The Eyethe-eye.eu/public/concen.org/Scientific American 1993...physician, and his mother devoted themselves to all aspects of culture. By his early teens Weil had become Òpas-sionately

JUNE 1994$3.95

Starfire laser beam creates a guide star

for adjusting a flexible telescope mirror.

Was there a race to the moon?

How the brain makes emotional memories.

Genetic testing: boon or bane?

Copyright 1994 Scientific American, Inc.

June 1994 Volume 270 Number 6

36

44

50

60

Was the Race to the Moon Real?John M. Logsdon and Alain Dupas

The Classical Limit of an AtomMichael Nauenberg, Carlos Stroud and John Yeazell

Emotion, Memory and the BrainJoseph E. LeDoux

4

66 Early Andean CitiesShelia Pozorski and Thomas Pozorski

Adaptive OpticsJohn W. Hardy

Did the Soviet Union really try to put humans on the moon before the U.S. did? Af-ter the Apollo landing, the Kremlin denied that the U.S.S.R. had been in the race. Butrecollections by former leaders of the Soviet space program, declassiÞed documentsand other primary evidence show otherwise. Internecine battles and high-level inde-cision Þnally defeated MoscowÕs attempts to capture the lunar high ground.

Quantum physics should blend seamlessly into classical physics. After all, billiardballs, Great Attractors, satellites and golden retrievers are made of electrons, pro-tons, neutrons and other particles. Yet the frontier between the microscopic andmacroscopic universes has resisted experimental probingÑuntil now. Pulses oflaser light make giant atoms whose properties come from both worlds.

A sight, a smell or a chord from a melody can evoke an emotional memory. Howdoes the brain recall such emotions? Experiments with rodents model the process.Nerve impulses from sounds that cause fear in rats have been traced along the au-ditory pathway to the thalamus, the cortex and the amygdala, arousing a memorythat leads to a higher heart rate and the cessation of movement.

Atmospheric turbulence hampers earthbound telescopes by distorting the lightfrom near and deep space. Even building observatories on mountains does notsolve the problem, and putting instruments in orbit is expensive. So mirrors are be-ing fabricated that change shape to compensate for the eÝects of troubled air. Muchof the technology grew out of eÝorts to design laser-based antimissile weapons.

A desert site at Pampa de las Llamas-Moxeke reveals evidence of a highly organizedcity whose 2,000 inhabitants bustled more than 3,500 years ago, well before theearliest known great civilizations of pre-Columbian Peru. The economic, social andtheocratic order of this and neighboring communities powerfully inßuenced the de-velopment and character of later Andean urban cultures.


74

82

88

The Ethnobotanical Approach to Drug DiscoveryPaul Alan Cox and Michael J. Balick

DEPARTMENTS

50 and 100 Years Ago1944: Television for peace.1894: Mortal incubator.

116

98

108

112

14 10

12

5

Letters to the EditorsThe hawks v. the owls. . . . A high-energy defense of physics.

Science and the Citizen

Science and BusinessBook ReviewsAlbert in ßagrante. . . . Buoyantwhales. . . . Of ßies and men.

Essay : Anne Eisenberg

ÒNot even false,Ó and otherartful scientiÞc insults.

The Amateur ScientistHow to mess with DNA in theprivacy of your own home.

TRENDS IN GENETICS

Grading the Gene TestsJohn Rennie, staÝ writer

The Sensory Basis of the HoneybeeÕs Dance LanguageWolfgang H. Kirchner and William F. Towne

Scientific American (ISSN 0036-8733), published monthly by Scientific American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111. Copyright © 1994 by Scientific American, 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 ina retrieval system, transmitted or otherwise copied for public or private use without written permission of the publisher. Second-class postage paid at New York, N.Y., and at additionalmailing offices. Canada Post International Publications Mail (Canadian Distribution) Sales Agreement No. 242764. Canadian GST No. R 127387652. Subscription rates: one year $36 (out-side U.S. and possessions add $11 per year for postage). Subscription inquiries: U.S. and Canada (800) 333-1199; other (515) 247-7631. Postmaster: Send address changes to ScientificAmerican, Box 3187, Harlan, Iowa 51537. Reprints available: Write Reprint Department, Scientific American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111, or fax: (212) 355-0408.

How do honeybees tell their nestmates where food outside the hive lies? The ques-tion has been debated since Aristotle Þrst observed apian communication. Contem-porary study of potential foragers responding to a robotic bee indicates that soundand the elaborately choreographed dance carry the message together.

Plants make many chemicals that protect them from infection, predation and otherharm. Biologists seeking new pharmaceutical compounds often screen ßora ran-domly for such agents. But there is a more eÛcient way: analyze plants alreadyused as drugs by indigenous cultures, particularly those of the rain forest.

An embryo can now be screened for genetic disease even before it is implanted inits motherÕs uterus. So the technology can help prevent the tragedy of a life doomedby heredity. But what constitutes a disease? Should genetic testing also be used toselect a childÕs sex or other characteristics? Who should know the results of genetictesting? A relative or Þanc�? An employer? An insurer ?

Cairo population summit. . . . Motherof attractors. . . . Unbound genes. . . .Just a phase. . . . Amazing vanishinglaser. . . . Gathering superstring. . . .Institutionalizing the environ-ment. . . . PROFILE: Andr� WeilÑa calculating life on the edge.

Cyberspace cadets. . . . GraÛti anti-dote. . . . Bioprospectors plunder theSouthern Hemisphere. . . . The spacestation: in the crosshairs. . . . Engi-neering universal immunization. . . .THE ANALYTICAL ECONOMIST: Privatizing eastern Europe.


36Ð37 Courtesy of Glenn Swanson, Quest magazine (left ), National Aeronautics and Space Administration (right )

38 NASA (top ), Sovfoto/Eastfoto (bottom)

39 NASA (top left ), UPI/Bett-mann (top center ), NASA (top right ), Tass, Sovfoto/Eastfoto (bottom)

40 NASA (top left ), AP/WorldWide Photos (top right ), Sovfoto/Eastfoto (bottom)

41 NASA (top ), Sovfoto/Eastfoto (bottom left ), A. Moklet Sov./NovostiPress Agency/StarlightPhoto Agency, Inc. (bottom center ), courtesy of AlainDupas (bottom right )

42 NASA (top ), courtesy of Glenn Swanson, Quest magazine (bottom left ), courtesy of Alain Dupas (bottom right )

43 NASA (top ), courtesy of SothebyÕs (bottom left ), Edwin Cameron (bottom center ), Tom StaÝord, Vance Brand/Starlight Photo Agency, Inc. (bottom right )

44Ð45 Ian Worpole

46 James Montanus, University of Rochester

47 Ian Worpole

48 Jack Harris/Visual Logic(top left ), Ian Worpole (top right and bottom )

49 Ian Worpole

51 Roberto Osti (drawings ), Andrew Leonard/APL Microscopic (photographs )

52 Roberto Osti (top ), Ian Worpole (bottom)

53 Ian Worpole

55 Ian Worpole (drawings ), Joseph E. LeDoux (photographs )

56Ð57 Roberto Osti

61 Roger Ressmeyer/StarlightPhoto Agency, Inc.

62 Jared Schneidman/JaredSchneidman Design

63 Jared Schneidman/JSD(drawings ), John W. Hardy(photographs )

64Ð65 Jared Schneidman/JSD

66Ð67 Steven N. Patricia (top ),Gabor Kiss (middle )

69 Shelia Pozorski and Thomas Pozorski (left ), Steven N. Patricia (right )

70Ð72 Shelia Pozorski and Thomas Pozorski

74Ð75 Mark MoÝett/Minden Pictures

76 William F. Towne

77 Tomo Narashima

78 Tomo Narashima (top ), Mark MoÝett/MindenPictures (bottom )

80 Thomas Seeley

83 Michael J. Balick

84 Michael J. Balick (top ), Roberto Osti (bottom )

85 Gregory Shropshire, Ix Chel Tropical Research Foundation, Belize

87 Paul Alan Cox

88Ð89 Courtesy of Susan Lanzen-dorf, Jones Institute, East-ern Virginia Medical School (left ), Hank Morgan (right )

90Ð91 Jared Schneidman/JSD(top ), Gabor Kiss (bottom )

92 UPI/Bettmann

94 Jared Schneidman/JSD

95 Lester Sloan

96 Abraham Menashe

108Ð111 Kathy Konkle

THE ILLUSTRATIONSCover photograph © 1994 by Roger Ressmeyer/Starlight Photo Agency, Inc.

8 SCIENTIFIC AMERICAN June 1994

THE COVER photograph shows the power-ful Starfire laser beam generated at the U.S.Air ForceÕs Phillips Laboratory in New Mexi-co. The beam, when reflected in the upperatmosphere, creates an artificial guide starthat is used to calibrate the Starfire tele-scopeÕs flexible mirror to compensate foratmospheric turbulence. The man seated atthe foot of the dome is a spotter who warnsof approaching aircraft so that the beamcan be shut down to protect the airplaneÕscrew and instruments (see ÒAdaptive Op-tics,Ó by John W. Hardy, page 60).

