reboot the entire col-lection. Then, if a glitch has affected onlya few parts of the system, restarting justthose isolated elements might solve theproblem. Should that prove ineffective,reinitializing a larger set of subcompo-nents might work. The trick is to be ableto restart one module without acciden-tally confusing its peers into thinkingsomething is really wrong, which is akinto swapping out the bottom plate in astack without allowing the other plates tofall—a difficult but doable feat.

George Candea and James Cutler,Stanford graduate students on our team,have focused on developing this indepen-dent-rebooting technique, which we callmicro-rebooting. Cutler’s experience in-cludes building systems that use inexpen-sive ground receivers assembled from off-the-shelf PCs, low-cost ham radios andexperimental software to capture incom-ing satellite data. In use, ground-stationfailures were common, and if a humanoperator was not available to reactivatethe equipment manually, the satellite sig-nal could be lost—and with it, all the datafor that orbit.

Last year Candea and Cutler testedmicro-rebooting on the ground-stationsoftware. They and others modified eachreceiving-station software module so thatit would not “panic” if other subcompo-nents were reinitialized. The students firstconsulted the human operators to learnabout the most frequent causes of failuresand then experimented to determine whichsets of subcomponents would have to bereinitialized to cure those specific mal-adies. They succeeded in automating the

recovery process for a range of recurringproblems, cutting the average restorationtime from 10 minutes to two—fast enoughfor a ground station that has faltered toreacquire the satellite signal and continuecollecting data for the latest orbit.

In dependability lingo, the percentageof time a computer system is functioningcorrectly is termed its availability, whichis typically measured in “nines.” A systemthat is functioning correctly 99.999 per-cent of the time, for example, has anavailability of “five nines,” correspondingto about two hours of downtime over 25years of operation. In contrast, well-man-aged mainstream computing systems areavailable only 99 to 99.9 percent of thetime (“two to three nines”). Going fromtwo nines to five nines would save almost90 hours of downtime a year, which is easyto appreciate when loss of service costs bigmoney [see box on opposite page]. Twomethods exist to achieve high dependabil-ity: ensuring fewer breakdowns or, failingthat, bringing systems back online faster.

For the satellite ground-station oper-ators, a fivefold-faster return to servicewas much more valuable than a fivefoldincrease in the time between failures (bet-ter reliability), even though either mea-sure would yield the same level of im-proved availability. We believe that a va-riety of computing systems exhibit sucha threshold.

Although much effort had to be ex-pended to modify the ground-station soft-ware manually, Candea and one of us(Fox) are now investigating whether thetechnique can be applied in an automat-ed fashion to Web sites programmed us-ing the Java 2 Enterprise Edition, a popu-lar development framework for Internetsoftware.

The most common way to fix Website faults today is to reboot the entire sys-tem, which takes anywhere from 10 sec-onds (if the application alone is rebooted)to a minute (if the whole thing is restart-


Methods to ACHIEVE HIGH DEPENDABILITY: ensurefewer breakdowns or bring systems back online faster.



SAM

UE

L VE

LASC

O

OLD METHOD NEW METHOD

THREE WEB USERS (A, B and C) access a site. A and B are usingsubcomponent 2; C is not.

Subcomponent 2 fails. Users A and B receive an error message. PinPoint notices and deduces that subcomponent 2 is most likely to be at fault from the fact that both A and B are using it.

2Subcomponent 2 fails. Users A and B receive an error message. The human operator cannot determine which subcomponent is at fault.

Because the faulty subcomponent cannot be located, the operator must reboot the entire system.

2

3

All the subcomponents had to be rebooted, so the system takes a long time to recover.

4

All the software subcomponents in the Internet service are working correctly. 1

Micro-Reboot consults its own database andlearns that when subcomponent 2fails, 1 is typically also affected.It decides to rebootboth. User C is not affected.

3 Pin-Point

Pin-Point

Because only two subcomponents were rebooted, the entire system returnsto function more quickly.

4

“When 2 fails, 1 is usually affected, too”

?

Micro-Reboot

Micro-Reboot

Pin-Point

“Looks likesubcomponent 2 has failed”

1

4

23

A

B

1

4

23

1

4

23

1

4

23

1

4

23

C

A

B

C

A

B

C

Error

Error

A

B

C

1

4

23A

B

C

1

4

23A

B

C

HUMAN OPERATOR

DOWNTIMECLOCKError

Error

SOFTWARE AGENTS

SOFTWARE SUBCOMPONENTINTERNET SITE SOFTWARE

Micro-Reboot

A

B

C

Error

Error

Error

Error

Error

HOW SELECTIVE REBOOTING CUTS DOWNTIME


ed). According to our initial results, micro-rebooting just the necessary subcompo-nents takes less than a second. Instead ofseeing an error message, a user would ex-perience a three-second delay followed byresumption of normal service.

Pinpointing ProblemsIT’S ONE THING to fix known or like-ly errors, but a tougher, related challengeis to find the unanticipated ones. Systemoperators could use assistance in trackingdown problems more quickly, the secondof our ROC principles.

When building a traditional high-de-pendability computer system, program-mers start with a complete description ofall hardware and software elements. Theythen follow with a carefully constructedfailure-analysis tree, which traces the myr-iad ways the system can break down—sothat those may be prevented or, if they dooccur, corrected. In contrast to single-source systems, Internet services are het-erogeneous, using components from mul-tiple vendors. Further, these modules of-ten change rapidly as the service evolves.Failures frequently arise from unexpect-ed interactions between components,rather than resulting from a bug in a sin-gle piece of software. When this kind ofdynamic fault occurs, a Web user whohappens to be accessing the service at thetime may receive an error message.

To help analyze these complex mal-functions, graduate students Emre Kici-man and Eugene Fratkin of Stanford andMike Chen of Berkeley created PinPoint,a ROC-based computer program that at-tempts to determine which componentsare at fault. Every time someone surfs toa PinPoint-enabled Web site, the programtraces which software components par-ticipated in delivering service to that user.When a particular access request fails—

for example, the user gets an error mes-sage from the site—PinPoint notes thisfact. Over time, the monitoring applica-

tion analyzes the mix of components thatwere activated in both failed and success-ful requests, using standard data-miningtechniques. By doing this, PinPoint canhelp find out which components are sus-pected of causing most of the failures. Theadditional information gathered by thefailure-analysis code slows the systemdown by at most 10 percent. Unlike thetraditional solution—which requires elab-orate preplanning every time the softwaresuite changes—PinPoint works with anycombination of software components.

Wiping Away ErrorsPERHAPS THE GREATEST challenge inboosting system reliability is ensuring amargin of safety against random errorsinputted by the operator; this rationaleunderlies our third ROC principle, whichconcerns the undo command. The firstword processors did not have this capa-bility, which made them frustrating, if notterrifying, to use. A single erroneous glob-al substitution could destroy an entire file.The undo function—which affords usersthe ability to cancel any command—re-moved the anxiety from word processing.

Operators of today’s large computingsystems have no such option. When thefoundations of information technologywere laid, no one considered it important

to be able to expunge mistakes. That’s be-cause an undo function requires morework to construct, consumes a significantamount of storage space and probablyslows systems down somewhat.

To demonstrate a better approach,our group is working on an undo capa-bility for e-mail systems that is aimed atthe place where messages are stored.Berkeley graduate student Aaron Brownand one of us (Patterson) have recentlycompleted the prototype of an e-mail sys-tem featuring an operator undo utility. Weare testing it now [see box on page 61].

Suppose a conventional e-mail storageserver gets infected by a virus. The systemoperator must disinfect the server, a la-borious job. Our system, however, wouldrecord all the server’s activities automat-ically, including discarded messages. If thesystem gets infected, the operator couldemploy the undo command to “turn backthe clock” to before the arrival of thevirus. Software that attacks that viruscould then be downloaded. Finally, theoperator could “play forward” all the e-mail messages created after the infection,returning the system to normal operation.The newly installed antivirus softwarewould filter all subsequent e-mail traffic.In this way, the operator could undo thedamage without losing important mes-


ARMANDO FOX and DAVID PATTERSON have studied how to improve the reliability of com-puting systems for many years. Fox has been assistant professor at Stanford Universitysince 1999. As a Ph.D. student at the University of California, Berkeley, he joined with Pro-fessor Eric A. Brewer in constructing prototypes of today’s clustered Internet services andshowing how they can support mobile computing applications, including the first graphicalWeb browser for personal digital assistants. Fox also helped to design microprocessors atIntel and later founded a small mobile computing company. He received his other degreesin electrical engineering and computer science from the Massachusetts Institute of Tech-nology and the University of Illinois. Patterson has spent the past quarter of a century asa professor at Berkeley. He received all his advanced degrees in computer science from theUniversity of California, Los Angeles. Patterson is best known for work on the simplificationof microprocessor architecture, for building reliable digital storage systems and for co-authoring the classic books Computer Architecture: A Quantitative Approach and ComputerOrganization and Design: The Hardware/Software Interface. He first published in the pagesof Scientific American two decades ago.

THE

AU

THO

RS

Computers may end up 10,000 TIMES FASTERyet no more dependable than today’s machines.


sages. To prevent potential confusionamong users—who might notice thatsome e-mails had been eradicated—thesystem could send a message saying thatnotes were deleted during an attempt tostop the spread of a virus.

Injecting Test ErrorsLAST AMONG OUR ROC principles isthe idea of harnessing errors to do good:we advocate testing the system periodical-ly by inserting artificial faults. This prac-tice would aid in evaluating the recoveryperformance of a system and also in com-ing up with new methods to make it morerobust. In an industry analogy, designersof microprocessor chips have regularlyadded circuits to simplify the testing ofchips, even though these additions in-crease chip size and remain unused afterthe microprocessors leave the factory.Manufacturers consider these test circuitsworth the effort; they lower the cost of en-suring that the completed chips work asplanned. Part of this benefit arises becausethe test circuits allow designers to inject“failures” artificially to verify that the chipdetects and recovers from them correctly.

Our group proposes a software equiv-alent for computer systems. When opera-tors employed the selective-rebootingstrategy, for example, test errors wouldhelp determine which software compo-nents to reinitialize to counter a particu-lar kind of failure. If the problem propa-gated to only one or two other elements,operators could reboot only those. If theflaw came to involve a great many com-ponents, it might be more sensible to re-boot the entire system. We have startedusing error injection to characterize thefault-propagation behavior of Internetsites built using Java 2 Enterprise Edition.

Another version of this softwarecould permit potential buyers to see howa computer system handled failures—abenchmark program that could informtheir purchase decisions. Developed byBerkeley graduate students Pete Broad-well, Naveen Sastry and Jonathan Traup-man, the Fig application tests the abilityof programs to cope correctly with un-expected errors in the standard C library,a part of the operating system used bynearly all software programs. “Fig”


SAM

UE

L VE

LASC

O

STRIVING FOR DEPENDABILITYCOMPUTER SYSTEMS and their “organs”—microprocessors, applications andcommunications networks—are becoming ever more powerful. But they are alsobecoming ever more complex and therefore more susceptible to failure. As the costs ofadministration, oversight and downtime expand in response, scientists and engineersin the computer industry are working to enhance the dependability of their products.Significantly, many of their efforts aim to take humans (and the errors they inevitablyengender) out of the loop.

Concerned about security holes, bugs and other weaknesses in its presentproduct line, Microsoft’s management took the unusual step recently of haltingsoftware development for an entire month to focus on what it calls TrustworthyComputing. The issue of dependability has grown in importance as moreadministrators adopt the company’s Windows operating system to run Web servers.Operating-system developers at Microsoft attended classes to learn techniques thatimprove security and reliability for desktop systems. They are now refining Windows forthe next version, called Palladium. Engineers plan to cull potential weak points fromcurrent products while developing new features that boost defenses against hackers.

Little research exists on ways to reduce the lifetime cost ofcomputer ownership—the price individuals and companies pay forrunning their systems. Programmers at Hewlett-PackardLaboratories and IBM Research are working to cut thoseexpenses by adding new capabilitiesand developing products that canmanage themselves. Hewlett-Packardofficials envision a globally networkedsystem of computational and storageresources that monitor, heal andadapt themselves without operatorintervention. HP’s PlanetaryComputing project concentrates ondeveloping corporate data centersthat could contain as many as 50,000 individual office computers, a collection 10times as large as today’s counterparts.

IBM’s scheme borrows ideas from control theory (the use of feedback to stabilizeclosed-loop systems) and artificial intelligence (mimicking or otherwise capturingexpert human skills or intelligence to solve complex problems). These concepts willhelp create data centers that can diagnose problems on their own, adjust theirconfigurations to match changes in demand, repair themselves and defend againsthacker attacks. Drawing an analogy with the body’s autonomic nervous system, IBM’smanagement calls this goal Autonomic Computing.

When designers of other engineering systems have discovered a propensity foroperator error, they often have attempted to remove the need for human input.Removing human operators can lead to a well-established pitfall known as theAutomation Irony. Because designers can typically reduce but not eliminate the needfor human intervention, such efforts frequently make things worse. That’s becauseengineers generally automate the tasks that are easy, leaving the hard jobs forpeople. These measures mean that administrators must carry out difficult tasksintermittently on unfamiliar systems—a sure recipe for failure.

Will the path toward truly dependable computing be additional automation,leading to hands-off computing, or will it be streamlined design combined with toolsthat dramatically improve the effectiveness and productivity of human operators?Only time will tell. —A.F. and D.P.


stands for “Fault Injection in glibc” (aversion of the standard C library used bynumerous programmers).

Error injection would also permitcomputer programmers to test their re-pair mechanisms, which is difficult to do.Fig would allow operators to try theirhands at diagnosing and repairing fail-ures, perhaps on a small experimental sys-tem if not on the real thing. The programhas been used several times and is avail-able for no charge on our ROC Web site[see More to Explore below].