Page Source Page Source

¨

Established 1845

EDITOR: Jonathan Piel

BOARD OF EDITORS: Michelle Press , ManagingEditor ; John Rennie, Associate Editor; TimothyM. Beardsley; W. Wayt Gibbs; Marguerite Hollo-way ; John Horgan, Senior Writer ; Kristin Leut-wyler; Philip Morrison, Book Editor; MadhusreeMukerjee; Corey S. Powell ; Ricki L . Rusting;Gary Stix ; Paul Wallich; Philip M. Yam

ART : Joan Starwood, Art Director ; Edward Bell ,Art Director , Graphics Systems; Jessie Nathans,Associate Art Director ; Johnny Johnson, AssistantArt Director, Graphics Systems; Nisa Geller, Pho-tography Editor ; Lisa Burnett, Production Editor

COPY: Maria-Christina Keller, Copy Chief; NancyL . Freireich; Molly K. Frances; Daniel C. SchlenoÝ

PRODUCTION : Richard Sasso, Vice President,Production; William Sherman, Production Man-ager ; Managers: Carol Albert, Print Production;Janet Cermak, Makeup & Quality Control ;Tanya DeSilva , Prepress; Carol Hansen, Compo-sition; Madelyn Keyes, Systems; Eric Marquard,Special Projects; Ad TraÛc: Carl Cherebin; KellyAnn Mercado

CIRCULATION: Lorraine Leib Terlecki, AssociatePublisher/Circulation Director ; Katherine Robold,Circulation Manager; Joanne Guralnick, Circula-tion Promotion Manager ; Rosa Davis, FulÞllmentManager

ADVERTISING: Kate Dobson, Associate Publish-er/Advertising Director. OFFICES: NEW YORK:Meryle Lowenthal, New York Advertising Man-ager ; William Buchanan, Manager, CorporateAdvertising ; Randy James, Elizabeth Ryan. CHI-CAGO: 333 N. Michigan Ave., Chicago, IL 60601;Patrick Bachler, Advertising Manager. DETROIT:3000 Town Center, Suite 1435, SouthÞeld, MI48075; Edward A. Bartley, Detroit Manager.WEST COAST: 1554 S. Sepulveda Blvd., Suite212, Los Angeles, CA 90025; Lisa K. Carden,Advertising Manager ; Tonia Wendt. 235 Mont-gomery St., Suite 724, San Francisco, CA 94104;Lianne Bloomer. CANADA: Fenn Company, Inc.DALLAS: GriÛth Group

MARKETING SERVICES: Laura Salant, MarketingDirector ; Diane Schube, Promotion Manager ;Ethel D. Little, Advertising Coordinator

INTERNATIONAL: EUROPE: Roy Edwards, Inter-national Advertising Manager, London; VivienneDavidson, Linda Kaufman, Intermedia Ltd., Par-is; Karin OhÝ, Groupe Expansion, Frankfurt ;Barth David Schwartz, Director, Special Proj-ects, Amsterdam. SEOUL: Biscom, Inc. TOKYO:Nikkei International Ltd.; SINGAPORE: Hoo SiewSai , Major Media Singapore Pte. Ltd.

ADMINISTRATION: John J. Moeling, Jr., Publisher;Marie M. Beaumonte, General Manager

SCIENTIFIC AMERICAN, INC.415 Madison Avenue, New York, NY 10017-1111

(212) 754-0550

CHAIRMAN AND CHIEF EXECUTIVE OFFICER:John J. Hanley

CO-CHAIRMAN: Dr. Pierre Gerckens

CHAIRMAN EMERITUS : Gerard Piel

DIRECTOR, ELECTRONIC PUBLISHING: Martin Paul

CORPORATE OFFICERS: President, John J. Moeling,Jr.; Chief Financial OÛcer, R. Vincent Barger ;Vice Presidents, Robert L. Biewen, Jonathan Piel

PRINTED IN U.S.A.


LETTERS TO THE EDITORS

UnÞnished Business

In ÒParticle MetaphysicsÓ [SCIENTIFIC

AMERICAN, February], John Horgan ar-gues that we particle physicists havebankrupted ourselves by our own suc-cesses. A ÒdesertÓ of physics betweenthe Large Electron-Positron Collider en-ergies and the scale of grand uniÞedtheories means that our most beautifultheories are inaccessible to experiment,and thus our Þeld is nearing a deadend. This is like saying that biology is awaste of time because the mystery oflife is too diÛcult to comprehend.

Despite the data doldrums of the1980s, the pages of the Physical Review

are Þlled with experimental results inthe physics of heavy quarks and lep-tons, tests of fundamental symmetries,searches for new phenomena and muchmore. Particle physics is as interestingand stimulating as it has ever been.Our successes have only added to thatrichness and to say otherwise reveals ashallow heart.

The best argument for the continuedfunding of particle physics experimen-tation is the one rooted in the truestrengths of our Þeld: its far-reachingbeauty and profound implications. Theexperience of selling the Superconduct-ing Super Collider to ourselves and tothe country has left many of us cynicaland unenthusiastic. But this is not thefault of the ÞeldÑonly of the times. El-ementary particle physics will not dieas long as we remember why we arepursuing it in the Þrst place.

ALAN J. WEINSTEINLaboratory of High Energy PhysicsCalifornia Institute of Technology

The Best Defense

In ÒThe Future of American DefenseÓ[SCIENTIFIC AMERICAN, February], Phil-ip Morrison, Kosta Tsipis and JeromeWiesner argue that collective security,such as coalition forces, can meet anyfuture military challenges. That is sim-ply not so, and the example of the Per-sian Gulf War, to which the authorspoint, demonstrates it. The U.S. tookmonths to build up suÛcient strengthto attack Iraqi forces in Kuwait. The sea-lift capability of the U.S. is sadly lack-ing. The U.S. merchant ßeet is practical-ly nonexistent. The airlift capacity was

stressed to the point that part of theCivil Reserve Air Fleet was required.

If the active forces are to be signiÞ-cantly reduced, then the reserve forcesmust be increased to retain qualiÞedpersonnel for future conßicts. Addition-ally, the industrial base must be main-tained and available to provide for arapid buildup if needed.

The military still has valid Ònonmili-taryÓ missions around the world and athome. The basic rule for oÝensive op-erations is a three-to-one advantage inpersonnel and equipment. Perhaps alittle more consideration is needed be-fore the U.S. military shrinks away pastthe point of recovery.

( I am a major in the U.S. Army and agraduate of the U.S. Army Commandand General StaÝ College. These viewsare strictly my own and do not reßectthe oÛcial positions of the U.S. govern-ment, the Department of Defense orthe Department of the Army.)

NIELS J. ZUSSBLATT

ChesterÞeld, Mo.

The U.S. does not have excessive air-lift and sea-lift capability when it comesto addressing ÒbrushÞreÓ wars. Becausewe can only guess where we will con-front aggression next, there should bean emphasis on weapons and equip-ment that make possible a powerful,conventional response in hours or daysrather than weeks or months. It takesdecades to introduce new weapons sys-tems and considerable time to bring oldones out of mothballs; defense reduc-tions will eÝectively be irreversible. Weshould resist the temptation to baseour decision for our future defense onbean counting and wishful thinking.

CHRISTOPHER ROSEBERRYRowlett, Tex.

I agree with the authors that thereshould be some kind of drawdownfrom the years of the Reagan militarybuildup, but the plan proposed in thearticle should be sent back to the draw-ing board. Planning based on the as-sumption that the U.S. has only fourpotential adversaries ( Iran, Iraq, NorthKorea and Libya) is an exercise withblinders. NATO has been paralyzed bythe dilemma of whether to intervene inthe Yugoslavian civil war. Some Penta-gon planners thought that Þghting in

the mountainous terrain would requiremore combat personnel than had Op-eration Desert Storm. What wonderfulglue holds Ukraine or Belarus together?How big a peacekeeping force would itrequire to sort out a civil war there pat-terned on Serbia versus Bosnia?

W. D. KELLY

Houston, Tex.

The authors reply:It is conÞdence in our strategic-warn-

ing capabilities and in the prodigiouscapabilities of the U.S. Marine Corps,not bean counting or wishful thinking,that led us to our recommendations. Inthe Gulf, the U.S. was able to insert trip-wire forces in Saudi Arabia promptly, aswas urgently needed, and then to buildup to win. We agree that sea lift and air-lift should be maintained and that re-serve forces should be augmented as welower active strength. In addition, air-refueling tankers, now not needed forstrategic missions, can support a U.S.air presence over many distant battle-Þelds more cheaply than maintaining12 carrier task forces.

Because few people foresee that theU.S. will be the aggressor anywhere inthe world, we do not provide for sud-den oÝensive operations requiring athree-to-one advantage. Finally, we donot believe the U.S. should be involvedin every civil war conceivable, certainlynot without our allies. What threatensUkraine or Belarus most is not war buteconomic collapse, which we shouldhelp prevent with a policy requiring po-litical leadership, even generosity, andnot guns.

Letters selected for publication may

be edited for length and clarity. Man-

uscripts will not be returned or ac-knowledged without a stamped, self-ad-

dressed envelope.


ERRATA

The credit for the illustration on page28 of the March issue should read ÒAn-drew Hanson/© Wolfram Research.Ó

On page 58 of the April issue, the topleft magnetic resonance image scan mis-takenly lists the numbers identifying theother scans in reverse order. The slicesshould be numbered Ò1 2 3 4 5 6.Ó



50 AND 100 YEARS AGO

JUNE 1944

ÒTelevision oÝers the soundest basisfor world peace that has yet been pre-sented. Peace must be created on thebulwark of understanding. Internation-al television will knit together the peo-ples of the world in bonds of mutualrespect; its possibilities are vast, in-deed.ÑNorman D. Waters, President,

American Television Society.Ó

ÒStatistics show that, while much hasbeen done to reduce industrial acci-dents, there is a long way to go. For ex-ample, from Pearl Harbor until January1, 1944, 32,078 soldiers, sailors, andmarines died as war casualties; 94,000workers were killed in accidents. Thenumber of workers injured will dwarfthe total of war wounded: 45,595 ofour armed men were wounded up toJanuary 1, 1944, while 8,800,000 work-ers were injured.Ó

ÒOne of the most persistent enemiesof safe ßyingÑformation of ice on pro-pellers of planes in ßightÑis now beingovercome by a new electrically heatedpropeller ÔskinÕ that enables the propel-ler surface to warm up like a sick-bedheating pad. The skin is made by twokinds of synthetic rubber, the outer sur-face being a thin coating that is tailor-made to conduct electricity instead ofblocking its ßow.Ó

JUNE 1894

ÒThe tendency of the present day isthat the horse must go, must go meta-phorically, for his days of labor seemnearly passed.Ó

ÒThe theory is advanced by S. E.Christian, in Popular Astronomy, thatstellar scintillation is caused largely byinconceivable numbers of small mete-oric bodies, which are constantly pass-ing between the stars and our earth.Momentary oscillation of the stars bythese bodies would certainly occur ifthese bodies were numerous enough,and recent investigation seems to pointto the fact that they are.Ó

ÒMr. Francis Galton aÛrms that Ôthe

patterns of the papillary ridges uponthe bulbous palmar surfaces of the ter-minal phalanges of the Þngers andthumbs are absolutely unchangeablethroughout life, and show in diÝerentindividuals an inÞnite variety of formsand peculiarities. The chance of twoÞnger-prints being identical is less thanone in sixty-four thousand millions. If,therefore, two Þnger-prints are com-pared and found to coincide exactly, itis practically certain that they are printsof the same Þnger of the same person;if they diÝer, they are made by diÝer-ent Þngers.ÕÑLancet.Ó