Benchmarking RecoveryTHE HISTORY OF the computer indus-try makes us strong believers in the im-portance of measuring technical progressand publicizing it. When computer firmsfinally adopted standard benchmarkingprograms to compare performance (aftermany years of delay), customers could atlast see the relative merits of each prod-uct clearly. Companies that trailed intechnology were forced to spend more onengineering and subsequently couldgauge the effect of their innovations us-ing standard evaluation metrics. The re-sulting test data led to a continuing

cascade of performance improvements.By focusing on evaluating recovery

time, Brown and fellow Berkeley graduatestudent David Oppenheimer, togetherwith Patterson, are working to re-createthat kind of competition for computersystem dependability. Products that wereshown to come back from crashes quick-est would win greater sales. We envisiona test suite that would incorporate com-mon failures, including errors caused byhumans, software and hardware. Pro-spective buyers could insert these faultsinto systems and then monitor the time torecovery. It is noteworthy that currentcomputer-marketing efforts more typi-

cally quote availability (percentage down-time) metrics, which are much more diffi-cult to measure than recovery time.

When scientists and engineers focustheir efforts, they can achieve amazingprogress in a relatively short time. The me-teoric 30-year rise of both computer per-formance and cost-effectiveness proves it.If the industry continues traveling blindlydown the current path, computers in 2023may end up another 10,000 times fasteryet no more dependable than today’s ma-chines. But with dependability-enhancingsoftware tools and benchmarks to inspireus, computing may one day become as re-liable as users expect it to be.


SAM

UE

L VE

LASC

O

To Engineer Is Human: The Role of Failure in Successful Design. H. Petroski. Vintage Books, 1992.

The Mythical Man-Month: Essays on Software Engineering. Second edition. Frederick P. Brooks. Addison-Wesley, 1995.

Managing the Risks of Organizational Accidents. James Reason. Ashgate Publishing Company, 1997.

Autonomic Computing. IBM Research, 2001. Available atwww.research.ibm.com/autonomic/manifesto Building a Secure Platform for Trustworthy Computing. Microsoft, December 2002. Available at www.microsoft.com/security/whitepapers/secure–platform.aspPlanetary Computing Web site at HP Labs: www.hpl.hp.com/org/issl/current–research–themes.htmRecovery Oriented Computing Web site at University of California at Berkeley and StanfordUniversity: http: //roc.cs.berkeley.eduTuring. Christos Papadimitriou. MIT Press (in press).


SPRING FORWARD The operator now instructs the undo module to “redo” the “undone” actions, causing all messages to be redelivered to the e-mail server. This time, however, the new virus filter intercepts and deletes the infected message. The intended recipient of the e-mail receives notice.

3

Undo!

Redo!“An infected e-mail

message was deleted

to protect you from

a virus”

UndoModule

E-mailServer

E-mailServer UndoModule

MARK TIME An e-mail user sends out a virus-laden message to a friend. The human operator of the e-mail system cannot protect the server from an undetected virus. But because the “undo” module monitors and logs all messages traveling in and out (including the infected one), it can fix these kinds of problems once they become apparent.

1

FALL BACK The operator hears that a new virus is circulating and suspects that the server may be infected. Fortunately, the infected message has not yet been forwarded to its intended recipient. The operator tells the undo module to use its undo/redo log data to “undeliver” all messages that arrived since the virus appeared. The operator installs a virus filter.

2

INFECTED SENDER VIRUS

UndoModule

INTENDED MESSAGE RECIPIENTHUMAN OPERATOR

E-mailServer

VIRUS FILTER

Undo/RedoLogs

Undo/RedoLogs

Undo/RedoLogs

TURNING THE CLOCK BACK ON PROBLEMS



RYS

ZAR

D H

OR

OW

ITZ

Cor

bis


On July 25, a once unique person will turn 25.This nursery school aide in the west of England seems like an av-erage young woman, a quiet, shy blonde who enjoys an occa-sional round of darts at the neighborhood pub. But LouiseBrown’s birth was greeted by newspaper headlines calling her the“baby of the century.” Brown was the world’s first test tube baby.

Today people may remember Brown’s name, or that she wasBritish, or that her doctors, Steptoe and Edwards, soundedvaguely like a vaudeville act. But the past quarter of a centuryhas dimmed the memory of one of the most important aspectsof her arrival: many people were horrified by it. Even some sci-entists feared that Patrick Steptoe and Robert Edwards mighthave brewed pestilence in a petri dish. Would the child be nor-mal, or would the laboratory manipulations leave dreadful ge-netic derangements? Would she be psychologically scarred bythe knowledge of how bizarrely she had been created? And wasshe a harbinger of a race of unnatural beings who might even-tually be fashioned specifically as a means to nefarious ends?

Now that in vitro fertilization (IVF) has led to the birth ofan estimated one million babies worldwide, these fears and spec-ulations may seem quaint and even absurd. But the same con-cerns once raised about IVF are being voiced, sometimes almostverbatim, about human cloning. Will cloning go the way of

IVF, morphing from the monstrous to the mundane? Andif human cloning, as well as other genetic interventions

on the embryo, does someday become as com-monplace as test tube baby–making, is that

to be feared—or embraced? Thelessons that have been

In vitro fertilization was onceconsidered by some to be athreat to our very humanity.Cloning inspires similar fears

BY ROBIN MARANT Z HENIG

PANDORA’S

BABY



PAS

CAL

GO

ETG

HE

LUC

K S

cien

ce P

hot

o Li

bra

ry (

left

); R

AY F

OLI

Cor

bis

-Bet

tma

nn

(ri

gh

t)

learned from the IVF experience can illu-minate the next decisions to be made.

Then and NowAS IVF MOVED FROM the hypotheticalto the actual, some considered it to benothing more than scientists showing off:“The development of test tube babies,”one critic remarked, “can be compared tothe perfecting of wing transplants so thatpigs might fly.” But others thought of IVFas a perilous insult to nature. The Britishmagazine Nova ran a cover story in thespring of 1972 suggesting that test tubebabies were “the biggest threat since theatom bomb” and demanding that thepublic rein in the unpredictable scientists.“If today we do not accept the responsi-bility for directing the biologist,” theNova editors wrote, “tomorrow we maypay a bitter price—the loss of free choiceand, with it, our humanity. We don’thave much time left.”

A prominent early enemy of IVF wasLeon Kass, a biologist at the University ofChicago who took a professional interestin the emerging field of bioethics. If soci-ety allowed IVF to proceed, he wrote

shortly after Louise Brown’s birth, someenormous issues were at stake: “the ideaof the humanness of our human life andthe meaning of our embodiment, our sex-ual being, and our relation to ancestorsand descendants.”

Now read Kass, a leading detractorof every new form of reproductive tech-nology for the past 30 years, in 2003:“[Cloning] threatens the dignity of hu-man procreation, giving one generationunprecedented genetic control over thenext,” he wrote in the New York Times.“It is the first step toward a eugenic worldin which children become objects of ma-nipulation and products of will.” Suchcommentary coming from Kass is partic-ularly noteworthy because of his uniqueposition: for the past two years he hasbeen the head of President George W.Bush’s Council on Bioethics, whose firsttask was to offer advice on how to regu-late human cloning.

Of course, IVF did not wind up cre-ating legions of less than human children,nor did it play a role in the disintegrationof the nuclear family, consequences thatpeople like Kass feared. And so many

newer, more advanced methods of assist-ed reproduction have been introduced inthe past decade that the “basic IVF” thatproduced Louise Brown now seems pos-itively routine. One early prediction, how-ever, did turn out to contain more than akernel of truth. In the 1970s critics cau-tioned that IVF would set us tumblingdown the proverbial slippery slope to-ward more sophisticated and, to some,objectionable forms of reproductive tech-nology—and that once we opened thefloodgates by allowing human eggs to befertilized in the laboratory, there wouldbe no stopping our descent.

If you consider all the techniques thatmight soon be available to manipulate adeveloping embryo, it could appear thatthe IVF naysayers were correct in their as-sessment of the slipperiness of the slope.After all, none of the genetic interventionsnow being debated—prenatal genetic di-agnosis, gene insertions in sex cells or em-bryos to correct disease, the creation ofnew embryonic stem cell lines and, theelephant in the living room, cloning—

would even be potentialities had scientistsnot first learned how to fertilize humaneggs in a laboratory dish.

But does the existence of a such a slip-pery slope mean that present reproductivetechnology research will lead inevitably todevelopments that some find odious, suchas embryos for tissue harvesting, or theeven more abhorrent manufacture of hu-man-nonhuman hybrids and humanclones? Many people clearly fear so,which explains the current U.S. efforts to

■ Many arguments against in vitro fertilization in the past and cloning todayemphasize a vague threat to the very nature of humanity.

■ Critics of IVF attempted to keep the federal government from supporting theresearch and thus ironically allowed it to flourish with little oversight.

■ Because of the lack of oversight, it is only in the past few years that the increasedrate of birth defects and low birth weight related to IVF have come to light.

Overview/In Vitro Veritas

MICRONEEDLE INJECTS a sperm’s package of DNA directly into a human egg,thereby achieving in vitro fertilization (left). The first human being born asa result of IVF, Louise Brown was 14 months old when she frolicked on the

set of the Donahue television program (right). With her was VanderbiltUniversity IVF researcher Pierre Soupart, who predicted that “by the timeLouise is 15, there will be so many others it won’t be remarkable anymore.”


curtail scientists’ ability to manipulateembryos even before the work gets underway. But those efforts raise the questionof whether science that has profoundmoral and ethical implications should sim-ply never be done. Or should such scienceproceed, with careful attention paid to theearly evolution of certain areas of researchso that society can make informed deci-sions about whether regulation is needed?

IVF UnboundTHE FRENZY TO TRY to regulate oreven outlaw cloning is in part a deliberateattempt not to let it go the way of IVF,which has been a hodgepodge of unregu-lated activities with no governmental orethical oversight and no scientific coordi-nation. Ironically, the reason IVF becameso ubiquitous and uncontrolled in theU.S. was that its opponents, particularlyantiabortion activists, were trying to stopit completely. Antiabortion activists’ pri-mary objection to IVF was that it in-volved the creation of extra embryos thatwould ultimately be unceremoniously de-stroyed—a genocide worse than at anyabortion clinic, they believed. According-ly, they thought that their best strategywould be to keep the federal governmentfrom financing IVF research.

A succession of presidential commis-

sions starting in 1973 debated the ethicsof IVF but failed to clarify matters. Someof the commissions got so bogged downin abortion politics that they never man-aged to hold a single meeting. Others con-cluded that IVF research was ethically ac-ceptable as long as scientists honored theembryo’s unique status as a “potentialhuman life,” a statement rather than apractical guideline. In 1974 the govern-ment banned federal funding for fetal re-search. It also forbade funding for re-search on the human embryo (defined asa fetus less than eight weeks old), whichincludes IVF. In 1993 President Bill Clin-ton signed the NIH Revitalization Act,which allowed federal funding of IVF re-search. (In 1996, however, Congress againbanned embryo research.) The bottomline is that despite a series of recommen-dations from federal bioethics panels stat-ing that taxpayer support of IVF researchwould be acceptable with certain safe-guards in place, the government has nev-

er sponsored a single research grant forhuman IVF.

This lack of government involve-ment—which would also have served todirect the course of IVF research—led toa funding vacuum, into which rushed en-trepreneurial scientists supported by pri-vate money. These free agents did essen-tially whatever they wanted and whatev-er the market would bear, turning IVFinto a cowboy science driven by the mar-ketplace and undertaken without guid-ance. The profession attempted to regu-late itself—in 1986, for example, theAmerican Fertility Society issued ethicaland clinical guidelines for its members—

but voluntary oversight was only sporad-ically effective. The quality of clinics, ofwhich there were more than 160 by 1990,remained spotty, and those seeking IVFhad little in the way of objective informa-tion to help them choose the best ones.

Today, in what appears to be an effortto avoid the mistakes made with IVF, thefederal government is actively involved inregulating cloning. With the announce-ment in 1997 of the birth of Dolly, the firstmammal cloned from an adult cell, Presi-dent Clinton established mechanisms,which remain in place, to prohibit such ac-tivities in humans. Congress has made sev-eral attempts to outlaw human cloning,most recently with a bill that would makeany form of human cloning punishable bya $1-million fine and up to 10 years inprison. (The House of Representativespassed this bill this past winter, but theSenate has yet to debate it.) Politicians thuslumped together two types of cloning thatscientists have tried to keep separate:“therapeutic,” or “research,” cloning, de-signed to produce embryonic stem cellsthat might eventually mature into spe-cialized human tissues to treat degenera-tive diseases; and “reproductive” cloning,undertaken specifically to bring forth acloned human being. A second bill now

WIL

LIAM

PLO

WM

AN G

etty


ROBIN MARANTZ HENIG has written seven books, most recently The Monk in the Garden: TheLost and Found Genius of Gregor Mendel. Her articles have appeared in the New York TimesMagazine, Civilization and Discover, among other publications. Her honors include an AliciaPatterson Foundation fellowship and a nomination for a National Book Critics Circle Award.She lives in New York City with her husband, Jeffrey R. Henig, a political science professor atColumbia University; they have two nearly grown daughters. Her next book, entitled Pan-dora’s Baby, is about the early days of in vitro fertilization research.

THE

AU

THO

R

MEMBERS of the Christian Defense Coalition and the National Clergy Council protest Advanced CellTechnologies’s human cloning research outside the biotechnology firm’s headquarters in Worcester,Mass., on November 30, 2001. Similar protests against IVF occurred in the 1970s.



AP P

HO

TO

before the Senate would explicitly protectresearch cloning while making reproduc-tive cloning a federal offense.

IVF Risks RevealedONE RESULT OF the unregulated na-ture of IVF is that it took nearly 25 yearsto recognize that IVF children are at in-creased medical risk. For most of the1980s and 1990s, IVF was thought tohave no effect on birth outcomes, with theexception of problems associated withmultiple births: one third of all IVF preg-nancies resulted in twins or triplets, theunintended consequence of the wide-spread practice of implanting six or eightor even 10 embryos into the womb duringeach IVF cycle, in the hope that at leastone of them would “take.” (This brute-force method also leads to the occasionalset of quadruplets.) When early studiesraised concerns about the safety of IVF—

showing a doubling of the miscarriagerate, a tripling of the rate of stillbirths andneonatal deaths, and a fivefold increase inectopic pregnancies—many people attrib-uted the problems not to IVF itself but toits association with multiple pregnancies.

By last year, however, IVF’s medicaldark side became undeniable. In March2002 the New England Journal of Medi-cine published two studies that controlledfor the increased rate of multiple birthsamong IVF babies and still found prob-lems. One study compared the birthweights of more than 42,000 babies con-ceived through assisted reproductive tech-nology, including IVF, in the U.S. in 1996and 1997 with the weights of more thanthree million babies conceived naturally.Excluding both premature births andmultiple births, the test tube babies werestill two and a half times as likely to havelow birth weights, defined as less than2,500 grams, or about five and a halfpounds. The other study looked at morethan 5,000 babies born in Australia be-tween 1993 and 1997, including 22 per-cent born as a result of IVF. It found thatIVF babies were twice as likely as natu-rally conceived infants to have multiplemajor birth defects, in particular chro-mosomal and musculoskeletal abnormal-ities. The Australian researchers speculatethat these problems may be a consequence

From Outrage to ApprovalTHE STORY of Doris Del-Zio demonstrates the ironies resulting from society’schanging attitude toward IVF in the 1970s. After years of failure to conceive a child,Del-Zio and her husband turned to Landrum Shettles of what is now known as theColumbia Presbyterian Medical Center. In the fall of 1973 Shettles prepared toattempt a hasty IVF procedure on the couple. The operation was abruptly terminatedby Shettles’s superior, Raymond Vande Wiele, who was outraged at Shettles’saudacity and who questioned the medical ethics of IVF. Vande Wiele confiscated andfroze the container holding the Del-Zios’ eggs and sperm. As far as the Del-Zios wereconcerned, Vande Wiele had committed murder: they sued him and his employers for$1.5 million.

By coincidence, the Del-Zios’ case against Vande Wiele was finally brought totrial in July 1978, the same month that Louise Brown was born. The birth of theworld’s first test tube baby put Shettles’s early IVF attempt in a different light. AfterBrown’s appearance, most people—including the two men and four women on theDel-Zio jury—seemed much more inclined to think of IVF as a medical miracle than asa threat to civilized society.

The trial lasted six weeks, each side making its case about the wisdom, safetyand propriety of IVF. In the end, Vande Wiele was found to be at fault for “arbitraryand malicious” behavior, and he and his co-defendants were ordered to pay DorisDel-Zio $50,000.

IVF developed rapidly after the trial, and 200 more test tube babies—includingLouise Brown’s sister, Natalie—were born over the next five years. (Natalie is now amother, having conceived naturally, and is the first IVF baby to have a child.) Seeingso many healthy-looking test tube babies worldwide changed Vande Wiele’s opinion,a change that paralleled the transformation in feeling about IVF that was occurringin the public at large. When Columbia University opened the first IVF clinic in NewYork City in 1983, its co-director was Raymond Vande Wiele. —R.M.H.

COURTING JUSTICE: Doris Del-Zio and her attorney, Michael Dennis, outside U.S. district court inNew York City on July 17, 1978, after a session of jury selection. Del-Zio and her husband, John,sued physician Raymond Vande Wiele for derailing their early attempt at in vitro fertilization.


of the drugs used to induce ovulation orto maintain pregnancy in its early stages.In addition, factors contributing to infer-tility may increase the risk of birth defects.The technique of IVF itself also might beto blame. A flawed sperm injected into anegg, as it is in one IVF variation, may havebeen unable to penetrate the egg on itsown and is thus given a chance it wouldotherwise not have to produce a babywith a developmental abnormality.

Clearly, these risks could remain hid-den during more than two decades of ex-perience with IVF only because no systemwas ever put in place to track results. “Ifthe government had supported IVF, thefield would have made much more rapidprogress,” says Duane Alexander, direc-tor of the National Institute of ChildHealth and Human Development. “Butas it is, the institute has never funded hu-man IVF research of any form”—a recordthat Alexander calls both incredible andembarrassing.

Although the medical downsides ofIVF are finally coming to light, many ofthe more alarmist predictions about whereIVF would lead never came to pass. Forexample, one scenario was that it wouldbring us “wombs for hire,” an oppressedunderclass of women paid to bear thechildren of the infertile rich. But surrogatemotherhood turned out to be expensiveand emotionally complex for all parties,and it never became widespread.

Human cloning, too, might turn outto be less frightening than we currentlyimagine. Market forces might make re-productive cloning impractical, and sci-entific advancement might make it un-necessary. For example, people unable toproduce eggs or sperm might pondercloning to produce offspring. But thetechnology developed for cloning couldmake it possible to create artificial eggs orsperm containing the woman’s or man’sown DNA, which could then be com-bined with the sperm or egg of a partner.In the future, “cloning” might refer onlyto what is now being called therapeuticcloning, and it might eventually be trulytherapeutic: a laboratory technique formaking cells for the regeneration of dam-

aged organs, for example. And some ob-servers believe that the most common useof cloning technology will ultimately notinvolve human cells at all: the creaturemost likely to be cloned may wind up be-ing a favorite family dog or cat.

The history of IVF reveals the pitfallsfacing cloning if decision making is sim-ply avoided. But despite similarities in so-cietal reactions to IVF and cloning, thetwo technologies are philosophicallyquite different. The goal of IVF is to en-able sexual reproduction in order to pro-duce a genetically unique human being.

Only the site of conception changes, afterwhich events proceed much the way theynormally do. Cloning disregards sexualreproduction, its goal being to mimic notthe process but the already existing livingentity. Perhaps the biggest difference be-tween IVF and cloning, however, is thefocus of our anxieties. In the 1970s thegreatest fear related to in vitro fertiliza-tion was that it would fail, leading to sor-row, disappointment and possibly thebirth of grotesquely abnormal babies. To-day the greatest fear about human cloningis that it may succeed.


ADR

IAN

AR

BIB

Cor

bis

M O R E T O E X P L O R EMoving toward Clonal Man: Is This What We Want? James D. Watson in Atlantic Monthly, Vol. 227,No. 5, pages 50–53; May 1971.

The Frankenstein Myth Becomes a Reality: We Have the Awful Knowledge to Make Exact Copiesof Human Beings. Willard Gaylin in New York Times Magazine, pages 12–13, 41–46; March 5, 1972.

The Embryo Sweepstakes. David Rorvik in New York Times Magazine, pages 17, 50–62; September 15, 1974.

Remaking Eden: How Genetic Engineering and Cloning Will Transform the American Family. Lee M. Silver. Avon Books, 1998.

The Clone Age: Adventures in the New World of Reproductive Technology. Lori B. Andrews. Henry Holt and Company, 1999.

Free to Be Me: Would-Be Cloners Pushing the Debate. Rick Weiss in Washington Post, page A1; May 12, 2002.

THE DAY AFTER her 20th birthday, Louise Brownposes at home with her parents.


68 S C I E N T I F I C A M E R I C A N

PHYSICS BEYOND THE The Dawn of

By Gordon Kane


STANDARD MODEL

S C I E N T I F I C A M E R I C A N 69

The Standard Modelof particle physics isat a pivotal momentin its history: it isboth at the height ofits success and onthe verge of beingsurpassed

A NEW ERA IN PARTICLE PHYSICS could soon be heralded by the detectionof supersymmetric particles at the Tevatron collider at Fermi NationalAccelerator Laboratory in Batavia, Ill. A quark and an antiquark (red andblue) smashing head-on would form two heavy supersymmetric particles(pale magenta). Those would decay into W and Z particles (orange) andtwo lighter supersymmetric particles (dark magenta). The W and Z wouldin turn decay into an electron, an antielectron and a muon (all green),which would all be detected, and an invisible antineutrino (gray).

BR

YAN

CH

RIS

TIE

DE

SIG

N


Today, centuries after the search began for the funda-mental constituents that make up all the complexity and beau-ty of the everyday world, we have an astonishingly simple an-swer—it takes just six particles: the electron, the up and thedown quarks, the gluon, the photon and the Higgs boson.Eleven additional particles suffice to describe all the esoteric phe-nomena studied by particle physicists [see box at right]. This isnot speculation akin to the ancient Greeks’ four elements ofearth, air, water and fire. Rather it is a conclusion embodied inthe most sophisticated mathematical theory of nature in histo-ry, the Standard Model of particle physics. Despite the word“model” in its name, the Standard Model is a comprehensivetheory that identifies the basic particles and specifies how theyinteract. Everything that happens in our world (except for theeffects of gravity) results from Standard Model particles inter-acting according to its rules and equations.

The Standard Model was formulated in the 1970s and ten-tatively established by experiments in the early 1980s. Nearlythree decades of exacting experiments have tested and verifiedthe theory in meticulous detail, confirming all of its predictions.In one respect, this success is rewarding because it confirms thatwe really understand, at a deeper level than ever before, how na-ture works. Paradoxically, the success has also been frustrating.Before the advent of the Standard Model, physicists had becomeused to experiments producing unexpected new particles or oth-er signposts to a new theory almost before the chalk dust hadsettled on the old one. They have been waiting 30 years for thatto happen with the Standard Model.

Their wait should soon be over. Experiments that achieve col-lisions that are higher in energy than ever before or that study cer-


■ The Standard Model of particle physics is the mostsuccessful theory of nature in history, but increasinglythere are signs that it must be extended by adding newparticles that play roles in high-energy reactions.

■ Major experiments are on the verge of providing directevidence of these new particles. After 30 years ofconsolidation, particle physics is entering a new era ofdiscovery. Many profound mysteries could be resolved bypost–Standard Model physics.

■ One element of the Standard Model—a particle called theHiggs boson—also remains to be observed. The Tevatroncollider at Fermilab could detect Higgs bosons within thenext few years.

Overview/A New Era

The Particles ALTHOUGH THE STANDARD MODEL needs to be extended, its particlessuffice to describe the everyday world (except for gravity) and almostall data collected by particle physicists.

In the Standard Model, thefundamental particles of ordinary matter are the electron, the up quark(u) and the down quark (d). Triplets of quarks bind together to formprotons (uud) and neutrons (udd), which in turn make up atomic nuclei( above). The electron and the up and the down quarks, together withthe electron-neutrino, form the first of three groups of particles calledgenerations. Each generation is identical in every respect except for themasses of the particles ( grid at right). The values of the neutrinomasses in the chart are speculative but chosen to be consistent withobservations.

The Standard Model describes three ofthe four known forces: electromagnetism, the weak force (which isinvolved in the formation of the chemical elements) and the strong force(which holds protons, neutrons and nuclei together). The forces aremediated by force particles: photons for electromagnetism, the W and Zbosons for the weak force, and gluons for the strong force. For gravity,gravitons are postulated, but the Standard Model does not includegravity. The Standard Model partially unifies the electromagnetic andweak forces—they are facets of one “electroweak” force at high energiesor, equivalently, at distances smaller than the diameter of protons.

One of the greatest successes of the Standard Model is that theforms of the forces—the detailed structure of the equations describingthem—are largely determined by general principles embodied in thetheory rather than being chosen in an ad hoc fashion to match acollection of empirical data. For electromagnetism, for example, thevalidity of relativistic quantum field theory (on which the StandardModel is based) and the existence of the electron imply that the photonmust also exist and interact in the way that it does—we finallyunderstand light. Similar arguments predicted the existence andproperties, later confirmed, of gluons and the W and Z particles.

THE STANDARD MODEL

MATTER PARTICLES (FERMIONS)

FORCE CARRIERS (BOSONS)

Atom

Nucleus

Gluon

Up quark Down quark

Electron

Proton



In addition to the particles describedabove, the Standard Model predicts the existence of the Higgsboson, which has not yet been directly detected by experiment.The Higgs interacts with the other particles in a special mannerthat gives them mass.

Might the Standard Model be superseded by atheory in which quarks and electrons are made up of morefundamental particles? Almost certainly not. Experiments haveprobed much more deeply than ever before without finding a hintof additional structure. More important, the Standard Model is aconsistent theory that makes sense if electrons and quarks arefundamental. There are no loose ends hinting at a deeperunderlying structure. Further, all the forces become similar at highenergies, particularly if supersymmetry is true [see box on nextpage]. If electrons and quarks are composite, this unificationfails: the forces do not become equal. Relativistic quantum fieldtheory views electrons and quarks as being pointlike—they arestructureless. In the future, they might be thought of as tinystrings or membranes (as in string theory), but they will still beelectrons and quarks, with all the known Standard Modelproperties of these objects at low energies.

FERMIONS* BOSONS

FirstGeneration

SecondGeneration

Mas

s (g

iga-

elec

tron

-vol

ts)

ThirdGeneration

MASSLESS BOSONS

Photon

Electron

Muon-neutrino

Muon

Charm quark

Tau

Top quarkZ

W

Higgs

Down quark

Gluon

Electron-neutrino

Tau-neutrino

Bottom quark

Strange quark

Up quark

103

102

101

100

10–1

10–2

10–3

10–4

10–10

10–11

10–12

The Rules of the GameTHE STANDARD MODEL describes thefundamental particles and how they interact.For a full understanding of nature, we alsoneed to know what rules to use to calculatethe results of the interactions. An examplethat helps to elucidate this point is Newton’slaw, F = ma. F is any force, m is the mass ofany particle, and a is the acceleration of theparticle induced by the force. Even if you knowthe particles and the forces acting on them,you cannot calculate how the particlesbehave unless you also know the rule F = ma.The modern version of the rules is relativisticquantum field theory, which was invented inthe first half of the 20th century. In thesecond half of the 20th century thedevelopment of the Standard Model taughtresearchers about the nature of the particlesand forces that were playing by the rules ofquantum field theory. The classical concept ofa force is also extended by the StandardModel: in addition to pushing and pulling onone another, when particles interact they canchange their identity and be created ordestroyed.