ÒThe Medical Record tells of a wom-an in Ohio who utilized the high tem-perature of her phthisical husband foreight weeks before his death, by usinghim as an incubator for hensÕ eggs. Shetook 50 eggs, and wrapping each onein cotton batting, laid them alongsidethe body of her husband in the bed, he

being unable to resist or move a limb.After three weeks she was rewardedwith forty-six lively young chickens.Ó

ÒWe publish to-day an engraving (forwhich we are indebted to the Illustrirte

Zeitung) of the gigantic orang-outangin the Zoological Garden at Leipsic,Germany. This and two others that diedlast winter from the eÝects of the severeweather are the only full-grown orang-outangs that have ever reached Europealive. The animal is not as tall as onewould suppose from a Þrst glance, forhe measures only a little over 4 feet.The orang-outang shown has lost oneof his upper eye teeth. Many scars onhis hands and feet show that he hasled an eventful life and received honor-able wounds. His left thumb is bent andone of his toes is crippled. In captivityhe eats soaked rice, milk, raw eggs, or-anges, dates, and he is very fond of ba-nanas and white bread.Ó

The new orang-outang in the Leipsic Zoological Garden


Population SummitWomenÕs health and rightsshape Cairo document

This fall in Cairo the United Na-tions will hold its once-a-decadeconference on population. And if

the third and Þnal preparatory meetingheld in April at U.N. headquarters isany indication, the plan the confereeswill consider could diÝer radically fromits predecessors. Women in the hun-dredsÑand in the cloth and color of ev-ery cultureÑtook over the halls of theU.N., shaping, with unprecedented force,the so-called plan of action that willemerge from the Cairo meeting in Sep-tember. This document will provide aframework for the next 10 years of U.N.population programs. The Cairo meet-ing will presumably ratify it, and gov-ernments will pledge funding.

The Cairo text covers many of thesame issues as did the 1974 Bucharestand 1984 Mexico City plans. The targetsinclude stabilizing the worldÕs popula-tion, currently 5.7 billion people, at 7.8billion by 2050, instead of the projected12.5 billion. Providing family-planning

services to the 350 million couples whowant but cannot obtain them continuesto be a crucial goal as well.

But the draft plan of action also in-cluded phrases and words that neversaw the light of day in previous U.N.population documents: reproductiverights, sexual health, female genital mu-tilation and gender equity. This new em-phasis reßects the belief of womenÕshealth organizations and family-plan-ning experts that to address issues ofpopulation, governments have to ad-dress the health of women and theireconomic and social well-being; coercivenational family-planning programs orservices that do not take a clientÕs needsor culture into account are doomed tofail. ÒThe Þeld is getting much moresophisticated,Ó notes Joan Dunlop of theInternational WomenÕs Health Coalition.

Experts say the reason for the changeat the U.N. lies in the novel role womenand nongovernmental organizations(NGOs) are playing in the diplomaticprocess. ÒThere are far fewer gray suits,Ócomments Sally Ethelston of Popula-tion Action International. ÒWhat we areseeing is that the [1992 Earth Summit]opened the doors for NGOs. Particular-ly in the Þeld of family planning, there

is a recognition on the part of the dele-gates that the NGOs are most innova-tive. They are the ones that pioneereddoor-to-door delivery of contraceptivesin Bangladesh.Ó Some 900 NGOs wereaccredited to attend the Þnal prepara-tory meeting; many delegations in-clude NGO representatives.

The document, as it stood in earlyApril, oÝered several fresh approaches.They included improving girlsÕ accessto education and addressing the con-traceptive needs of adolescents as wellas the responsibility of men for popula-tion growth, their sexual behavior andfertility. Because men stay fertile muchlonger than women do, the averageman, by the end of his lifetime, couldbe responsible for more children thanthe average woman, according to AaronSachs of the Worldwatch Institute. Forinstance, Òmen in Kenya have more chil-dren than women do,Ó Dunlop adds.ÒThat is stating the obvious, but it is avery new thought.Ó

But in their eÝorts to change dramat-ically the focus of the text, some NGOshave had to battle the tireless eÝorts ofthe Vatican to inßuence the summit.Certain NGO leaders assert that the Vat-icanÕs attacks on family planning and

SCIENCE AND THE CITIZEN


HEALTH SERVICES FOR WOMEN, such as this family-plan-ning clinic in Egypt, are the focus of the document that will be

Þnalized at the United NationÕs International Conference onPopulation and Development in Cairo this September.

DO

NN

A D

EC

ES

AR

E Im

pact

Vis

uals


abortion seem especially Þerce this time,possibly because the oÛcial support itenjoyed from presidents Ronald Rea-gan and George Bush no longer exists.

Prior to the New York meeting, PopeJohn Paul II issued a statement callingthe International Conference on Popu-lation and Development a project to allow the Òsystematic death of the un-born.Ó The Pope has also written tomany national leaders urging them tocombat some goals of the conference.At the session itself, the Vatican dele-gation, led by Monsignor DiarmuidMartin, requested that many referencesto women and all references to abor-tion and contraception be bracketedÑthat is, reserved from approval.

The VaticanÕs oÝensive has encoun-tered deeply felt opposition. ÒOne ofthe extraordinary breakthroughs hasbeen the degree to which women havebeen outspoken about their distastefor and opposition to the Vatican,ÓDunlop explains. Some women fromcountries that are largely Catholic havedenounced the VaticanÕs claim to rep-resent their sex. Many of these womenhave presented data on the schisms ap-parent between the churchÕs male lead-

ership and its followers. In the U.S., forexample, 87 percent of Catholics be-lieve couples should make their owndecisions about birth control, accord-ing to a Gallup poll; 84 percent believeabortion should be legal in all or somecircumstances.

In a tactical session, Frances Kissling,director of the Washington, D.C.ÐbasedCatholics for a Free Choice, wearing ablack dress that resembled a priestÕsrobe, urged humor in dealing with theVatican. Other NGOs have questionedthe right of the Vatican to maintain per-manent observer status at the U.N., giv-en that Jews, Muslims, Buddhists, Epis-copalians and other religious groups donot have the same privilege.

Nevertheless, the VaticanÕs successin bracketing many terms could ulti-mately mean that the Þnal language ofthe plan of action is not as far-reachingas some family-planning experts andwomenÕs health advocates would like.If phrases addressing the need for safeabortionsÑeven in countries where thepractice is illegalÑremain bracketedwhen they appear in Cairo, the confer-ence may become focused on the abor-tion debate rather than on population

issues. (A study presented at the prepa-ratory meeting by the Alan GuttmacherInstitute reported that every year about2.8 million women have abortions and550,000 are hospitalized for relatedcomplications in six of the Latin Amer-ican countries where the practice is ille-gal: Brazil, Peru, Chile, Colombia, theDominican Republic and Mexico.)

The ultimate outcome of the strugglebetween some NGOs and the Vaticanwill only become clear in September inCairo. Much of the implementation ofthe plan will depend on how forthcom-ing governments are with money. TheU.N. Population Fund anticipates thatthe broad-based plan will cost morethan $13 billion a year by 2000Ñsome$4 billion is currently spent every year.

In the meantime, the U.N. is a diÝer-ent place. Children sleep on chairs inthe corners of conference rooms whiletheir mothers lead discussions on thedangers of self-induced abortion or theinformal economic sector. In hallways,men stand out because they seem rareand exotic against the backdrop of blueand gold saris, green and yellow head-dresses and the rainbow textiles ofLatin America. ÑMarguerite Holloway


Standing a safe distance outside a black hole, toss in acoin. As it nears the black hole’s horizon—the point of

no return—the coin will seem to fall ever more slowly un-til it hardly moves. Now suppose that the elementary par-ticles making up the coin resemble not points but tiny bitsof string. As they fall in, the strings grow continuouslylonger. They wind around until they encase the black holein a giant spaghettilike entanglement.

Odd? An inevitable blend of black hole physics andstring theory, says Leonard Susskind of Stanford Universi-ty. The black hole warps the space-time around it soacutely that time stretches out as in a slow-motionmovie—one microsecond for the coin seems to us to beseveral days or years. Even though the coin does fall intothe black hole, we can only see it slow down and come toa stop at the horizon.

Moreover, a string, like the wings of a hummingbird, isalways vibrating. Most of the time such movement is justa blur. But catch it in a slow-motion movie, and the vibrat-ing object suddenly looks opaque—and larger. So, too, astring; it grows longer if we are able to see it sloweddown. Further, a string vibrates in many different ways.Thus, as it falls toward the black hole, and its microsec-onds stretch out into minutes or days, it seems from ourpoint of view to elongate endlessly.

This picture would be merely a curiosity if it did notpromise to solve what Susskind calls “a puzzle as deep asthe constancy of the speed of light was” at the turn of thelast century. The puzzle is the information paradox. Firstposed in 1974 by Stephen W. Hawking of the University ofCambridge, the information paradox notes that objects

such as encyclopedias or elephants can fall into a blackhole, never to be seen again. What happens to the knowl-edge they carried, the details about the atoms they weremade of? If, as Hawking believed, these are lost forever,then physics is in trouble. Whereas in practice informationcan be irretrievable, Gerard ’t Hooft of Utrecht Universityhas explained, quantum mechanics dictates that in princi-ple the information should still be there in some form.

“Theoretical physicists have been very thoroughly con-fused for some time,” says Edward Witten of the Institutefor Advanced Study in Princeton, N.J. One suggested wayout of the paradox is that as the coin falls toward the blackhole’s horizon, its information is somehow scrambled andsent back to us as radiation. Still, the horizon can hold aninfinite amount of ordinary matter. Within its finite lifetime,how can the black hole possibly emit the infinite amountsof information the matter must have carried in?

This is where string theory holds out some hope. Ifstrings make up matter, they will spread out and take upall the room at the horizon—allowing the black hole to ab-sorb only a finite amount of material. Presumably infor-mation carried in could be encoded in radiation that thestrings emit as they fan out.