Feynman diagrams (a–g, at right), firstdevised by physicist Richard P. Feynman,serve as useful shorthand to describeinteractions in quantum field theory. Thestraight lines represent the trajectories ofmatter particles; the wavy lines representthose of force particles. Electromagnetism isproduced by the emission or absorption ofphotons by any charged particle, such as anelectron or a quark. In a, the incoming electronemits a photon and travels off in a newdirection. The strong force involves gluonsemitted (b) or absorbed by quarks. The weakforce involves W and Z particles (c, d), which areemitted or absorbed by both quarks andleptons (electrons, muons, taus and neutrinos).Notice how the W causes the electron tochange identity. Gluons (e) and Ws and Zs ( f)also self-interact, but photons do not.

Diagrams a through f are calledinteraction vertices. Forces are produced bycombining two or more vertices. For example,the electromagnetic force between an electron and a quark islargely generated by the transfer of a photon (g). Everything thathappens in our world, except for gravity, is the result ofcombinations of these vertices. —G.K.

DEEPER LEVELS?

THE SOURCE OF MASS

Electron

Electron

Electron

Quark

Quark

W boson

W boson

Z boson

W boson

Z boson

Electron-neutrino

Neutrino

Photon

Photon

Gluon

Gluon

a

b

c

d

e

f

g

* T h e f e r m i o n s a r e s u b d i v i d e d i n t o q u a r k s a n d l e p t o n s , w i t h l e p t o n si n c l u d i n g e l e c t r o n s , m u o n s , t a u s a n d t h r e e f o r m s o f n e u t r i n o .

BR

YAN

CH

RIS

TIE

DE

SIG

N


tain key phenomena with greater precision are on the verge ofgoing beyond the Standard Model. These results will not over-turn the Standard Model. Instead they will extend it by uncov-ering particles and forces not described by it. The most impor-tant experiment is occurring at the upgraded Tevatron colliderat Fermi National Accelerator Laboratory in Batavia, Ill., whichbegan taking data in 2001. It might produce directly the still elu-sive particles that complete the Standard Model (Higgs bosons)and those predicted by the most compelling extensions of the the-ory (the so-called superpartners of the known particles).

Significant information is also beginning to come from “Bfactories,” particle colliders running in California and Japanconfigured to create billions of b quarks (one of the 11 additionalparticles) and their antimatter equivalents to study a phen-omenon called CP violation. CP (charge-parity) is the symmetryrelating matter to antimatter, and CP violation means that anti-matter does not exactly mirror matter in its behavior. The amountof CP violation observed so far in particle decays can be accom-modated by the Standard Model, but we have reasons to expectmuch more CP violation than it can produce. Physics that goesbeyond the Standard Model can generate additional CP violation.

Physicists are also studying the precise electric and magnet-ic properties of particles. The Standard Model predicts that elec-trons and quarks behave as microscopic magnets with a specif-ic strength and that their behavior in an electric field is deter-mined purely by their electric charge. Most extensions of theStandard Model predict a slightly different magnetic strengthand electrical behavior. Experiments are beginning to collectdata with enough sensitivity to see the tiny effects predicted.

Looking beyond the earth, scientists studying solar neutri-nos and cosmic-ray neutrinos, ghostly particles that barely inter-act at all, have recently established that neutrinos have masses,a result long expected by theorists studying extensions of the Stan-dard Model [see “Solving the Solar Neutrino Problem,” by ArthurB. McDonald, Joshua R. Klein and David L. Wark; ScientificAmerican, April]. The next round of experiments will clarify theform of theory needed to explain the observed neutrino masses.

In addition, experiments are under way to detect mysteriousparticles that form the cold dark matter of the universe and toexamine protons at higher levels of sensitivity to learn whetherthey decay. Success in either project would be a landmark ofpost–Standard Model physics.

As all this research proceeds, it is ushering in a new, data-richera in particle physics. Joining the fray by about 2007 will be theLarge Hadron Collider (LHC), a machine 27 kilometers in cir-cumference now under construction at CERN, the European lab-oratory for particle physics near Geneva [see “The Large HadronCollider,” by Chris Llewellyn Smith; Scientific American,July 2000]. A 30-kilometer-long linear electron-positron collid-er that will complement the LHC’s results is in the design stages.

As the first hints of post–Standard Model physics areglimpsed, news reports often make it sound as if the StandardModel has been found to be wrong, as if it were broken andready to be discarded, but that is not the right way to thinkabout it. Take the example of Maxwell’s equations, written


NIN

A FI

NK

EL

Strong Force

Weak Force

Electromagnetic Force

Forc

e St

reng

th

102 104 106 108 10101012 1014

1016

Interaction Energy (GeV)

STANDARD MODEL

Strong Force

Weak Force

Electromagnetic Force

Forc

e St

reng

th

102 104 106 108 10101012 1014

1016

MINIMAL SUPERSYMMETRIC STANDARD MODEL

Interaction Energy (GeV)

GORDON KANE, a particle theorist, is Victor Weisskopf CollegiateProfessor of Physics at the University of Michigan at Ann Arbor. Hiswork explores ways to test and extend the Standard Model of par-ticle physics. In particular he studies Higgs physics and the Stan-dard Model’s supersymmetric extension, with a focus on relatingtheory and experiment and on the implications of supersymme-try for particle physics and cosmology. His hobbies include play-ing squash, exploring the history of ideas, and seeking to under-stand why science flourishes in some cultures but not others.

THE

AU

THO

REvidence for Supersymmetry

THE MOST WIDELY FAVORED THEORY to supersede the StandardModel is the Minimal Supersymmetric Standard Model. In thismodel, every known particle species has a superpartner particlethat is related to it by supersymmetry. Particles come in twobroad classes: bosons (such as the force particles), which cangather en masse in a single state, and fermions (such asquarks and leptons), which avoid having identical states. Thesuperpartner of a fermion is always a boson and vice versa.

Indirect evidence for supersymmetry comes from theextrapolation of interactions to high energies. In the StandardModel, the three forces become similar but not equal in strength(top). The existence of superpartners changes the extrapolationso that the forces all coincide at one energy (bottom)—a cluethat they become unified if supersymmetry is true.


down in the late 19th century to describe the electromagneticforce. In the early 20th century we learned that at atomic sizesa quantum version of Maxwell’s equations is needed. Later theStandard Model included these quantum Maxwell’s equationsas a subset of its equations. In neither case do we say Maxwell’sequations are wrong. They are extended. (And they are still usedto design innumerable electronic technologies.)

A Permanent EdificeS IMILARLY, THE STANDARD MODEL is here to stay. It isa full mathematical theory—a multiply connected and highlystable edifice. It will turn out to be one piece of a larger such ed-ifice, but it cannot be “wrong.” No part of the theory can failwithout a collapse of the entire structure. If the theory werewrong, many successful tests would be accidents. It will contin-ue to describe strong, weak and electromagnetic interactions atlow energies.

The Standard Model is very well tested. It predicted the ex-istence of the W and Z bosons, the gluon and two of the heav-ier quarks (the charm and the top quark). All these particles weresubsequently found, with precisely the predicted properties.

A second major test involves the electroweak mixing angle,a parameter that plays a role in describing the weak and elec-tromagnetic interactions. That mixing angle must have the samevalue for every electroweak process. If the Standard Model werewrong, the mixing angle could have one value for one process,a different value for another and so on. It is observed to have thesame value everywhere, to an accuracy of about 1 percent.

Third, the Large Electron-Positron (LEP) collider at CERN,which ran from 1989 to 2000, looked at about 20 million Zbosons. Essentially every one of them decayed in the manner ex-pected by the Standard Model, which predicted the number ofinstances of each kind of decay as well as details of the energiesand directions of the outgoing particles. These tests are but a fewof the many that have solidly confirmed the Standard Model.

In its full glory, the Standard Model has 17 particles andabout as many free parameters—quantities such as particle mass-es and strengths of interactions [see box on pages 70 and 71].These quantities can in principle take any value, and we learn thecorrect values only by making measurements. Armchair crit-ics sometimes compare the Standard Model’s many parameterswith the epicycles on epicycles that medieval theorists used todescribe planetary orbits. They imagine that the Standard Mod-el has limited predictive power, or that its content is arbitrary,or that it can explain anything by adjusting of some parameter.

The opposite is actually true: once the masses and interac-tion strengths are measured in any process, they are fixed forthe whole theory and for any other experiment, leaving no free-dom at all. Moreover, the detailed forms of all the StandardModel’s equations are determined by the theory. Every parameter but the Higgs boson mass has been measured. Untilwe go beyond the Standard Model, the only thing that canchange with new results is the precision of our knowledge of theparameters, and as that improves it becomes harder, not easier,for all the experimental data to remain consistent, because

measured quantities must agree to higher levels of precision. Adding further particles and interactions to extend the Stan-

dard Model might seem to introduce a lot more freedom, butthis is not necessarily the case. The most widely favored exten-sion is the Minimal Supersymmetric Standard Model (MSSM).Supersymmetry assigns a superpartner particle to every parti-cle species. We know little about the masses of those superpart-ners, but their interactions are constrained by the supersymme-try. Once the masses are measured, the predictions of the MSSMwill be even more tightly constrained than the Standard Modelbecause of the mathematical relations of supersymmetry.

Ten MysteriesIF THE STANDARD MODEL works so well, why must it beextended? A big hint arises when we pursue the long-standinggoal of unifying the forces of nature. In the Standard Model, wecan extrapolate the forces and ask how they would behave atmuch higher energies. For example, what were the forces like inthe extremely high temperatures extant soon after the big bang?At low energies the strong force is about 30 times as powerfulas the weak force and more than 100 times as powerful as elec-tromagnetism. When we extrapolate, we find that the strengthsof these three forces become very similar but are never all exact-ly the same. If we extend the Standard Model to the MSSM, theforces become essentially identical at a specific high energy [seebox on opposite page]. Even better, the gravitational force ap-proaches the same strength at a slightly higher energy, suggest-ing a connection between the Standard Model forces and grav-ity. These results seem like strong clues in favor of the MSSM.

Other reasons for extending the Standard Model arise fromphenomena it cannot explain or cannot even accommodate:

1. All our theories today seem to imply that the universeshould contain a tremendous concentration of energy, evenin the emptiest regions of space. The gravitational effects ofthis so-called vacuum energy would have either quicklycurled up the universe long ago or expanded it to muchgreater size. The Standard Model cannot help us understandthis puzzle, called the cosmological constant problem.

2. The expansion of the universe was long believed to beslowing down because of the mutual gravitational attractionof all the matter in the universe. We now know that the ex-pansion is accelerating and that whatever causes the accel-eration (dubbed “dark energy”) cannot be Standard Modelphysics.

3. There is very good evidence that in the first fraction of asecond of the big bang the universe went through a stage ofextremely rapid expansion called inflation. The fields re-sponsible for inflation cannot be Standard Model ones.

4. If the universe began in the big bang as a huge burst of en-ergy, it should have evolved into equal parts matter and an-timatter (CP symmetry). But instead the stars and nebulae

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 73COPYRIGHT 2003 SCIENTIFIC AMERICAN, INC.

are made of protons, neutrons and electrons and not their an-tiparticles (their antimatter equivalents). This matter asym-metry cannot be explained by the Standard Model.

5. About a quarter of the universe is invisible cold dark mat-ter that cannot be particles of the Standard Model.

6. In the Standard Model, interactions with the Higgs field(which is associated with the Higgs boson) cause particlesto have mass. The Standard Model cannot explain the veryspecial forms that the Higgs interactions must take.

7. Quantum corrections apparently make the calculatedHiggs boson mass huge, which in turn would make all parti-cle masses huge. That result cannot be avoided in the Stan-dard Model and thus causes a serious conceptual problem.

8. The Standard Model cannot include gravity, because itdoes not have the same structure as the other three forces.

9. The values of the masses of the quarks and leptons (suchas the electron and neutrinos) cannot be explained by theStandard Model.

10. The Standard Model has three “generations” of particles.The everyday world is made up entirely of first-generationparticles, and that generation appears to form a consistenttheory on its own. The Standard Model describes all three gen-erations, but it cannot explain why more than one exists.

In expressing these mysteries, when I say the Standard Mod-el cannot explain a given phenomenon, I do not mean that thetheory has not yet explained it but might do so one day. TheStandard Model is a highly constrained theory, and it cannotever explain the phenomena listed above. Possible explanationsdo exist. One reason the supersymmetric extension is attractiveto many physicists is that it can address all but the second andthe last three of these mysteries. String theory (in which parti-cles are represented by tiny, one-dimensional entities instead ofpoint objects) addresses the last three [see “The Theory FormerlyKnown as Strings,” by Michael J. Duff; Scientific American,February 1998]. The phenomena that the Standard Model can-not explain are clues to how it will be extended.

It is not surprising that there are questions that the StandardModel cannot answer—every successful theory in science has in-creased the number of answered questions but has left some unan-swered. And even though improved understanding has led tonew questions that could not be formulated earlier, the numberof unanswered fundamental questions has continued to decrease.

Some of these 10 mysteries demonstrate another reason whyparticle physics today is entering a new era. It has become clearthat many of the deepest problems in cosmology have their so-lutions in particle physics, so the fields have merged into “par-ticle cosmology.” Only from cosmological studies could welearn that the universe is matter (and not antimatter) or that the

universe is about a quarter cold dark matter. Any theoretical un-derstanding of these phenomena must explain how they arise aspart of the evolution of the universe after the big bang. But cos-mology alone cannot tell us what particles make up cold darkmatter, or how the matter asymmetry is actually generated, orhow inflation originates. Understanding of the largest and thesmallest phenomena must come together.

The HiggsPHYSICISTS ARE TACKLING all these post–Standard Mod-el mysteries, but one essential aspect of the Standard Model alsoremains to be completed. To give mass to leptons, quarks, andW and Z bosons, the theory relies on the Higgs field, which hasnot yet been directly detected.