So is the information paradox solved? “The scenario isplausible and attractive,” Witten says, “but there is nosmoking gun.” String theory is very far from being com-plete; no one can as yet do all the calculations needed toverify this solution. As Susskind puts it, “Strings can’tsolve the problems of black holes until they solve theirown first.” Spaghetti may be on the plate of theorists wellinto the next century. —Madhusree Mukerjee

Gathering String


SCIENTIFIC AMERICAN June 1994 19

Sanity CheckPuzzling observations of thingsthat go lump in the night

The farther astronomers peer intospace, the more they come to ap-preciate the intricate structure of

the universe at very large scales. In1987 a group of observers inferred thepresence of a vast accumulation of mat-ter, nicknamed the ÒGreat Attractor.ÓTwo years later another team discov-ered the ÒGreat Wall,Ó an aggregation ofgalaxies at least 500 million light-yearsacross. New celestial surveys that takein larger chunks of the universe hint atstill vaster gatherings of galaxies. Theo-rists Þnd themselves hard-pressed tounderstand the origin of such enormousstructures in a cosmos that, accordingto present knowledge, started out al-most perfectly uniform. ÒThe new sur-veys are very impressive,Ósays Margaret J. Geller of theHarvard-Smithsonian Centerfor Astrophysics, Òbut thestate of our ignorance isequally impressive.Ó

Geller should know. Overthe past decade, she and anumber of colleaguesÑmostnotably John P. Huchra, alsoat the Center for Astrophys-icsÑhave produced informa-tion that has challenged themost ingenious theorizing.What the researchers do ismeasure the redshift (thestretching of light caused bythe expansion of the uni-verse) of thousands of gal-axies. The redshift in turnindicates the galaxiesÕ ap-proximate distances fromthe earth.

Those eÝorts have led toan increasingly comprehen-sive set of maps that showgalaxies located along thebubblelike surfaces of enor-mous Òvoids.Ó These compar-atively empty regions mea-sure as much as 150 millionlight-years in diameter (forcomparison, the Milky Wayis only about 100,000 light-years across). The GreatWall is more like a sheet ofgalaxies that outlines voids.

The discovery of the GreatWall has raised two crucialquestions: Are such forma-tions typical of the universeas a whole, and does theuniverse contain even largerstructures? In their search

for an answer, researchers at the Cen-ter for Astrophysics teamed up with anumber of astronomers working in Ar-gentina, Chile and South Africa. Obser-vatories in those locations can scruti-nize southern parts of the sky that areinvisible from the Whipple Observatoryin Arizona, where most of the earliermapping was done. Luis Nicolaci daCosta of the Brazilian National Obser-vatory, a former graduate student atthe Center for Astrophysics, headed thegroup that conducted the mapping ofgalaxies in the southern sky.

Nearly 3,600 galaxies appear in thislatest survey. The distribution of galax-ies in the southern sky shows a ÒgrosssimilarityÓ to that seen in the north, Gel-ler reports. For example, da Costa andhis co-workers have uncovered a sec-ond feature much like the Great Wall,which is knownÑpredictablyÑas theSouthern Wall.

Yet statistical analysis reveals that

Òthere are some diÝerences in certainmeasures,Ó according to Geller. Such dif-ferences are signiÞcant because theyimply that parts of the universe containstructures even larger than the extent ofthe current north-south sky map. Oth-erwise, every section of the universeshould, when viewed in terms of statis-tical averages, look like any other sec-tion. Da Costa and his fellow teammembers conclude that the nature ofthe ÒshellsÓ of galaxies seen in the mapvaries over a scale of 300 million light-years or so. Even larger structures maybe out there, simply too large to showup in the current study.

In the past few years, several groupsof researchers have found that the uni-verse displays another, unexpectedkind of departure from uniformity. TheMilky Way and all the galaxies aroundus seem to be rushing headlong in thedirection of the constellation Leo; thatmotion appears superimposed on the

COSMIC ROAD MAP shows the irregular distribution of roughly 11,000 bright galaxies (bluedots); the newly discovered Southern Wall runs diagonally across the lower slice of sky.

HA

RV

AR

D-S

MIT

HS

ON

IAN

CE

NT

ER

FO

R A

ST

RO

PH

YS

ICS


more general cosmic expansion associ-ated with the big bang. In 1987 Alan M.Dressler of the Observatories of theCarnegie Institution of Washington andhis six collaborators (known as the Sev-en Samurai) analyzed those motionsand concluded that they result from thegravitational pull of some vast mass,which they called the Great Attractor.

Intrigued by that Þnding, Tod R.Lauer of the National Optical Astrono-my Observatories in Tucson and MarcPostman of the Space Telescope Sci-ence Institute in Baltimore began whatthey call a Òsanity checkÓ to make surethe Great Attractor is real. The two re-searchers measured the motions of gal-axies in a region 30 times the volumeof space examined by DresslerÕs group.If the Great Attractor is just a discrete,

local feature, Lauer explains, then itshould show up as a zone of aberrantgalaxy motions embedded within a larg-er group that shows no net motion.

Lauer and Postman studied the bright-est elliptical galaxies in 119 galaxy clus-ters lying at distances of up to 500 mil-lion light-years from the earth in all di-rections. Previous work has shown thatgiant elliptical galaxies have a fairlyconsistent intrinsic luminosity, so theirapparent brightness alone betrays theirdistance. The two researchers then mea-sured each galaxyÕs redshift, which re-veals its velocity, and compared it withthe value expected for an object at thatdistance.

Over very large scalesÑa billion light-years or soÑLauer and Postman, likemost of their colleagues, expected that

the spread of matter through the cos-mos would be very even. If so, the gal-axies should appear, on average, at restwith respect to the cosmic microwavebackground, relic radiation from thetime of the big bang that continues toÞll the universe.

When he and Lauer looked at their re-sults, Postman recalls, they were Òsur-prised, to say the leastÓ: the entire groupof galaxies appeared to be ßeeing in thedirection of the constellation Virgo at aspeed of roughly 700 kilometers persecond. The boggling implication is thatsome tremendous clump of matter lo-cated beyond the edge of the surveyedregion is pulling at all the galaxies Post-man and Lauer observed ( including, ofcourse, our own Milky Way). The GreatAttractor, it seems, is only a small partof an even greater conglomeration ofgalaxies. ÒItÕs a very diÛcult measure-ment, and theyÕve done a wonderfuljob,Ó concludes P. James E. Peebles ofPrinceton University.

Such huge structures perplex the cos-mologists who try to piece together thestory of how the modern universe cameto be. Data collected by the Cosmic Back-

ground Explorer satellite showed thatthe microwave radiation left over fromthe big bang (and, by extension, thematter that was embedded in that radi-ation) is very nearly featureless. Some-how gravity pulled together lumps andblobs of gas into galaxies, stars, planetsand people. Given enough time, gravitycould magnify extremely slight irregu-larities into distinct formations. But thelatest crop of walls and attractors in-tensifies the mystery of how so muchstructure could have formed within the15-billion-year age of the universe.

Many research teams around theworld are racing to collect more obser-vations in order to test the models andlearn more about the processes thattransformed the primordial blur intothe modern, highly organized cosmos.Lauer and Postman plan to expand thevolume of their survey Þvefold. Post-man also expresses great enthusiasmfor a massive, multi-institution digitalsky survey, headed by Donald G. Yorkof the University of Chicago, which willcollect data on one million galaxies,starting next year.

Cosmologists have frequently under-estimated the baÜing complexity ofthe universe, which is increasingly ev-ident through modern telescopes. ÒI re-ally donÕt think we understand howstructure forms in the universe,Ó saysGeller in a cautionary tone. ÒIt is atough, tough problem, much harderthan people realized when I was start-ing out. Answers are not just aroundthe corner.Ó ÑCorey S. Powell


Bright Spot

Here is another progress report from the “smaller, fewer, weirder” front inquantum physics. Researchers at AT&T Bell Laboratories have formed

what may be the smallest and certainly the most evanescent laser ever. Itconsists of a gallium arsenide quantum wire in which electrons can move inonly one dimension. The next step in the technology will meet the weirdnesscriterion.

The AT&T group, headed by Loren Pfeiffer, guessed that if energy werepumped into a one-dimensional space, or “wire,” in semiconducting material,the electrons and holes would have little choice but to bind to one anotherand form particles called excitons. The excitons, which would be in an ener-getic ground state, would collapse and emit photons at a single wavelength.Pumped with energy from laser light, and more recently powered by a bat-tery, the wire laser met the workers’ expectations. As they varied the pump-ing power by two orders of magnitude, the material emitted stable, mono-chromatic red light.

Because of their size and stability, these lasers may be able to transmitmore information with less interference than can their larger, three- and two-dimensional predecessors. They would also allow photonic technology tocomplement electronic technology on the quantum scale toward which com-puting and communications devices are shrinking. Striving for weirdnessmay prove eminently useful. “Now that we finally have a quantum wire laser,”Pfeiffer says, “we can measure whether it has useful properties or not.”

Indeed, making a quantum wire laser was a major challenge. The first step,using molecular-beam epitaxy (MBE), is to lay down a crystal film only a fewatoms thick. Such a film, called a quantum well, is thinner than an electron’swavelength is long. Thus, the particle has only two dimensions in which tomove. How can a second dimension be removed from such structure?

At the end of last year, Pfeiffer’s group reported a solution to the problem.Drawing on elementary geometry, his team formed a one-dimensional elec-tron conduit by growing quantum wells, each 70 angstroms wide, at rightangles to one another. The T-shaped intersection of the films is in effect acontinuous wire, 70 angstroms wide and some 600 microns long. “Ourmethod may not be feasible for large-scale production,” Pfeiffer says. “Wewere interested in making an ideal one-dimensional quantum wire so that wecould study its laser properties first.” He may have a point: MBE has alsobeen rendered by others as megabuck evaporation.

What’s next? Weirdness, of course, in the form of a zero-dimension, quan-tum dot laser. The group plans to grow a well across one end of a quantumwire. Three perpendicular quantum wells would then intersect at a singlepoint. “One of my goals this year is to see the luminescence from a quantumdot structure,” Pfeiffer says. For such a small feat, it would be a glowingachievement indeed. —Kristin Leutwyler



La RondeWhat goes around comes aroundfor lifeÕs master molecule

Evidence is rapidly accumulatingthat a blizzard of genetic materi-al blows freely through the micro-

bial worldÑnot only between bacteriaof the same species but also betweenmembers of distantly related speciesand between bacteria and viruses. ÒInterms of the ßux of DNA, the generalimpression is that it goes anywhere andeverywhere,Ó says Julian E. Davies, a mi-crobiologist at the University of BritishColumbia. And although the geneticmaterial of multicellular plants and an-imals tends to be better buttoned up,the exchange involves them, too.