The Higgs is fundamentally unlike any other field. To un-derstand how it is different, consider the electromagnetic field.Electric charges give rise to electromagnetic fields such as thoseall around us (just turn on a radio to sense them). Electromag-netic fields carry energy. A region of space has its lowest possi-ble energy when the electromagnetic field vanishes throughoutit. Zero field is the natural state in the absence of charged parti-cles. Surprisingly, the Standard Model requires that the lowestenergy occur when the Higgs field has a specific nonzero value.Consequently, a nonzero Higgs field permeates the universe, andparticles always interact with this field, traveling through it likepeople wading through water. The interaction gives them theirmass, their inertia.

Associated with the Higgs field is the Higgs boson. In theStandard Model, we cannot predict any particle masses fromfirst principles, including the mass of the Higgs boson itself.One can, however, use other measured quantities to calculatesome masses, such as those of the W and Z bosons and the topquark. Those predictions are confirmed, giving assurance to theunderlying Higgs physics.

Physicists do already know something about the Higgs mass.Experimenters at the LEP collider measured about 20 quantitiesthat are related to one another by the Standard Model. All theparameters needed to calculate predictions for those quantitiesare already measured—except for the Higgs boson mass. One cantherefore work backward from the data and ask which Higgsmass gives the best fit to the 20 quantities. The answer is that theHiggs mass is less than about 200 giga-electron-volts (GeV). (Theproton mass is about 0.9 GeV; the top quark 174 GeV.) Thatthere is an answer at all is strong evidence that the Higgs exists.If the Higgs did not exist and the Standard Model were wrong,it would take a remarkable coincidence for the 20 quantities tobe related in the right way to be consistent with a specific Higgsmass. Our confidence in this procedure is bolstered because asimilar approach accurately predicted the top quark mass be-fore any top quarks had been detected directly.

LEP also conducted a direct search for Higgs particles, butit could search only up to a mass of about 115 GeV. At that veryupper limit of LEP’s reach, a small number of events involvedparticles that behaved as Higgs bosons should. But there werenot enough data to be sure a Higgs boson was actually discov-

74 S C I E N T I F I C A M E R I C A N J U N E 2 0 0 3COPYRIGHT 2003 SCIENTIFIC AMERICAN, INC.

ered. Together the results suggest the Higgs mass lies between115 and 200 GeV.

LEP is now dismantled to make way for the construction ofthe LHC, which is scheduled to begin taking data in four years.In the meantime the search for the Higgs continues at the Teva-tron at Fermilab [see illustration above]. If the Tevatron oper-ates at its design intensity and energy and does not lose runningtime because of technical or funding difficulties, it could confirmthe 115-GeV Higgs boson in about two to three years. If theHiggs is heavier, it will take longer for a clear signal to emergefrom the background. The Tevatron will produce more than10,000 Higgs bosons altogether if it runs as planned, and itcould test whether the Higgs boson behaves as predicted. TheLHC will be a “factory” for Higgs bosons, producing millionsof them and allowing extensive studies.

There are also good arguments that some of the lighter su-perpartner particles predicted by the MSSM have masses smallenough so that they could be produced at the Tevatron as well.Direct confirmation of supersymmetry could come in the nextfew years. The lightest superpartner is a prime candidate tomake up the cold dark matter of the universe—it could be di-rectly observed for the first time by the Tevatron. The LHC willproduce large numbers of superpartners if they exist, definitive-ly testing whether supersymmetry is part of nature.

Effective TheoriesTO FULLY GRASP the relation of the Standard Model to therest of physics, and its strengths and limitations, it is useful tothink in terms of effective theories. An effective theory is a de-scription of an aspect of nature that has inputs that are, in prin-ciple at least, calculable using a deeper theory. For example, innuclear physics one takes the mass, charge and spin of the pro-

ton as inputs. In the Standard Model, one can calculate thosequantities, using properties of quarks and gluons as inputs. Nu-clear physics is an effective theory of nuclei, whereas the Stan-dard Model is the effective theory of quarks and gluons.

From this point of view, every effective theory is open-end-ed and equally fundamental—that is, not truly fundamental atall. Will the ladder of effective theories continue? The MSSMsolves a number of problems the Standard Model does not solve,but it is also an effective theory because it has inputs as well. Itsinputs might be calculable in string theory.

Even from the perspective of effective theories, particle phys-ics may have special status. Particle physics might increase ourunderstanding of nature to the point where the theory can beformulated with no inputs. String theory or one of its cousinsmight allow the calculation of all inputs—not only the electronmass and such quantities but also the existence of spacetime andthe rules of quantum theory. But we are still an effective theoryor two away from achieving that goal.


FER

MIL

AB P

HO

TO

The Particle Garden. Gordon Kane. Perseus Publishing, 1996.

The Rise of the Standard Model: A History of Particle Physics from1964 to 1979. Edited by Lillian Hoddeson, Laurie M. Brown, MichaelRiordan and Max Dresden. Cambridge University Press, 1997.

The Little Book of the Big Bang: A Cosmic Primer. Craig J. Hogan.Copernicus Books, 1998.

Supersymmetry: Unveiling the Ultimate Laws of Nature. Gordon Kane.Perseus Publishing, 2001.

An excellent collection of particle physics Web sites is listed atparticleadventure.org/particleadventure/other/othersites.html


UPGRADING of the Tevatron’s huge particle detectors, which was carried out by physicists at Fermilab from 1996 to 2000, has primed the facility forobserving Higgs bosons and supersymmetry.


A study of chain letters shows how to infer the family tree of anything that evolves over time,

from biological genomes to languages to plagiarized schoolwork

BY CHARLES H. BENNETT, MING LI AND BIN MA

&76 S C I E N T I F I C A M E R I C A N J U N E 2 0 0 3

ChainEvolutionary Histories

Letters



SAR

A C

HE

N

collected between 1980 and 1995, when photocopiers, but note-mail, were in widespread use by the general public. These let-ters have passed from host to host, mutating and evolving. Likea gene, their average length is about 2,000 characters. Like a po-tent virus, the letter threatens to kill you and induces you to passit on to your “friends and associates”—some variation of thisletter has probably reached millions of people. Like an inheri-table trait, it promises benefits for you and the people you passit on to. Like genomes, chain letters undergo natural selectionand sometimes parts even get transferred between coexisting“species.” Unlike DNA, however, these letters are easy to read.Indeed, their readability makes them especially suitable for class-room teaching of phylogeny (evolutionary history) free from thearcana of molecular biology.

The letters are an intriguing social phenomenon, but we arealso interested in them because they provide a test bed for the al-gorithms used in molecular biology to infer phylogenetic treesfrom the genomes of existing organisms. We believe that if thesealgorithms are to be trusted, they should produce good resultswhen applied to chain letters. Using a new algorithm that is gen-eral enough to have wide applicability to such problems, wehave reconstructed the evolutionary history of our 33 letters [seeillustration on page 79]. The standard methods do not work aswell on these letters. Originally developed for genomes, our al-gorithm has also been applied to languages and used to detectplagiarism in student assignments: anything involving a se-quence of symbols is grist for its mill.

A Mind VirusTHE 33 CHAIN LETTERS are a fascinating collection. We la-beled them arbitrarily from L1 to L33. The letters differ signif-icantly. There are 15 titles, 23 names for “an office employee,”and 25 names for the original author of the letter. Misspellings,swapped sentences, and missing or added phrases, sentences andparagraphs are common. (A typical letter is shown on the nextpage, along with some of the many variations.) Nearly all aremore or less faded photocopies of typescripts, leading us to sur-mise that the mutations arose by an intermittent process, where-by a letter would be photocopied for several generations untilits legibility was so reduced that the next recipient decided to re-type it, introducing new errors and variations.

All but three of the letters we received were unique; for L4,L6 and L22, a second copy arrived within a few months of thefirst. Besides the 33 English-language letters, we received (butdid not include in our study) four in French and one each inDutch and German, all clearly sharing a common ancestry withthe ones in English.

To analyze the letters, we retyped them into computer filesentirely in lowercase, ignoring extra information such as datesand marginal notes as well as the division of the text into linesand paragraphs. Each letter became one continuous string ofcharacters.

Before we applied our new algorithm, we tried analyzing theletters with a procedure called multiple alignment, which iswidely used for examining genes to infer phylogeny. This meth-od attempts to line up as many matching sections of all the let-ters as possible. The amount of matching between any pair of let-ters defines their similarity, and from that data another algorithmconstructs an evolutionary tree. Unfortunately, multiple align-ment only finds matches with everything remaining in the sameorder, so it gets confused by L12 and L26, in which the orderof sentences has been rearranged. For the same reason, the tech-nique is known to work better within individual genes than forwhole genomes, in which such translocations are more common.

We tried omitting L12 and L26 and then performing multi-ple alignment on the remaining 31 letters. Even with this trun-cated set, the resulting tree seemed wrong, classifying L6, L7 andL13 as closely related. This error occurred because those threeletters are all relatively short, giving them a correspondinglysmall number of differences. This problem can arise in geneticsas well: merely counting differences can overestimate the simi-larity of short genomes while underestimating that of long ones.A proper measure should give more weight to a small differencein a small genome than to a small difference in a big genome.

We turned to devising our own similarity measure, one thatcould be applied to genomes, chain letters or any other type ofdata that might be stored as a computer file. We wanted to makeour new similarity measure insensitive to minor mix-ups such astranslocations, which represent only a small loss of informa-tional similarity. To cope with differences in length, we wantedour measure to assign two completely dissimilar data files a scoreof 0 and two identical ones a score of 1, regardless of their sizes.

The natural measure of the information content in a data fileis not its raw size in bits but rather the smallest size it can be com-pressed to by a file compression program such as zip or StuffIt.These programs are designed to save space on hard disks by find-ing and squeezing out the most common kinds of redundancy(for instance, repeated phrases), resulting in a smaller file fromwhich the original can be perfectly reconstructed when needed.

Something interesting happens if we compress two files to-gether so that both can be regenerated from the compressed file.If the two files share no information at all, the joint compressedfile will be as big as the two individual compressed files com-bined. But if the two files contain some of the same information,

IN OUR HANDSARE 33 VERSIONS OF

A CHAIN LETTER,


that repetition will be detected by a goodcompression program, and the joint com-pressed file will be smaller. In this way,the size of the joint compressed file com-pared with the sum of the individual com-pressed files provides a measure of thefiles’ similarity.

That measure is not yet a good one forour purposes, because two large files willtend to have greater similarity than twosmall files. To correct this problem, wedefine our “relatedness” measure to bethe proportion of shared information—

that is, the percentage by which the sum

of the separately compressed files exceedsthe size of the jointly compressed file. Thismakes the relatedness range from 0 forunrelated files to 1 (or 100 percent) foridentical files, regardless of length.

Which compression program shouldwe use? Obviously, our relatedness mea-sure will depend on that. Ideally, wewould want to use a program that com-presses every file to the smallest possiblesize. The study of information measuresdefined in terms of such ultimate com-pressibility forms an elegant branch of in-formation theory known as algorithmic

information theory or Kolmogorov com-plexity (after the late Russian mathemati-cian Andrei N. Kolmogorov, one of itsfounders). Unfortunately, informationtheorists have proved that such an idealzip program would take essentially an in-finitely long time to perform its task. Forour purposes, then, we decided to use aparticular compression algorithm calledGenCompress, created by Xin Chen of theUniversity of California at Santa Barbara.GenCompress was designed for genomesand works well on them. As we shall see,it also works well on chain letters.


THE CHAIN LETTERS ARE ESPECIALLY SUITABLE FOR CLASSROOM TEACHING

OF PHYLOGENY.

THEME AND VARIATIONSA SAMPLE CHAIN LETTER, labeled L11, illustrates some of the ways that related letters changed when they were retyped(presumably after photocopies became illegible). The greatest variations occurred with unfamiliar names and quantities ofmoney—errors in those elements are easily overlooked because they do not change the meaning of the letter.

Trust in the Lord with all your heart and he will light the way“And all things whatever ye shall ask in prayer, believing, ye shall receive.” (Matthew 21:22)With love all things are possibleKiss someone you love when you get this letter and make magic

the Netherlandsthe Netherland.

Good luck

Babbit BrandtBrentCraduitCradutDabbittDaddiDaddianDaddin DadditDadditoDadiott Daditt DavittDepot DoddsRaditt

AnalaAndyAriaArlaCarioCarlCarlaCarleCarol CharlesGorco

For no reason should this chain be broken.For no reason should this letter be broken.

$1,755$7,755$7,775$75,000$115,000$775,000$7,750,000$7,770,000$7,775,000

6life

WalchWales WalshWelschWelsh

his

GemGen.GeneralGeorge

$7,000,$70,000,$470,000

Air ForceA.F.A.R.F.A.R.P.R.A.RAFU.S.U.S.A.F.


How the Letters EvolvedTHE EVOLUTIONARY TREE of the chain letters deduced by theauthors’ automated “relatedness” measure shows a number ofinteresting features. (Some other features are discussed in themain text.) At point A, a phrase, “provided you in turn send it backout,” was adopted. At point B, a new title evolved: “And all thingswhatever ye shall ask in prayer, believing, ye shall receive.” At pointC, the title further mutated to “With love all things are possible.”Also at point C, “The Netherlands” changed to “New England,” and“General Welch” became “Gene Welch.” The sentence “For no reasonshould it be broken” got lost.

At point F, the title mutated to “Kiss someone you love whenyou get this letter and make magic.” After we had finished ouranalysis, we found a very interesting and comprehensive study ofmore than 460 chain letters of many varieties by mathematicianDaniel W. VanArsdale. His study raised the question, Which titlecame first: the “Kiss” or the “Love” one? We have not yet appliedour algorithm to the many chain letters in his collection, but judgingfrom our phylogeny, the “Love” title came first. This conclusion isfurther supported by the amount of money Gene Welch received: inall the “Kiss” letters (except for the mutation at G), he received$7,755, whereas in the “Love” group he got $7,755,000. In thegroup before C, the figure was $775,000. The mutation sequence$775,000 → $7,755,000 → $7,755 is clearly more parsimoniousthan $775,000 → $7,755 → $7,755,000.

At point D, $70,000 grew to $470,000, and “St. Jude” was

adopted. Neither $470,000 nor St. Jude appears outside of thisgroup. At point E, two interesting alterations happenedconcurrently: the California woman’s car story was added and thetime between Gene Welch’s receiving the letter and losing his wifechanges from six days to 51 days. This is consistent across allletters in the group rooted at E, except for L28, which requires aspecial explanation. The “car and 51” mutation doesn’t appearanywhere outside of the E group.

L28 provides evidence of horizontal transfer—that is, thetransfer of information from one “organism” to another in additionto simple inheritance. In the group rooted at D, every letter haseach R.A.F. officer receive $470,000 or $470, except for L28, inwhich the amount is $70,000. L28 also has Gen. Welch, whootherwise features only in letters before point C. All the letters inthe D group except L21 have the car story, and all but L21 and L28have the “51 days” mutation. It seems unbelievable to assume thatL28 gained “$70,000” and “Gen. Welch” by mutation independentlyof the other instances of these mutations elsewhere in the tree.One might try placing the “car and 51” mutation before the“$470,000 and St. Jude” mutation, but then L21 must undergo avery implausible genesis: either it must lose the whole car storyand mutate “51 days” back to “six days,” or it must gain “$470,000”and “St. Jude” independently. Apparently somebody had two lettersin his or her possession while composing L28 (or L21) and introduceda foreign gene from a letter before C. —C.H.B., M.L. and B.M.

1

1533

1426

17

29

34

2112

1127

2416

528

3113

3022

28

20

32

23

9

19

7

10

6

1825

A

B

C

DE

FG

The phrase “provided you in

turn send it back out” is added

“St. Jude” and “$470,000”

“Kiss” title appears

“And all things” title change

“Love” title

“car” and “51 days”mutation

“$50,000 lottery”

SAR

A C

HE

N


THE AUTHORS’ ANALYSIS of the similaritiesamong 33 randomly numbered chainletters produced this tree that tracks how they evolved over time. A few of themore significant mutations, or changes, are indicated at the points in the tree that they occurred.


Given a set of chain letters, it is astraightforward and entirely automaticprocess to calculate the relatedness of eachpair using the GenCompress program.The next step, converting the relatednessdata into an evolutionary tree, is alsolargely automatic (many software pack-ages exist for this purpose). The result caneither be a simple tree diagram with arbi-trary branch lengths of the kind shown onthe preceding page, indicating simply thequalitative pattern of descent, or a moredetailed diagram, with branch lengthsthat represent relatedness quantitatively.

In either case, the main human inputis deciding where to put the root of thetree, which represents the hypothesizedcommon ancestor of all the letters (orspecies). In biological phylogenies, theroot stands for an extinct species frommillions of years ago, so it should not betoo closely related to any of the branch-es denoting organisms alive today. In ourstudy, the chain letters were collectedover a 15-year period, and some of themwere dated near the beginning of thattime, so we chose to put the root near oneof these (L15). Unfortunately, most ofthe letters were collected without record-ing their postmark or date of receipt, re-flecting the fact that this project began asa hobby and only belatedly evolved intoscientific research.

St. Jude’s PhylogenyTHE EVOLUTIONARY TREE that wasinferred for the chain letters appears to bealmost a “perfect” phylogeny, in thesense that documents that share the samecharacteristic are always grouped to-gether. After the tree was built, we wereable to use it to help make numerous hy-

potheses about how the letters evolved.First we judge that the letters before

point C in our phylogeny are the oldest[see illustration on preceding page]. Thechief evidence is that the name “CarloDadditt” and the title of the letter had themost mutations in this group of letters.We expect such errors to be more com-mon in the oldest letters because photo-copiers were less available at that timeand retyping more frequent. In addition,among the 14 dated letters, the two thatoccur in the pre-C group (L4 and L15) arethe oldest. These older letters are all titledwith religious prayers, come from “TheNetherlands” and contain the sentence“For no reason should it be broken.”

Next we see an effect familiar frommolecular biology, in which differentparts of the genome have quite differentmutation rates. Active sites of enzymesmutate scarcely at all, whereas parts farfrom the active site continually undergorandom drift. Similarly, with chain lettersthe parts required for “viability” do notmutate at all, but more arbitrary parts,such as the types of mishaps that wouldbefall those who did not propagate thechain, mutate more. Parts with little in-trinsic meaning to help catch errors—forinstance, unfamiliar names like “GemWalsh” and “Carlo Craduit”—mutatethe most.

Another biological phenomenon ap-pearing in the chain letters is the occur-rence of parallel, compensating muta-tions: two mutations that would be detri-mental individually and that must occurtogether to be neutral or beneficial. Ex-cluding L12 and L26, in which no onedies, all letters before point C (exceptnearby L29) read:

General Welsh (or a variation) lost his life . . . however before his death . . .

On the other hand, letters after pointC read:

Gene Welch (or a variation) lost his wife . . . however before herdeath . . .

To preserve the sense, “his” mutates to“her” when “life” mutates to “wife.” [Seebox on preceding page for more of theseobservations.]

Mammals and PlagiarismBESIDES ANALYZING chain letters, ourrelatedness measure has been used in awide range of settings. In bioinformaticsitself, we used it to analyze the mitochon-drial genomes of 18 mammals. Mito-chondria are energy-producing organelleswithin cells whose genes are inheritedsolely from the mother (similar to how achain letter inherits from a single “par-ent”). Because no reshuffling of maternaland paternal genes occurs, the accumula-tion of mutations in the mitochondrial ge-nome acts as a clock measuring when anorganism’s ancestors diverged from relat-ed species.

Traditional methods applied to dif-ferent mitochondrial genes often give con-flicting evolutionary trees, and many, un-like our new measure, cannot be appliedsuccessfully to an entire genome becauseof problems such as translocations. Forexample, using the traditional methods,about half a dozen mitochondrial genesimply that primates, such as ourselves, aremore closely related to rodents than toferungulates, a diverse group that in-cludes cows, horses, whales, cats anddogs. Another half a dozen genes implythat primates and ferungulates are themore closely related pair, which is gener-ally believed to be correct based on othermultiple lines of evidence, such as non-mitochondrial genes and the fossil record.

THE ACCUMULATIONOF MUTATIONS IN A MITOCHONDRIAL GENOME (OR A CHAIN LETTER)

ACTS AS A CLOCK.

CHARLES H. BENNETT, MING LI and BIN MA’s joint work on chain letters began as a resultof a hike in the Hong Kong mountains during which Bennett mentioned his collection to Li.Bennett is an IBM Fellow at the company’s research division in Yorktown Heights, N.Y. Hefocuses on the physics of information processing and is a co-discoverer of quantum tele-portation. Li studies bioinformatics and Kolmogorov complexity as professor of computerscience at the University of Waterloo in Ontario. Ma works on bioinformatics and algorithmdesign as assistant professor of computer science at the University of Western Ontario.

THE

AU

THO

RS

80 S C I E N T I F I C A M E R I C A N J U N E 2 0 0 3COPYRIGHT 2003 SCIENTIFIC AMERICAN, INC.

When our method is applied to whole mi-tochondrial genomes, it produces this lat-ter evolutionary tree without needing anyad hoc tinkering to resolve ambiguities orcontradictions [see illustration above].

Taking the art of phylogenetic infer-ence to an audacious extreme, DarioBenedetto, Emanuele Caglioti and Vitto-rio Loreto of La Sapienza University inRome tried inferring a phylogeny of hu-man languages not by analyzing the lan-guages’ known literatures or history butmerely by applying a method similar toours to 52 translations of the UniversalDeclaration of Human Rights. The resultwas surprisingly good, considering thetiny body of evidence on which it wasbased. One notable mistake was the clas-sification of English as a romance lan-guage, closely related to French, whereashistorically English evolved within theGermanic group. This error arises becauseof the great many French words that Eng-lish acquired after the Norman Conquest(an example of parallel transfer).

Another application of our measurehas been the detection of plagiarism in

students’ homework assignments. In oneinstance, two assignments in a comput-er programming class were flagged as be-ing unusually alike, but the instructorcould not see any obvious evidence ofcopying when he examined them him-self. The two students were approachedand, in the interest of research, given im-munity to plagiarism charges in returnfor an honest account of whether theyhad collaborated. Apparently the twostudents had discussed the problem and

how they planned to tackle it but hadnot worked together beyond that level.If that is what happened, our distance al-gorithm detected the subtle similaritiesengendered by their discussion!

The automatic nature of our proce-dure is both an advantage and a disad-vantage. On the one hand, it yields objec-tive answers, free from the need to weighvarious lines of evidence (such as DNAversus the fossil record) or to take accountof which parts of the genome mutatefastest. On the other hand, it does notbenefit from the insights that might comefrom such additional data. All methods ofphylogenetic inference are imperfect andwill sometimes mistakenly infer a phy-logeny that differs from what actuallytook place historically. Like historiansand paleontologists, evolutionary molec-ular biologists have come to accept thatno matter how many lines of evidencethey consider, the full truth about the pastcan never be reconstructed. This is espe-cially true with regard to extinct species.Many species that once existed will nev-er be known, because they left neither fos-sils nor descendants. Likewise many lan-guages have become extinct, vanishingwithout a trace even in the past century.

In the realm of chain letters, it is cer-tain that many letters became extinctwhen too many recipients broke the chain.Like the lost plays of Sophocles, these let-ters’ texts may never be recovered, andeven their existence can only be surmisedfrom circumstantial evidence, such as arash of unemployment and expensive carrepairs occurring in California for no ap-parent reason.


SAR

A C

HE

N

M O R E T O E X P L O R EAn Introduction to Kolmogorov Complexity and Its Applications. Second edition. Ming Li and Paul M. Vitányi. Springer-Verlag, 1997.

An Information-Based Sequence Distance and Its Application to Whole Mitochondrial GenomePhylogeny. Ming Li, Xin Chen, Jonathan H. Badger, Sam Kwong, Paul Kearney and Haoyong Zhang in Bioinformatics, Vol. 17, No. 2, pages 149–154; February 2001.

Language Trees and Zipping. Dario Benedetto, Emanuele Caglioti and Vittorio Loreto in PhysicalReview Letters, Vol. 88, No. 4, pages 048702-1–048702-4; January 28, 2002.

Chain Letter Evolution. Daniel W. VanArsdale: www.silcom.com/~barnowl/chain-letter/evolution.html

The chain letters used in this article and other data are at www.math.uwaterloo.ca/~mli/chain.html

A discussion of phylogenetic inference methods is athelix.biology.mcmaster.ca/721/outline2/node47.html

Kolmogorov complexity is discussed at www.wikipedia.org/wiki/kolmogorov–complexity

The program for plagiarism detection is at http://dna.cs.ucsb.edu/SID/

RELATEDNESS OF MAMMALS

Ferungulates

Primates

Rodents

Marsupials

GibbonGorillaHumanPygmy chimpanzeeChimpanzeeOrangutanCatHarbor sealGray sealWhite rhinocerosHorseFinback whaleBlue whaleCowRatHouse mouseWallarooOpossum

Chicken

DIVERSE PROBLEMS can be analyzed with the authors’ relatedness measure. Appliedto whole mitochondrial genomes, it produced this phylogeny of mammals. Note howprimates are more closely related to ferungulates than to rodents—which is believedto be true. This degree of kinship is ambiguous when using traditional techniques.The tree is anchored by the chicken as the “outgroup.” Obviously the mammals did not evolve from the chicken, but the chicken and the mammals would have shared a common ancient reptilian ancestor.



ALIC

E Y

. C

HE

N

COCHLEAR IMPLANTS

We hear when the cochlea, in the inner ear, stimu-lates the auditory nerve. Deafness occurs most com-monly when tiny hair cells inside the cochlea are dam-aged as a result of a genetic defect, infection, loudnoise or aging. Cochlear implants bypass the damageby receiving and converting sound into signals sentalong electrodes to cells adjacent to the auditory nerve.

Worldwide, more than 40,000 children andadults depend on cochlear implants. Certain hair cellslining the cochlear ducts alongside auditory neuronsare tuned to respond to specific frequencies. There-fore, implants such as the Nucleus or the Clarionhave from eight to 22 electrodes that surgeons placeat different positions to maximize the range of fre-quency stimuli forwarded to the brain. Recent re-search indicates that more electrodes won’t improveperformance as much as optimizing their placement;most implant wearers perceive loudness properly butcan still have trouble sensing pitch correctly, makingspeech comprehension difficult. “Something is pre-venting the brain from extracting or assimilating allthe coding information,” says Philip Loizou, a pro-fessor of electrical engineering at the University ofTexas at Dallas. “We don’t know what yet.”

The sooner a person receives an implant after be-coming deaf, the more likely he will adapt to the newsound input; people who have been deaf for years donot respond as well because of degeneration in thecochlea or auditory nerve. Typically, success meansthe wearer can hear moderate and perhaps softsounds, can communicate without lip reading or sign-ing, and may be able to converse over the telephone.Completely normal hearing is still uncommon.

To help, engineers are trying, among other things,to tailor signal-processing algorithms for particularsituations. People with implants often have troubleperceiving speech clearly in noisy environments andappreciating music, which has complex waveforms.Today’s implant processors come with a general al-gorithm, but perhaps they could store specialized onesas well. “That way,” Loizou says, “an individual couldselect an algorithm depending on whether he was talk-ing at home, eating in a noisy restaurant or sitting ata concert.” —Mark Fischetti

To Hear Again

WORKINGKNOWLEDGE

Transmittingcoil

Microphone

Sound waves

Receiver

Eardrum

Malleus

OUTER EARcollects the pressure waves of a sound, which the eardrum converts into mechanical vibrations in tiny bones in the middle ear. The oscillating stapessets off pressure waves within the fluid in the cochlea,which in turn stimulate nerve cells on the auditorynerve that leads to the brain.