Recent research has revealed howsome of this promiscuity may comeabout. Since the 1920s bacteria havebeen known to exchange genetic mate-rial among their own kind. One method,conjugation, is the bacterial version ofsex: genes are transferred from onebacterium to another through a specialtube. In 1958 Joshua Lederberg shareda Nobel Prize for investigations thatmade use of bacterial conjugation.

In the 1980s conjugation began to attract more than just scholarly atten-tion when researchers found clues thatgenes were spreading between species.In 1985 Patrick Trieu-Cuot proved thatgenes were indeed moving between dis-tantly related bacteria by showing thatneomycin- and kanamycin-resistancegenes in three diÝerent species werevirtually identical. Often bacterial genesare transmitted on plasmids: small, par-asitic loops of DNA that are distinctfrom the bacterial chromosome. Somestriking Þndings have come from Abi-gail A. Salyers of the University of Illi-nois. She has shown that when bacteriaare exposed to the antibiotic tetracycline,they use a variety of methods, some stillmysterious, to accelerate the exchangeof genes for tetracycline resistance.

In the laboratory at least, environ-mental stress appears to enhance con-jugation across species lines. Germanworkers have found a possible explana-tion. Alfred P�hler and his colleaguesat the University of Bielefeld showedthat heat, acids, alkalis and alcohol allinhibit the action of enzymes in Coryne-

bacterium that cut up foreign DNA. Asa result, Corynebacterium subjected tosuch treatments became more accept-ing of DNA from Escherichia coli. P�h-

ler notes that if environmental stresspromotes gene exchange between bac-terial species, genes deliberately engi-neered into microorganisms mightspread more easily in nature than theydo in the laboratory.

Transformation is another mecha-nism that bacteria use to exchange DNA.It occurs when a bacterium absorbs na-ked DNA in the environment. The DNAmay have been left lying around eitherby an experimenter or by some otherorganism, possibly one that has died.Because DNA is chemically not very sta-ble outside of cells, transformation isprobably less important in nature thanis conjugation. Nevertheless, GuentherStotzky of New York University andMarilyn Khanna, now at Cornell Univer-sity Medical College, have shown thatmontmorilloniteÑa mineral betterknown as clayÑcan absorb and bindDNA in such a way that it is protected.The bound DNA can then be taken upby other bacteria.

The third major mechanism for DNAexchange in bacteria is transduction. Itoccurs when viruses that attack bacte-riaÑknown as bacteriophagesÑbringwith them DNA they have acquiredfrom their previous host. Because mostbacteriophages have a restricted num-ber of hosts, transduction probablydoes not routinely transmit genes be-tween distantly related species of bac-teria. Still, Gustaaf A. de Zoeten, chairof the botany and plant pathology de-partment at Michigan State University,says, ÒViruses are even worse than bac-teriaÑthey evolve by the exchange ofwhole functional genetic units.Ó Fungiand plants are by no means immune tothe pervasive DNA ßux. The bacteriumAgrobacterium tumefaciens has longbeen known to transfer plasmids toplants. And in 1989 Jack A. Heinemannof the University of Oregon proved thatbacterial plasmids could be transmit-ted to yeast through a process verymuch like conjugation.

Experiments reported in Science inMarch by Ann E. Greene and Richard F.Allison of Michigan State indicate thatplant viruses can combine the RNA thatconstitutes their genes with RNA cop-ied from the genes of genetically engi-neered plants. Although the situationGreene and Allison investigated wasartiÞcial, plants engineered to containuseful viral genes may be commerciallyavailable within a few years. De Zoetenbelieves Greene and AllisonÕs resultsmean more research is still necessaryto establish the safety of such plants.

So far the evidence is slight for mas-sive and long-lasting gene exchange be-tween diÝerent species of multicellularanimals or plants. But it would be un-

DNA PASSES through bridges linking individual bacteria in the process known asconjugation, shown here taking place between a ÒmaleÓ and two Òfemales.Ó Microbi-ologists have learned that conjugation also occurs between distantly related species.

L. C

AR

O S

cien

ce P

hoto

Lib

rary

/Pho

to R

esea

rche

rs, I

nc.


wise to assume that animals are com-pletely out of the loop. In 1985 Joe V.Bannister and his colleagues at the Uni-versity of Oxford found indicationsthat genes from a species of Þsh hadbeen transferred to bacteria. And genesthat are introduced into humans by vi-

ruses probably have their origins inother animals.

What are the implications of inter-speciÞc gene transfer for evolution? Al-though the phenomenon is plainly areal one, little is yet known about howoften it occurs. The standard neo-Dar-

winian picture in evolution, in whichmutation is the main engine for intro-ducing genetic novelty, has proved ex-tremely powerful over the past half acentury. But it seems evolution hassome wrinkles that even Charles Dar-win did not foresee. ÑTim Beardsley


Water, ice and steam might be the first examples thatcome to mind when describing various phases of

matter. But to a physicist, any unusual configuration ofparticles or entities may also qualify as a new state. Forexample, electrons might organize themselves into a pat-tern called a charge-density wave. Another phase is theLuttinger liquid. Although not something one can drink, itrepresents a unique collective behavior of electrons in aconducting medium.

Under normal circumstances, electrons in conductorsconstitute a Fermi liquid. They form a sea of negativecharge. In such a liquid, a single electron does not re-spond to the individual charges of other electrons presentin the material. In effect, the Fermi liquid consists of non-interacting particles, which enables an electron to roamfairly freely through the substance. This picture explainsin part how electrons in a conductor can transmit current.

During the 1960s, Joaquin M. Luttinger of Columbia Uni-versity explored situations in which electrons are forcedto interact with one another. For a simplified, one-dimen-sional case, he solved the equations that defined thisstate (a so-called many-body problem). There the mattermostly stayed until advances in technology and the dis-covery of high-temperature superconductivity led to an in-tense reexamination of the activity of electrons in solids.

Last year Charles L. Kane of the University of Pennsylva-nia and Matthew P. A. Fisher of the University of Californiaat Santa Barbara and their colleagues squeezed a verifi-able prediction from Luttinger’s calculations. At the Marchmeeting of the American Physical Society, Richard A. Webbof the University of Maryland present-ed the first experimental evidence.

“The theory is rather specific in howyou have to set the system up,” Webbobserves. The electrons must flowthrough a one-dimensional channelthat can be obstructed in the middle. Apoint contact can create this blockageby acting as an electrical vise. Research-ers simply apply a voltage to the pointcontact, which in essence pinches offthe channel and thereby reduces theconductance.

As a Fermi liquid, electrons wouldoccasionally tunnel through the ob-struction; some conductance would al-ways remain in the channel. Not so fora Luttinger liquid. At temperaturesnear absolute zero, each electron inthis state would feel the individualcharge forces from other electrons.This effect would serve to correlatetheir behavior. The correlation wouldmanifest itself as a characteristic drop

in the conductance through the point contact; eventuallyall the electrons would be reflected by the barrier.

“You would think the experiment is easy, but it’s not,”Webb says. “I worked on it on and off for two years.” Col-laborating with Frank P. Milliken and Corey P. Umbach ofthe IBM Thomas J. Watson Research Center, Webb finallycreated the Luttinger liquid in a semiconductor made ofgallium arsenide. The theory stated that the particular sig-nature of the liquid would appear only for ballistic elec-trons—that is, electrons that move in one direction with-out scattering. The source of the ballistic electrons wasthe fractional quantum Hall effect. This phenomenonrefers to the sideways drift of electrons as they movethrough a sample exposed to an external magnetic field.Xiao-Gang Wen of the Massachusetts Institute of Technol-ogy had pointed out that under the correct conditions,these electrons move ballistically.

The Luttinger liquid is not likely to find applications. Itdestabilizes at temperatures higher than one degreeabove absolute zero. Its real value may be that investiga-tors can now see how electrons truly interact with one an-other in a solid. Conventional methods of analyzingmany-body problems demand a mixture of intuition andan approximation scheme called perturbation theory.

“The beauty of the Luttinger liquid is that the electron-electron interaction can be treated exactly,” Kane explains.“It’s an example of a many-body problem that you can re-ally solve.” Webb concurs: “It is one more tool in our bagto understand the physics of small structures.” Now that’ssomething you can drink to. —Philip Yam

LUTTINGER LIQUID forms in a channel etched into a semiconductor chip. Inthis state, electrons are reßected by an electrical barrier erected by a point con-tact. In contrast, electrons in a Fermi liquid can tunnel through the obstacle.

ItÕs Just a Phase

ST

EV

EN

STA

NK

IEW

ICZ

POINT CONTACT

CURRENT

LUTTINGERLIQUID

ELECTRONS

FERMILIQUID

ELECTRONS

ONE-DIMENSIONAL

CHANNEL

ELECTRICALBARRIER

SEMICONDUCTOR CHIP


Shooting the RapidsA new environmental agencynavigates Potomac currents

Change in the natural world spansdecades, even centuries. It followsthat long-term monitoring is the

only way to identify harmful trends. Yethuman institutions operate on the ba-sis of months, or years at best. Membersof Congress run for reelection every twoor six years. Many corporate managersliveÑand dieÑby quarterly results. Ten-ured professors scramble annually forresearch grants. How, then, can exist-ing bodies identify environmental prob-lems and assess the eÝectiveness ofmeasures taken to mitigate them?

They cannot, argue the founders ofthe Committee for the National Insti-tute for the Environment (CNIE). Whatis needed, they suggest, is their epony-mous institution. The National Institutefor the Environment would be a newfederal agency that would sponsor re-search on critical environmental issues.Proponents say it could serve as an ear-ly-warning system for such ominous de-velopments as global warming, strato-spheric ozone depletion and the declineof biodiversity. Because the institutewould be governed by an independentboard, it would be relatively immune topolitical pressure.

The idea of such an organization wasconceived more than Þve years ago byHenry F. Howe of the University of Illi-nois and Stephen P. Hubbell of Prince-ton University. Recently the CNIE ap-pointed a high-proÞle president, Rich-ard E. Benedick. Benedick, a formerstate department ambassador, was theprincipal force behind the 1987 Mon-treal protocol on chemicals that harmthe ozone layer. He is currently a spe-cial adviser to the 1994 InternationalConference on Population and Develop-ment to be held in Cairo. The CNIE hasso far secured the support of morethan 6,000 scientists, numerous envi-ronmental organizations and at leastone senior government oÛcial, Secre-tary of the Interior Bruce Babbitt.

There is opposition. Robert T. Wat-son, associate director for environmentin the OÛce of Science and TechnologyPolicy, says he Òagrees completelyÓ withthe CNIE that current government re-search eÝorts are too short-term in fo-cus and poorly coordinated. But he sug-gests instead redirecting some of the$4 billion to $6 billion the governmentalready spends annually on environ-mental research. Watson maintains thata committee newly established underthe National Science and TechnologyCouncil, the Committee on the Environ-ment and Natural Resources, is a Òvir-tual agencyÓ that should achieve manyof the CNIEÕs goals.

Others wonder how a new agencywould be linked to existing institutions.Robert C. Szaro, a deputy research di-rector of the U.S. Forest Service, com-plains that the CNIE seems to lack Òanyreal recognition of what federal govern-ment scientists already do.Ó He adds: ÒIdonÕt think the CNIE supporters wouldhave the exclusive role they think theywould haveÓ in ecological research. TheNational Research Council issued a re-port last year that considered the CNIEÕsplan but came down in favor of a de-partment of the environment thatwould subsume several existing agen-cies. The Carnegie Commission on Sci-ence, Technology and Government haslikewise supported creating an agencyout of existing programs.

Benedick points out that an inde-pendent institute for the environmentwould have backers in Congress whocould ensure continued funding even ifa future administration were hostile.Furthermore, he says such an institutecould bring in funds from industry.

To move from president of a com-mittee to head of an institute, Benedickwill have to persuade 218 representa-tives and 51 senators of the wisdom ofthe CNIEÕs plan. Success, if it comes, isunlikely to be in 1994: for now, the peo-pleÕs elected oÛcials are far too busynavigating Whitewater, grappling withhealth care and courting a disgruntled,skittish electorate. ÑTim Beardsley


SCIENTISTS at the Dorset Research Center in Ontario test theacidity of Lake Muskoka, near the U.S. border. Sulfur dioxide

from burning fossil fuels has acidified many lakes in the U.S.and Canada. Long-term monitoring is needed to track changes.

SA

RA

H L

EE

N M

atrix



In 1939 a 33-year-old French mathe-matician proved that a profoundconjecture about the behavior dis-

played by prime numbers as they me-ander toward inÞnity holds true forcertain limited but crucial cases. Theachievement, which is known as theproof of the Riemann hypothesis onthe Zeta function for Þeldfunctions, is a jewel ofmodern number theory.It is all the more remark-able because its authorÞrst scribbled it down ina French military prison.

This is only one in a se-ries of extraordinary inci-dents in the life of Andr�Weil, who eventually lefthis prison cell to becomeone of the 20th centuryÕsgreatest mathematicians.Yet so isolated is mathe-matics from the rest ofhuman culture that Weil,now a professor emeritusat the Institute for Ad-vanced Study in Prince-ton, N.J., remains largelyunrecognized outside hisÞeld. When WeilÕs autobi-ography, The Apprentice-

ship of a Mathematician,

was published three yearsago, not a single non-mathematical publicationreviewed it. WeilÕs young-er sister, Simone Weil, aphilosopher and politicalactivist, is more widelyknown in spite of the factthat she died more than50 years ago.

Professional colleaguesare therefore eager to praise Weil. Theycall him the last of the great ÒuniversalÓmathematicians. They point out that hewas a founder of Bourbaki, a legendarygroup that in the guise of a ÞctitioussageÑNicolas BourbakiÑwrote a seriesof monumental treatises that broughtorder and unity to mathematics. Weilhimself navigated all the major tribu-taries of mathematicsÑnotably, numbertheory, algebraic geometry and topolo-gyÑerecting proofs and conjecturesthat, like levees, determined the futurecourse of inquiry. One of these conjec-

tures played a crucial role in the cele-brated ÒproofÓ of FermatÕs Last Theo-rem, perhaps the most famous unsolvedproblem in mathematics, announcedlast year by Andrew Wiles of PrincetonUniversity.

WeilÕs style has been as inßuential ashis speciÞc contributions. One number

theorist likens him to a medieval monkdoing work with Òtremendous simplici-ty and purity and no unnecessary orna-ment.Ó Weil Òwas always after what wasessential,Ó another agrees. Weil was re-portedly feared for his sharp tongue aswell as admired for his brilliance. Onecompatriot, comparing Weil to a violinwhose strings have been stretched tootightly, recalls that Òhe suÝered foolsvery badly.Ó The colleague suggests Weilmay have mellowed with age.

Indeed, Weil is 88 now, equipped witha hearing aid and plastic hip joints.

And during an interview at the Institutefor Advanced Study, he seems, at times,almost serene. Asked if he is botheredby the fact that so few people know ofhis work and even fewer can appreciateit, he gives a Gallic shrug. ÒWhy shouldI be?Ó he replies. ÒIn a way, that makesit more exciting.Ó

Unlike some modern purists, Weil isalso unconcerned by the growing col-laboration between mathematics and

physics (spurred in partby Edward Witten, a theo-retical physicist whose of-Þce abuts WeilÕs). ÒI havelived through a periodwhen physics was not im-portant for mathematics,ÓWeil comments. ÒNow weare coming back to a pe-riod where it is becomingimportant again, I think,and that is a perfectlyhealthy development.Ó

Yet there are ßashes ofacerbity. When asked hisopinion of WilesÕs assaulton FermatÕs Last Theo-rem, Weil jokes at Þrstthat centuries hence his-torians will think he andthe similarly named Wilesare the same person. Thenhis smile fades, and headds, ÒI am willing to be-lieve he has had somegood ideas in trying toconstruct the proof, butthe proof is not there.Also, to some extent,proving FermatÕs theoremis like climbing Everest. Ifa man wants to climbEverest and falls short ofit by 100 yards, he hasnot climbed Everest.Ó

Explaining why his au-tobiography describes his life onlythrough World War II, Weil oÝers an-other barbed response. ÒI had no storyto tell about my life after that,Ó he says.ÒSome of my colleagues have writtenso-called autobiographies, which I thinkare very boring. They consist entirely ofsaying, ÔIn the year such and such I wasappointed to such and such an institu-tion, and in such a year I proved this orthat theorem.Õ Ó

WeilÕs life, at least its Þrst half, wasalmost excessively eventful. He wasborn in Paris in 1906. Both his father, a

PROFILE: ANDR� WEIL

The Last Universal Mathematician

ANDR� WEIL: ÒAlways after what was essential.Ó

JAS

ON

GO

LTZ


physician, and his mother devotedthemselves to all aspects of culture. Byhis early teens Weil had become Òpas-sionately addictedÓ to mathematics. Hegraduated from the University of Parisin 1928, after having delivered a Ph.D.thesis that solved a 25-year-old prob-lem about elliptic curves posed by Hen-ri Poincar�.

Weil had renounced philosophy as afatuity years earlier, after he received agood grade on a philosophy test de-spite having read none of the relevanttexts. ÒIt seemed to me that a subjectin which one could do so well whilebarely knowing what one was talkingabout was hardly worthy of respect,Óhe wrote in his autobiography.

Not that he lacked other interests. Hisfascination with Indian cultureÑand inparticular the Hindu epic the BhagavadGitaÑcontributed to his decision to ac-cept a teaching position in India in1930. After two years, he became en-tangled in IndiaÕs arcane academic pol-itics and was Þred, but not before meet-ing Gandhi. Weil sipped tea with the In-dian leader as he was planning therevolt that toppled the British Raj.

On his return to France, Weil becamea professor at the University of Stras-bourg. In 1937 he married Eveline, whohad a son by a previous marriage (shedied in 1986). Two years later, as Ger-many grew increasingly belligerent, theFrench government ordered Weil to re-port for military service. Instead he ßedto Finland, which at that point the SovietUnion had not invaded. Weil admits tosome lingering ambivalence over his de-cision to avoid service. ÒMy basic idea,which was correct, I think, was that asa soldier I would be entirely useless, andas a mathematician I could be of someuse,Ó he says. ÒOf course, that was in thedays of Hitler, and I was entirely of theopinion that the world should not yieldto Hitler, but I couldnÕt see myself tak-ing part in that eÝort.Ó

Unfortunately, the young professortyping abstract symbols hour after hourin the countryside aroused the suspi-cions of the Finns, who were fearful ofa takeover by the Soviet Union. The Fin-nish police arrested Weil andÑaccord-ing to one account related to Weil sub-sequentlyÑnearly executed him beforelearning that he was merely a Frenchmathematician avoiding the draft. WeilÕstroubles did not end there. The Finnsturned him over to the French authori-ties, who promptly convicted him of de-sertion and imprisoned him again.

Weil spent six months in jail, wherehe created his theorem on the Riemannhypothesis, before being released in ex-change for agreeing to join the Frencharmy. His ability to make the most of

his incarceration provided much amuse-ment for colleagues in later years. Oncewhen Weil made an uncharacteristicmisstep during a lecture, the eminentmathematician Herman Weyl suggest-ed that Weil return to prison so hecould work out the problem.

After the Germans routed the Frencharmy, Weil ßed to England. He eventu-ally made his way with his wife andstepson to the U.S., where he begansearching for a job. Weil was alreadysuÛciently Þlled with self-regard thathe was chagrined when the only insti-tution that initially oÝered him a paidposition was Lehigh University in Penn-sylvania. On leaving Lehigh after twounhappy years in what they felt was anintellectual wasteland, he and his wifevowed never to utter its name again.Henceforth they would call it Òthe un-mentionable place.Ó In his autobiogra-phy, Weil uncharitably recalls Lehigh asa Òsecond-rate engineering school at-tached to Bethlehem Steel.Ó

In 1947, after a stint in Brazil, Weilmoved to the University of Chicago,where he resumed his work on Bourba-ki. The project had begun in the mid-1930s, when Weil and half a dozenFrench colleagues, concerned aboutwhat they felt was the lack of adequatetexts on mathematics, vowed to writetheir own. They decided that rather thanpublishing under their own names, theywould invent a pseudonymous Þgure-head: Nicolas Bourbaki, an eminentprofessor who hailed from the (alsoÞctitious) eastern European nation ofPoldavia.