➤ LOUD! Audiologists measure how loud a test signal must become

before a person can hear it. The normal threshold is anything up to 15

decibels. People who have “moderate” hearing loss require a signal

of 41 to 55 dB; “severe” loss, 71 to 90 dB; and “profound” loss, 91 dB

or greater. Typical conversation rings in at 40 to 50 dB; freeway traf-

fic at 50 feet, 70 dB; and a blender, 90 dB. Only people with severe or

profound loss in both ears qualify for cochlear implants.

➤ HAIRPIN TURN: Tipping the head forward, back or to the side tilts

a gelatinous substance inside the vestibule, beside the cochlea. The

shifting gel bends hair cells that tell attached nerve fibers which way

the head is headed, so the brain can keep the body upright. Similar-

ly, hair cells inside three fluid-filled semicircular canals, at right an-

gles to one another just above the cochlea, respond to sudden

changes in the head’s velocity, to maintain balance.

➤ HARD ON HEARING: In its “Position Statement on Cochlear Im-

plants,” the U.S. National Association of the Deaf notes that the tech-

nology doesn’t always work. It also discourages the use of implants

in children who are born deaf or become deaf before learning lan-

guage, because even with the technology, it is very hard for them to

develop the cognition for spoken language. Meanwhile the children

are often not taught visual, or sign, languages, which “can result in

developmental delays that can be extremely difficult to reverse.”

DID

YOU

KNOW

...

This month’s column was suggested by reader Thomas Boehm.

Have a topic for a future column? Send it to [email protected]

Incus

Cochlea

Stapes

Basilarmembrane

Electrodebundle

Electrode

Neuron

Hair cell

Auditorynerve

Fluid-filledduct

Processor (in shirt pocket)

Vestibule

Semicircularcanals

COCHLEAR IMPLANTuses a microphone worn behind the ear to pick up sound and send it to aprocessor, where integrated circuits and algorithms amplify, digitize andfilter the sound into a coded signal sent to the transmitter coil. The coilsends the signal via radio waves through the skin to an implanted receiver.The receiver converts the waves into electrical impulses that travel alongelectrodes that end at cells at certain points along the cochlea. A magnet in the transmitter holds it against the implanted receiver.

COCHLEAtransmits pressure wavesthrough its duct fluid, displacingthe basilar membrane, whichbends hair cells to variousdegrees. The cells releaseneurotransmitters that causeattached neurons to fire, tellingthe brain where along the ductthe bending has occurred (whichcorresponds to the frequency ofthe original sound) as well as theamplitude of the bending (whichindicates the loudness).


The song is irritatingly familiar. You’veheard it before, stacks of times. Its nameis on the tip of your tongue, but you justcan’t place it. Hum along a little, maybethat’ll help. No? Oh, what is the darnthing called?

Technology should be able to resolvethis dilemma. Songs have been availablein digital form for many years. Couldn’tyou identify a tune by matching it to asong stored in an enormous musical data-base? Alas, it’s not that easy. “When Istarted looking at this question, I was toldby all the experts in the field that this wasa next-to-impossible task,” says AveryWang, chief scientist at Shazam Enter-tainment, an English company that spe-cializes in music-recognition software.

The problem is the huge number ofpossible combinations of musical notes.The potential database is so large thateven the best software algorithms wouldbe hard put to search it. Some research in-stitutions have tried to employ probabil-ity analysis to ease the task of distin-guishing songs. “With every note that isplayed, the next note can be either high-er, lower or the same,” explains MarkSandler of Queen Mary College at theUniversity of London. “By analyzing thepatterns of vast amounts of existing mu-sic, you can use the preceding notes towork out the probability of where thatnext note will be on the scale.” Suchanalysis, however, takes time and pains-taking effort. And one must also take intoaccount the individual interpretationsthat composers, conductors and musi-cians make, the changes to tempo andstyle that can completely alter the texture

of a piece, and the fact that the copyingand sampling of certain passages of mu-sic within other pieces is commonplace.

Nevertheless, the first breakthrough ishere. Shazam Entertainment has createda commercial music-recognition system,now available in Britain, that allows peo-ple to identify tunes using their cellphones. When you hear a song—on theradio, over a public address system, in abar or on television—you punch in theShazam service’s four-digit code numberon the handset, point the phone at the

source of the sound and hold it there for15 seconds. Within a few minutes, the ser-vice should return a text message givingthe name of the song and the artist.

An American, Chris Barton, co-found-ed Shazam three years ago after hedreamed up the idea while studying at theLondon Business School. Wang wasworking on sound analysis and manipu-lation at Stanford University when Bartoncaught up with him. “At first, I resisted,”Wang says. “The general current of aca-demic thought was that this was such a

TECHNICALITIES


JOH

N F

RAS

ER

Name That TuneMUSIC LOVERS USE CELL PHONES TO IDENTIFY SONGS BY FIONA HARVEY

THIS IS YOUR SONG: A new music-recognition system developed by Shazam Entertainment is nowavailable in Britain. Just point your cell phone at the speakers, and you receive a text message givingthe name of the song and the artist—as long as the background noise is not too loud.


very hard problem that it would take yearsof work.” But then Wang had a brain-storm. Instead of messing around withthe probabilities of different notes follow-ing one another, he plotted graphs of mu-sical frequency against time. Creating dig-ital signatures of musical pieces had beendone before—you can do it on a home

computer with ordinary audio software.Such signatures are complex, though, andfinding a match requires the entire graphto be compared with the graphs of all oth-er tunes, which takes time.

Wang’s innovation was to eliminateas much information from the signaturesas he could while ensuring that they werestill recognizable and unique to each pieceof music. For each song, Wang recordedonly the peak and trough of frequency inevery 10-millisecond segment. Shazam’sdatabase currently contains the stripped-down digital signatures of 1.7 millionsongs. When a user with a mobile phoneinputs 15 seconds of music, the systemcreates a digital signature for that snippetand then looks for a matching pattern inthe database. Wang offers an analogy:“Imagine 306 million seconds of songstranslated into constellations of dots andprinted out in a long line. Next, imaginethat the constellation of 15 seconds thatyou want to match is printed on a piece oftransparent plastic. You slide the plasticalong the whole line of dots until you findthe same pattern.”

Wang calculates that the probabilityof finding a false match is less than threein a million, which means that the system

can work even when only 1 percent of thesong is audible. To minimize the retrievalspeed, all the data are kept in active mem-ory (rather than on hard disks) on a dis-tributed computer system consisting ofabout 70 PCs.

In practice, Shazam is easy and fun touse. I tried it out in cafés and bars, in busy

shopping malls, in the backseats of auto-mobiles while the radio was playing andin my own home. Most times, the servicesuccessfully recognized the music, return-ing the name of the song by text messagewithin two or three minutes. Not surpris-ingly, I got the best results when the mu-sic was clearest—from the radio or CDplayer at home. But Shazam identifiedBlondie’s “Heart of Glass” even whenplayed on a taxicab’s crackly radio andthe Beatles’s “Hey Jude” when broadcastover the public address system of a shopin London’s West End. I found that thebest way to use the service was to pointthe phone along an obstacle-free path to-ward the source of the sound, avoidingpeople engaged in loud conversations. (Igot some funny looks from passersby asI earnestly held out my mobile phone inthe middle of shopping malls.) Each 15-second call costs 50 pence, or about 75cents, which is simply added to your cell-phone bill.

Shazam’s database focuses on newermusic. When I tried out a couple ofrecordings from the 1950s by blues singerBlind Willie McTell, the system wasstumped. Shazam has not yet includedany classical music in its database either,


Shazam’s database focuses on newer music. But the system does not always work

optimally in the noisy haunts of music lovers.


because the company believes that itsmain users will be young people seekingto identify popular songs. But the systemdoes not always work optimally in thenoisy haunts of music lovers. I got no re-sponse from Shazam when I attempted toget it to recognize Otis Redding’s “Sittingon the Dock of the Bay” and DustySpringfield’s “I Only Want to Be withYou” in a busy pub.

These failures were most likely the re-sult of excessive background noise anddistortions in the frequency of the songs.Background noise can be a problem be-cause European mobile phones always fo-cus on the loudest sounds and throwaway the quieter signals. And when discjockeys speed up or slow down songs,they alter the frequency profiles that Sha-zam relies on. Some tolerance is built intothe system’s search engine, though; Wangsays that Shazam can recognize songsplayed within about 10 percent of theiroriginal speed.

Despite the shortfalls, Shazam’s tech-nology is impressive. The company isplanning to offer new services that will al-low users to identify a musical piece sim-ply by humming or maybe even whistlingit into the handset. And Shazam is notalone. Philips, the Dutch electronics giant,is developing a prototype system designedto recognize hummed versions of songs.Philips could not say when such a prod-uct might be commercially available,however.

About 300,000 people in Britain havetried Shazam’s service since it was intro-duced last August. The company ispreparing to launch the service in Ger-many soon and to make it available in theU.S. and some Pacific Rim countries in thenext 12 to 18 months. By then, Shazam’smusical database should be even bigger—

perhaps even big enough to recognizeBlind Willie McTell.

Fiona Harvey is a journalist based inLondon who writes about all aspects of technology for the Financial Times.

J U N E 2 0 0 3

TECHNICALITIES


REVIEWSC

OU

RTE

SY O

F H

EN

RY

HO

LT A

ND

CO

MP

ANY


Chelsea Thomas was born with Möbiussyndrome, in which a nerve that trans-mits commands from the brain to the fa-cial muscles is missing. As a result, for herfirst seven years Chelsea looked perpet-ually grumpy. Then surgeons transplant-ed nerves from Chelsea’s leg to both sidesof her mouth, and today Chelsea can dowhat most people in the world take forgranted. She can smile.

Meanwhile, thousands of adults arebotoxing the nerves that allow them tofrown. Actors who do so cannot conveyanger or fear, and some botoxed moth-ers complain that their children no longertake their admonitions seriously, accom-panied as they are by the mothers’ blandexpressions.

Paul Ekman would not be surprised.He has been studying facial expression ofemotions for some 30 years, in the nobletradition of Aristotle, who first observedthe characteristic facial expressions ofanger, fear “and all the other passions,”and Charles Darwin, who added an evo-lutionary explanation. Darwin’s theoryof the universality of emotional expres-sion was unpopular in the 1960s, whenEkman began his research. It was the era

of the tabula rasa in social science; Ek-man was to emotion what Harry Harlowwas to love, swimming against the aca-demic tides. As a graduate student at thetime, I was in that tide up to my neck,and I remember how vehemently psy-chologists protested the idea that any as-pect of human behavior might have ahardwired element. Facial expressions?Clearly cultural. Don’t the Japanese cool-ly suppress any sign of emotion, and don’tthe Italians exuberantly reveal theirs?

Over the next decades, Ekman andhis colleagues gathered evidence of theuniversality of seven facial expressions ofemotion: anger, happiness, fear, surprise,disgust, sadness and contempt. In everyculture they studied—in Japan, through-out Europe and the U.S., and among thenonliterate Fore of New Guinea—a largemajority could recognize the basic emo-tional expressions portrayed by people inother cultures, and others could recognizetheirs.

Yet, as Ekman also showed,cultures do differ widely in the“display rules” of emotional ex-pression. Certain emotions areuniversal, hardwired into facialexpressions and the brain; how-ever, emotional expressions areculture-specific. People smile ordisplay anger for many reasons,and they don’t reveal theseemotions when such displayswould be considered rude or inappropriate.

Ekman and his collabora-tor Wallace Friesen created acoding system that identifies

each of the nearly 80 muscles of the face,as well as the thousands of combinationsof muscles associated with various emo-tions. (Ekman can do all of them himself.)When people try to hide their feelings or“put on” an emotion, Ekman found, theyuse different groups of muscles than theydo for authentic feelings. For example,authentic smiles of joy involve the mus-cles surrounding the eyes; false or socialsmiles bypass the eyes completely.

In Emotions Revealed, Ekman, whois a professor of psychology at the Uni-versity of California at San Francisco,beautifully interweaves his research withanecdotes, recommendations, and the be-hind-the-scenes flubs, accidental discov-eries and debates that never make theirway into published articles but that arethe essence of scientific inquiry. He re-views what is known about the triggers,automatic and learned, that set off anemotion and how we might learn tomanage or even get rid of them. He then

A Polite Smile or the Real McCoy?HOW FACIAL MUSCLES REVEAL—AND HIDE—EMOTION BY CAROL TAVRIS

EMOTIONS REVEALED:RECOGNIZING FACES ANDFEELINGS TO IMPROVECOMMUNICATION AND EMOTIONAL LIFEBy Paul EkmanTimes Books, Henry Holtand Company, New York,2003 ($25)

TRUE SMILE of enjoyment (the one on the right, in these twophotographs of author Paul Ekman) involves the muscle thatorbits the eye and is almost impossible to fake.



REVIEWS

examines five emotions in detail: sadness,anger, fear, disgust and contempt, andthe “enjoyable emotions.” I was charmedto find naches on the list (the Yiddishword—it rhymes with “Loch Ness”—forthe pleasure and pride that “parents feelwhen their child accomplishes somethingimportant”), along with “wonder,” de-fined in terms of “its rarity and the feel-ing of being overwhelmed by something incomprehensible.”

Because of Ekman’s emphasis on theuniversality of emotions, especially thosewritten on the face, readers will not learnmuch about the raging debate about emo-tions that do not necessarily have partic-ular facial expressions, such as pride,envy, jealousy, compassion, and roman-tic or parental love (Ekman does not con-sider these to be “emotions,” althoughother researchers do). Nor will readerslearn much about the origins of emotionblends (such as naches, wonder, longing,the feeling of “bittersweet,” and schaden-freude), which are more varied across cul-tures and individuals and which appear tobe uniquely human, involving as they dohigher cognitive processes.

Readers will enjoy seeing the many fa-cial expressions of Ekman’s favorite pho-tographic subject, his daughter, Eve, whomust have received ample compensationin fatherly naches for her ability to isolateand vary her facial muscles to reveal eachbasic emotion. These photographs servebrilliantly for scientific research, butwhether they will help readers becomebetter at accurately detecting another’semotion is doubtful. As research by oth-ers in this field has shown, when we readanother’s emotion, we do so through thefilters and blinders of culture, the imme-diate situation, status, our own history,and degree of familiarity with the target.The face reveals, and the face lies. And as Ekman himself once observed, wewouldn’t want it otherwise.

Carol Tavris, a social psychologist, isauthor of Anger: The MisunderstoodEmotion (Touchstone Books, 1989).

WHERE IS EVERYBODY?by Stephen Webb. Copernicus Books, New York, 2002 ($27.50)On the way to lunch at Los Alamos Scientific Laboratory one day in 1950, Enrico Fermiand three other physicists—Emil Konopinski, Edward Teller and HerbertYork—chatted about flying saucers. At lunch, when the talk had turned to othermatters, Fermi suddenly said, “Where is everybody?” His companions realizedthat the talk of flying saucers had turned his mind to the possibility that thereis intelligent life elsewhere in the universe and that he was asking why, if thereis, we have seen no sign of it. The question encapsulates what is now knownas the Fermi paradox. Webb, lecturer in physics at the Open University inEngland, presents 49 solutions that have been proposed for the paradox,grouping them according to whether they hold that intelligent extraterrestrialsare here, exist but have not communicated, or do not exist. He makes a splendidand enlightening story of it, concluding with his own solution, the 50th: “We are alone.”

SETI 2020: A ROADMAP FOR THE SEARCH FOR EXTRATERRESTRIAL INTELLIGENCEedited by Ronald D. Ekers, D. Kent Cullers, John Billingham and Louis K. Scheffer. SETI Press, Mountain View, Calif., 2002 ($25)

The point of view of the SETI project, of course, differs from Webb’s. Formore than 50 years, SETI researchers have been inspired by the followingreasoning: “In a cosmos filled with billions of galaxies containing trillionsof stars, is it possible that Earth, a world of inconsequential size andordinary position, is alone in housing life that can discern the naturalorder? It is deeply incongruous to suppose that our enormous Universeis only sparsely occupied. To do so requires the belief that humans areexceedingly special. In view of astronomy’s history, such a view is clearlysuspect.” This book reviews the history of the search and lays out a plan

for SETI’s next 18 years. The book presents sober science that holds outan awesome prospect should the search succeed: “We would come to view

ourselves and our place in the Universe in a very different light. The changes would be asprofound as those resulting from the revolutionary discoveries of Galileo and Darwin, whichchanged our understanding of our place in the solar system and of our biological evo-lutionary heritage.”

WHAT DOES A MARTIAN LOOK LIKE? THE SCIENCE OF EXTRATERRESTRIAL LIFEby Jack Cohen and Ian Stewart. John Wiley & Sons, Hoboken, N.J., 2002 ($27.95)The authors of this book also see a possibility of extraterrestrial life. “We now know thatthere are many planets out there in the galaxy, and we have good grounds for supposingthat a number of these will have life.” It may be strange life, though, they say, nothing likewhat we know on Earth. On those principles, Cohen and Stewart (respectively, areproductive biologist and a professor of mathematics at WarwickUniversity in England) lay a basis for what they call xenoscience—

knowledge of the strange. They draw on serious science—biology,chemistry, astronomy and physics—and also on science fiction, becausethe best of it has “made some useful contributions to the scientificunderstanding of possibilities for alien lifeforms.”

All the books reviewed are available for purchase through www.sciam.com

THE EDITORS RECOMMEND . . .. . . THREE BOOKS ON EXTRATERRESTRIAL LIFE



JOH

NN

Y JO

HN

SON

PUZZLINGADVENTURES

Prime Squares BY DENNIS E. SHASHA

First move, Player 1 Score: 0 to 0

Fifth move, Player 1 Score: 3 to 0

Sixth move, Player 2 Score: 3 to 2

Seventh move, Player 1 Score: 6 to 2

Eighth move, Player 2Score: 6 to 3

Ninth move, Player 2Final score: 6 to 6

Second move, Player 2 Score: 0 to 0

Third move, Player 1 Score: 1 to 0

Fourth move, Player 2 Score: 1 to 0

9

5

6

9

53

3

11

3

6 7

9

3

1

3

6 7

9

9

353

6

6

9

6 77

5

5

11

3

6

5

9

335

5

3

5

6

1

The grid directly below is a prime square: thenumbers in the rows (769, 953, 797) and columns(797, 659, 937) are all prime, none begins with zero,and no two row or column primes are the same. The

grid is also “ambidextrous,” be-cause the numbers in the rows arestill prime if read backward (fromright to left). The square is not,however, “omnidextrous.” To earnthat label, its columns, too, would

have to be prime when read backward (bottom totop), and its diagonals would have to be prime inboth directions [see squares at far right ]. In thiscase, 956 is even and therefore not prime.

This puzzle asks you to create other primesquares. To warm up, build a prime 4-square (con-sisting of four-digit numbers) that uses nine distinctdigits. This square need not be ambidextrous. (You can finda table of prime numbers atwww.sciam.com/ontheweb.)See one solution at the right.

Now for the problems. De-sign an omnidextrous prime

3-square that uses as few distinct dig-its as possible. Then try to make aprime 5-square that uses all 10 dig-its; this one does not have to be om-nidextrous or ambidextrous. Now itgets harder. Can you find the set ofnumbers n having prime n squares(whether plain, ambidextrous oromnidextrous)? I won’t have an an-swer for that one; researchers are stillworking on it. But if you enjoy play-ing with primes, here’s a new gameto try: tic-tac-prime.

A move in tic-tac-prime consistsof placing a single digit in an emptysquare of a three-by-three grid. Thetwo players alternate moves except that the secondplayer gets the ninth move. When one player’s dig-it completes a segment of three digits in a line, thatplayer gets a point for each three-digit prime (read-ing in any direction) that includes this last-placeddigit (so two points for numbers that are primebackward and forward). Can you discover a win-ning strategy for either player?

Answer to LastMonth’s PuzzleWhen the possibleoffers are $1 and $5,the best strategy isto take the first 50offers no matter whatand then take only $5offers. If the ticketexchanger offers $1for each ticket andthen halts thetrading, your regretratio would be 1.8($90 divided by $50).If the exchangerinitially offers $1 andthen switches to $5,the regret ratio wouldstill be 1.8 ($450divided by $250). If the possible offersare $1 and $1 million,take the first 45offers no matter what and then waitfor $1-million offers.This strategy yields a regret ratio of 2.

Web SolutionFor a peek at theanswer to thismonth’s problem, visitwww.sciam.com

7 6 9

9 5 3

7 9 7

3 2 5

8 4 1

8 2 6

3

9

3

9 3 7 1

A SAMPLE GAME OF TIC-TAC-PRIME

PLAIN

AMBIDEXTROUS

OMNIDEXTROUS


ANTIGRAVITY

Tennyson’s well-known description ofnature as “red in tooth and claw” con-jures up dramatic moments in the prey-predator relationship. But nature’s paletteis rich, and Lord Alfred neglected to con-template the grays, browns, blacks, greensand even blues to be found along the trail-side. My interest in rectifying the poet’smonochromatic view recently led meto the Loxahatchee National WildlifeRefuge in Boynton Beach, Fla., for anintroductory session on the science ofidentifying scat.

If any of the 15 other attendeesexpected to be taught how to distin-guish Ella Fitzgerald from Mel Tormé,they were sorely disappointed. Therewas singing, however. For the bene-fit of the six children in the room,workshop leader Serena Rinker war-bled: “It begins with an ‘S’ and endswith a ‘T,’ it comes out of you and itcomes out of me, I know what you’rethinking, but don’t say that, ’cause tobe scientific we call it scat.”

The ditty’s silly but accurate: scatstudy truly is scientific. As authorJames Halfpenny notes in his book AField Guide to Mammal Tracking inNorth America, “Scat, the feces of mam-mals, provides . . . an important windowinto the mammalian world. Informationwe gather from scat includes mammalidentification and presence, location ofactivities, composition of diet, seasonaldiet changes, and samples of prey species.”Such species-by-feces analysis works fornonmammals, too: should I have an inter-est in, for example, the many large alli-gators at the Loxahatchee refuge, scat is

a safe window through which to gaze.Rinker wears the uniform of the U.S.

Fish and Wildlife Service and describesherself as “only a part-time amateur scat-ologist.” (Really, does anyone want to bea full-time professional scatologist?) Shepresented a brief primer on scat basics, re-plete with a table set with samples from

her collection, after which we attemptedto identify the icky. Size and shape give ob-vious evidence, as big animals tend to dobig things. Contents are likewise clear indi-cators of who’s who by what’s what: fishscales and crayfish parts suggest otter, saysHalfpenny in his field guide. Cameras andcell phones might therefore suggest bear.

“Color also provides important clues,”Rinker observed. A gator contributionwas both large and gray, indicating a

(possibly literally) heady mixture of hairand meat. Consuming grass can producegreenish scat; white scat may be old anddried-out or high in phosphates found inthe bones of a prey animal that began itsjourney at the other end of the alimenta-ry canal; brown indicates vegetable mat-ter, even wood; blue means a critter

passed the berries. Black scat has seensome fat, the result of a purely meatdiet, although Rinker pointed outthat a very dark sample on the tablewas in fact the result of coprophagy.I leave it to the more adventurousreaders to look up the definition of co-prophagy and then to wonder at na-ture’s rich tapestry.

After the lecture, Rinker led us ona scat search along a boardwalk thatsnakes through a cypress swamp sec-tion of the refuge. I once encountered aplayful otter wandering along theboardwalk, but on this day Rinker wasdelighted to find relatively fresh otteroeuvre, luminous with fish fins andcrayfish shell shards. “Oh, my gosh!”she exclaimed. “I actually should comeback and collect it, because this is oneI don’t have.”When the walk was over and the class

dismissed, Rinker and I returned to thescene of the grime. She collected the scatand explained that the next step was to letit dry on newspaper, a plan that seemedquite sensible to me as a magazine writer.I then left the refuge, mulling over how ascientific outlook can make finding a pileof poop cause for celebration. And mus-ing that the scatologist’s motto must be“Good thing we didn’t step in it.”


ILLU

STR

ATIO

N B

Y M

ATT

CO

LLIN

S; P

HO

TOG

RAP

H B

Y FR

ANK

VE

RO

NSK

Y

Dropping BySOMETIMES A NATURALIST CAN GET AHEAD BY ASSESSING WHAT’S BEEN LEFT BEHIND BY STEVE MIRSKY


Q

Q


MAT

T C

OLL

INS

Sant P. Singh, a professor and chief of endocrinology, diabetesand metabolism at Chicago Medical School, offers this answer:

Several factors appear to be involved in getting a hang-over—the unpleasant consequence visited on 75 percent ofthose who drink alcohol to intoxication. The effects includeheadache, nausea, vomiting, thirst, dryness of the mouth,tremors, dizziness, fatigue and muscle cramps. Often there isan accompanying slump in cognitive and visual-spatial skills.

A hangover has been suggested to be an early stage of al-cohol withdrawal. Mildshakiness and sweats canoccur; some people mayeven hallucinate. Acet-aldehyde, a toxic break-down product of alcoholmetabolism, plays a rolein producing symptoms.Chemicals known as con-geners that are formedduring alcohol processingand maturation also in-crease the likelihood andseverity of a hangover; asa rule of thumb, the dark-er the liquor, the morecongeners it contains. The toxins in congeners are distributedthroughout the body as the liver breaks down the alcohol. Last,hangovers cause changes in the blood levels of various hor-mones, which are responsible for some symptoms. For exam-ple, alcohol inhibits antidiuretic hormone, which leads to ex-cessive urination and dehydration. Blood aldosterone and reninlevels also increase with a hangover—but unlike antidiuretichormone, they do not correlate well with symptomatic severi-ty, so their role is less clear.

Individuals are more prone to develop a hangover if theydrink alcohol rapidly, mix different types of drinks, and do notdilute the absorption of liquor by eating food or drinking non-alcoholic beverages. Sugar and fluids can help overcome the en-suing hypoglycemia and dehydration, and antacids can reducenausea. To reduce headache, anti-inflammatory drugs should

be used cautiously: aspirin may irritate the stomach, and thetoxic effects of acetaminophen on the liver can be amplified byalcohol. Other drugs have been used to treat hangovers, butmost have questionable value.

Why does shaking a can of coffee cause the larger grainsto move to the surface?

—H. Kanchwala, Pune, IndiaHeinrich M. Jaeger, a professor of physics at the University of Chicago, explains:

The phenomenon by which larger coffee grains move upand smaller ones travel down when shaking a can is called gran-ular-size separation. It is often referred to as the Brazil nut ef-fect, because the same thing will happen when you jiggle a canof mixed nuts. This occurs for two main reasons.

First, during a shaking cycle—as the material lifts off thebottom of the can and then collides with the base again—larg-er particles briefly separate from smaller ones, leaving gaps un-derneath. The tinier bits then slip into the gaps. When the shak-ing cycle finishes, the large particles cannot return to their orig-inal positions, and therefore the bigger particles slowly “ratch-et” upward.

The second action at work is called a convective mecha-nism. When a can shakes, the coffee rubs against the sides. Fric-tion causes a net downward motion of the grains along thewalls, which is balanced by a net upward flow in the center—

setting up a convection roll pattern. The downward flow is con-fined to a narrow region only a few (small) particle diametersin width. Once the large java grains reach the top, they movetoward the side walls. If they are too large, they cannot fit intothe region of downward flow and, after a few shakes, they ag-gregate near the top. Typically this mechanism dominates un-less friction with the side walls is carefully minimized.

Why do hangovers occur?—P. Bouchard, Orange, Calif.

ASK THE EXPERTS

For a complete text of these and other answers from scientists in diverse fields, visit www.sciam.com/askexpert


radlab.cs.berkeley.eduradlab.cs.berkeley.edu/people/fox/static/pubs/pdf/n07a.pdf · windblown...

Documents