At Þrst, few people beyond their im-mediate circle guessed the true identityof Bourbaki. As the group churned outvast treatises on virtually every Þeld inmathematics, however, doubts grew. In1949 Ralph Boas proclaimed in an arti-cle in the Encyclopaedia Britannica year-book that Bourbaki was a pseudonymand did not exist. Weil wrote a letter, inhigh dudgeon, denying the accusation.BourbakiÕs members then began circu-lating rumors that Boas did not exist.

Although younger mathematicianshave continued to perpetuate the lega-cy of Bourbaki, its inßuence has waned.Weil himself, who resigned from thegroup in the late 1950s, thinks Òin someways the inßuence has been good. Insome ways it has not been good.Ó Per-haps the most important contributionof Bourbaki was to carry out a famousproposal made by the great Germanmathematician David Hilbert in 1900that mathematics be placed on a moresecure foundation. ÒHilbert just saidso, and Bourbaki did it,Ó Weil declares.BourbakiÕs emphasis on abstraction andaxiomatics was sometimes carried too

far, but Weil emphasizes that it was notBourbaki itself but its followers whoperpetrated these crimes.

Weil dismisses the argument of somephilosophers that a celebrated theoremproved by Kurt G�del in the 1930sshows that attempts to systematizemathematics are ultimately futile. ÒItÕsa perfectly good mathematical proof,Óhe says. ÒThe philosophical importanceis something else that does not interestme.Ó So averse is Weil to philosophizingthat he even claims to be an agnostic onthe old question of whether mathemat-ical truths are discovered or invented.In his autobiography, Weil describesÒthe state of lucid exaltation in whichone thought succeeds another as if mi-raculously, and in which the uncon-scious (however one interprets thatword) seems to play a role.Ó Yet he de-nies that such inspiration might stemfrom an external or even divine source.Tapping his forehead, he exclaims, ÒIthink itÕs there!Ó

In 1958 Weil came to the Institute forAdvanced Study, where he kept prob-ing for deep links between arithmetic,algebra, geometry and topology. TheseuniÞcation eÝorts spawned what hasbecome arguably the most vital Þeld ofinquiry in modern mathematics. Al-though he oÛcially retired from the in-stitute in 1976, Weil still goes to his of-Þce almost every day. There he pursuesan old passion, the history of mathe-matics. He is currently helping to editthe works of two previous French giants,Jacques Bernoulli and Pierre de Fermat.

The last universalist confesses he hasdiÛculty following recent developmentsin mathematics: ÒMathematics haspassed me by, which is as it should be,of course.Ó Although he thinks comput-ers can be useful tools, he rejects thesuggestion that they may become cru-cial for constructing proofs as mathe-matics becomes more complex. He con-tends that the use of computations incertain proofsÑsuch as the famousfour-color theoremÑis only a tempo-rary crutch. ÒIÕm sure when somethingis proved by computers it will later beproved without computers.Ó

On the other hand, Weil doubts wheth-er any human can ever again have agrasp of all of mathematics. One prob-lem, he says, may be that there are toomany mathematicians, especially goodones. ÒWhen I was much younger, Ithought there was a danger that math-ematics would be stißed by the abun-dance of mediocre mathematics beingproduced. And now I am inclined tothink that its greatest danger is that toomuch good mathematics is produced.Things are going too fast. Nobody cankeep up with it all.Ó ÑJohn Horgan

34 SCIENTIFIC AMERICAN June 1994 Copyright 1994 Scientific American, Inc.

Twenty-Þve years ago, on July 20,1969, Neil A. Armstrong took theÞrst footsteps on the surface of

the moon. That event marked a politi-cal and technological victory for the U.S.in its cold war rivalry with the U.S.S.R.In the years that followed, the Sovietgovernment insisted that the SovietUnion had never planned a lunar land-ing. Hence, it argued, the contest tosend humans to the moon was a one-sided exercise. The reality is otherwise;recently declassiÞed information fromthat era and testimony of key partici-pants in the Soviet space program un-der Khrushchev and Brezhnev provethat the moon race was indeed real.

New evidence reveals that personalrivalries, shifting political alliances andbureaucratic ineÛciencies bred failureand delays within the Soviet lunar-land-ing program. In contrast, the AmericaneÝort received consistently strong po-litical and public support. The NationalAeronautics and Space Administrationand its contractor teams also beneÞtedfrom a pool of skilled and highly moti-

vated workers and managers. Despitean early Soviet lead in human space exploration, these factors, along withmore generous and eÝective allocationof resources, enabled the U.S. to win thecompetition to be Þrst to the moon.

Soviet capability in space became clearto the world in October 1957, when theU.S.S.R. lofted Sputnik 1, the Þrst artiÞ-cial satellite. Two years later the Sovietslaunched a probe that returned close-up images of the lunar surface. And onApril 12, 1961, cosmonaut Yuri A. Gaga-rin became the Þrst human in space.Soviet oÛcials cited each accomplish-ment as evidence that communism wasa superior form of social and economicorganization. The Soviet advantage inspace rocketry underlined fears in theU.S. that a missile gap existed betweenit and its adversary, an issue that Ken-nedy belabored in the 1960 presiden-tial campaign.

At Þrst, the shape that a U.S.-Soviet space race might take was not clear. Indeed, if President Dwight

D. Eisenhower had had his way, theremight not have been one at all. Eisen-hower rejected the idea that spectacu-lar space achievements had anything todo with the fundamental strength of a country; he consistently refused toapprove space programs justiÞed onpurely political grounds. In July 1958,however, he created NASA, an agencythat brought together the resources toestablish a U.S. civilian space program.

Was the Race to the Moon Real?In 1961 President John F. Kennedy made the

goal to be first on the moon a matter of nationalhonor. But were the Soviets truly in the running?

by John M. Logsdon and Alain Dupas

GIANT ROCKETS needed to transporthumans to the moon were developed inboth the U.S.S.R. and the U.S. The SovietN-1 rocket (opposite page ) failed in eachof its four test launches before its de-velopment was canceled. The U.S. SaturnV (left ), in contrast, proceeded roughlyon schedule and successfully carriedAmericans to the moon in July 1969.


It was inevitable, perhaps, that NASA

would argue that such a programshould be ambitious.

EisenhowerÕs successor, PresidentJohn F. Kennedy, perceived a muchmore direct link between space explo-ration and global leadership. Stimulat-ed by the worldwide excitement gener-ated by the Gagarin ßight, Kennedy de-cided that the U.S. had to surpass theSoviets in human spaceßight.

On April 20, 1961, just eight days af-ter the Gagarin ßight, Kennedy askedVice President Lyndon B. Johnson, ÒIsthere any.. .space program that promis-es dramatic results in which we couldwin?Ó In particular, Kennedy inquired,ÒDo we have a chance of beating theSoviets by putting a laboratory in space,or by a trip around the moon, or by arocket to land on the moon, or by arocket to go to the moon and back witha man?Ó Johnson, whom Kennedy hadnamed his primary adviser on spacepolicy, promptly organized an intensetwo-week assessment of the feasibilityof these and other alternatives. A seriesof memoranda trace the evolving re-sponse to KennedyÕs questions.

One of the many people Johnson con-sulted was Wernher von Braun, leaderof a team of rocket engineers whom theU.S. Army had spirited out of Germanyduring the last days of the Third Reich.In a memorandum dated April 29, vonBraun told the vice president that Òwedo not have a good chance of beatingthe Soviets to a manned laboratory inspace,Ó but Òwe have a sporting chanceof sending a three-man crew aroundthe moon ahead of the Soviets,Ó andÒwe have an excellent chance of beatingthe Soviets to the Þrst landing of a crewon the moon.Ó

Von Braun judged that a lunar land-ing oÝered the U.S. the best opportuni-ty to surpass the Soviets because Òaperformance jump by a factor 10 over

their present rockets is necessary to ac-complish this feat. While today we donot have such a rocket, it is unlikelythat the Soviets have it.Ó He suggestedthat Òwith an all-out crash eÝort, I thinkwe could accomplish this objective in1967/1968.Ó

On May 8, 1961, Johnson presentedKennedy with a memorandum that re-ßected the results of his investigation.It was signed by James Webb, the NASA

administrator, and Robert S. McNama-ra, the secretary of defense. Webb andMcNamara recommended that the U.S.should set the objective of manned lu-nar exploration Òbefore the end of thisdecade.Ó They argued that Òthis nationneeds to make a positive decision topursue projects aimed at enhancingnational prestige. Our attainments area major element in the internationalcompetition between the Soviet systemand our own.Ó The two men cited lunarand planetary exploration as Òpart ofthe battle along the ßuid front of thecold war.Ó

Kennedy accepted these recommen-dations and presented them to a jointsession of Congress on May 25. Thepresident said, ÒI believe we should goto the moon.. . . No single space projectin this period will be more exciting, ormore impressive to mankind. . . . Whilewe cannot guarantee that we shall oneday be Þrst, we can guarantee that anyfailure to make this eÝort will Þnd uslast.Ó Kennedy vowed that Americanswould set foot on the moon Òbeforethis decade is out.Ó

The presidentÕs call to action strucka responsive chord in the U.S. populace.There was little public or political de-bate over the wisdom of the lunar com-mitment in the weeks following Ken-nedyÕs speech. Within months Congressincreased NASAÕs budget by 89 percent;another 101 percent increase came thenext year. Between 1961 and 1963NASAÕs payroll swelled from 16,500people to more than 28,000, and thenumber of contractors working on thespace program grew from less than60,000 to more than 200,000.

During the Þrst year after KennedyÕsannouncement, a Þerce technical de-bate erupted that threatened to delayprogress in getting to the moon. Thedispute centered on the most eÛcientstrategy for sending people to the lu-nar surface and back to the earth. Onepossibility was to use several rocketsto launch pieces of a lunar spacecraftseparately into earth orbit, where theywould be assembled and directed on tothe moon. Jerome Weisner, the presi-dentÕs science adviser, and some ele-ments within NASA initially inclined to-ward this Òearth orbit rendezvousÓ plan.

JOHN M. LOGSDON and ALAIN DUPASoften work together in the analysis ofthe worldÕs space programs. Logsdon isthe director of the Space Policy Instituteof George Washington UniversityÕs ElliottSchool of International Affairs, where heis a professor of political science and in-ternational aÝairs. His research interestsinclude space policy, the history of theU.S. space program and international sci-ence and technology policy. He is a mem-ber of the Aeronautics and Space Engi-neering Board of the National ResearchCouncil and sits on the board of directorsof the National Space Society. Dupas isan international space policy strategistfor CNES, the French space agency. He isalso a partner at L.D. Associates, a stra-tegic management consulting Þrm.

SCIENTIFIC AMERICAN June 1994 37Copyright 1994 Scientific American, Inc.

McNamara was also intrigued by thepotential military applications of earth-orbiting missions.

As they examined how best to meetKennedyÕs goal of getting to the moonbefore the end of 1969, a growing num-ber of engineers within NASA favoredanother approach, called lunar orbitrendezvous. In this scheme, the entireApollo spacecraft would be sent intospace in a single launch and would ßydirectly into orbit around the moon; asmall landing craft would detach fromthe main spaceship and ferry the astro-nauts from lunar orbit to the moonÕssurface and then back to the mothership, which would then return to earth.

Lunar orbit rendezvous dramaticallylowered the overall weight of the Apol-lo spacecraft. Consequently, the Apollomission could be carried out using asingle Saturn V rocket. After fending oÝWeisnerÕs objections, NASA oÛcials ap-proved lunar orbit rendezvous, realiz-ing that it oÝered the greatest likeli-hood of reaching the lunar surface ac-cording to KennedyÕs schedule. By theend of 1962 the U.S. was well on its wayto the moon. Not so the Soviet Union.

Until a few years ago, the Soviets of-ficially claimed that the U.S. was thesole participant in the race to the moon.The very existence of the Soviet lunarprogram was a tightly held secret. As a

result of glasnost and the collapse ofthe U.S.S.R., that situation has signiÞ-cantly changed. Several crucial playersin the space program of the 1960s(most notably Vasily P. Mishin, whoheaded the Soviet human spaceßighteÝort from 1966 to 1974) have Þnallybeen allowed to place their recollec-tions of the period in the public record.On August 18, 1989, the Soviet news-paper Izvestia printed a lengthy and un-precedentedly frank account of the na-tionÕs unsuccessful assault on the moon.And an increasing number of photo-graphs and engineering descriptions ofSoviet lunar hardware have becomeavailable to Western analysts and spaceobservers. A recent study by ChristianLardier, a French space researcher, hasbeen particularly valuable in bringingsuch information to light. The result isa much clearer picture of just how ex-tensive the Soviet lunar program was.

In June 1961, at his Þrst summitmeeting with Soviet premier NikitaS. Khrushchev, Kennedy twice raised

the possibility that the U.S. and theU.S.S.R. might travel to the moon to-gether. Khrushchev was unresponsive,at least in part because KennedyÕs lu-nar-landing announcement had caughtthe Soviet Union by surprise. The Sovi-et leadership was so conÞdent in the

countryÕs space prowess that it had notanticipated that the U.S. might actuallytry to compete in that arena.

More than three years of political de-bate dragged on before the Kremlin de-cided, and then only tentatively, thatthe Soviet Union should also have a lu-nar-landing program. During that time,powerful and entrenched leaders of theSoviet design bureaus ( industrial orga-nizations in which the Soviet technicalcapabilities for space resided) struggledfor priority and for resources related topossible lunar missions. Those conßictspresented a roadblock to establishing asingle, coordinated plan of action forreaching the moon.

Sergei P. Korolev, the top space engi-neer, headed one of the design bureaus.He was, in many ways, the Russianequivalent of Wernher von Braun. Ko-rolev had both designed the rocket usedfor all Soviet space launches to thatpoint and had managed the programsresponsible for developing most of thepayloads lofted by those rockets. Hewas also an energetic and enthusiasticproponent of space travel. Such secre-cy surrounded his work that Korolevwas identiÞed only as the ÒChief De-signerÓ; his name was not publicly re-vealed until after his death.

Unfortunately for the Soviet space ef-fort, in the early 1960s Korolev became


1961–1962, UNITED STATESJust four months after his inaugura-

tion, President John F. Kennedy vowedto land Americans on the moon by theend of the decade. That goal had beensuggested by, among others, Wernhervon Braun, the German-born rocketengineer. At the same time, the U.S.raced to catch up with the Soviets.Alan B. Shepard became the firstAmerican in space; nine months laterJohn H. Glenn matched Gagarin’s feat.

1957–1962, SOVIET UNIONThe launch of Sputnik 1, the first ar-

tificial satellite, captured the world’sattention. Subsequent flights lofteddogs into space, paving the way forhumans to follow. On April 12, 1961,Yuri A. Gagarin circled the globe in theVostok 1 spacecraft, solidifying the So-viet lead in space. “Let the capitalistcountries catch up with our country!”boasted Soviet premier Nikita S.Khrushchev.

The Space Race betweenthe U.S. and the U.S.S.R.

The competition between the U.S. and theSoviet Union in space grew out of the cold warconflict between the two nations. Early Sovietspace achievements included the first satelliteand the first human to orbit the earth. An ag-gressive, well-funded U.S. effort to place a hu-man on the moon attempted to negate thepropaganda value of these Soviet successes. Bythe mid-1960s the Soviets had initiated a se-cret, parallel program, setting the stage for arace to the moon.

EngineerreadiesSputnik 1 forflight (1957)

Malyshkaduring pre-

flight testing(1958)


embroiled in a personal and organiza-tional conßict with Valentin P. Glushko,the head of the Gas Dynamics Labora-tory and the primary designer of Sovietrocket engines. Disputes between thetwo dated to the 1930s, when Glushkowas one of those whose testimonyhelped to send Korolev to a forced-laborcamp. The two men clashed over theconcept of the rocket engines for thenext generation of Soviet space launch-ers. Korolev wanted to use high-energyliquid hydrogen as a fuel (the choice theU.S. made for the upper stages of Sat-urn V). Glushko was only interested indesigning an engine fueled by storablebut highly toxic hypergolic compounds,such as hydrazine and nitrogen tetra-oxide, that ignited on contact.

The dispute grew so bitter that Glush-ko refused to work with Korolev in thecreation of a new rocket. Instead Glush-ko allied his laboratory with anotherdesign bureau, headed by Vladimir N.Chelomei, to compete for the lunar as-signment. ChelomeiÕs group had devel-oped military missiles but had no expe-rience with rockets for outer space. Onthe other hand, one of ChelomeiÕs dep-uties was KhrushchevÕs son, Sergei. Thatfamily link offered a great advantage ina system where such personal connec-tions were often all-important. Chelomeihad ambitions to expand his bureauÕs

works into what had been KorolevÕs turf.On major technical issues such as

space exploration, the Soviet leadershiprelied on recommendations from theSoviet Academy of Sciences. Mstislav V.Keldysh, the president of the academy,was given the task of advising the gov-ernment on the technical merits ofcompeting proposals for future eÝortsin space. Keldysh and his associatestook the path of least political resis-tance and did not fully support eitherKorolev or his competitors until afterKhrushchev was removed from power.

From late 1961 on, ChelomeiÕs designbureau devoted most of its attentionnot to landing on the moon but to send-ing cosmonauts on a ßight around themoon without even going into lunar or-bit. This mission was to use a UR-500rocket (later known as Proton), derivedfrom one of ChelomeiÕs failed designsfor an intercontinental ballistic missile( ICBM). Chelomei also promoted anoverly ambitious plan for a reusablerocket airplane that could reach themoon and even the other planets.

In August 1964 the Chelomei designbureau received Kremlin approval tobuild both a spacecraft and the UR-500rocket to send cosmonauts on a circum-lunar mission by October 1967, the 50thanniversary of the Bolshevik Revolution.But ChelomeiÕs apparent victory over

Korolev was short-lived. The Politburoremoved Khrushchev from power inOctober 1964.

The post-Khrushchev leadershipquickly discovered that little progresshad been made by the organization thathad been receiving the lionÕs share offunding related to possible lunar mis-sions. The Chelomei design bureau soonfell from favor, and its contract for thecircumlunar program was canceled.

Korolev, meanwhile, had not been en-tirely shut out of the Soviet space pro-gram. After his successful eÝorts in us-ing a converted ICBM to carry out theinitial Soviet forays into space, he hadbeen thinking about the design of anew heavy-lift space launcher, which hehad designated N-1. In mid-1962 theKeldysh commission authorized the de-velopment of a version of the N-1 thatcould launch 75 tons into earth orbit,but the commission did not approveKorolevÕs plan to utilize the N-1 for alunar mission structured around earth-orbit rendezvous.

The N-1 rocket was supposed to beready for ßight testing by 1965. Be-cause he did not have access to the ex-pertise of GlushkoÕs Gas Dynamics Lab-oratory, Korolev had to Þnd an alterna-tive source of rocket engines. He turnedto the design bureau led by Nikolai D. Kuznetsov, which had previously


Yuri Gagarin about to orbit the earth (April 12, 1961)

Gagarin (center ) celebrates his achievement with Nikita Khrushchev (left) and Leonid Brezhnev (May 1, 1961)

Alan Shepard preparesfor suborbital flight

(May 5, 1961)

John Glenn enters the Mercury capsule( February 20, 1962)

John Kennedy announces lunar plansto Congress (May 25, 1961)

Wernher von Braun


worked on airplane engines. Kuznet-sovÕs group had to begin its work onspace propulsion systems basicallyfrom scratch. In the limited time avail-able, Kuznetsov was able to developonly a conventionally fueled motor ofrather little power. To achieve suÛcientlifting power for a lunar mission, theN-1 ultimately needed 30 such enginesin its Þrst stage. (The American SaturnV had just Þve Þrst-stage engines.)

After the fall of Khrushchev, the So-viet space program changed direction.Probably because it no longer fearedangering Khrushchev, by December1964 the Keldysh commission Þnallygave preliminary approval to a Korolevplan for placing cosmonauts on themoon. KorolevÕs revised lunar missionutilized a redesigned, more powerfulN-1 rocket and the same lunar orbitalrendezvous approach adopted for theApollo mission. In May 1965 the Sovietgovernment created the Ministry ofGeneral Machine Building to oversee thenationÕs space program; the ministrygave KorolevÕs lunar mission its high-est priority. The oÛcial plan called fora Þrst landing attempt in 1968, in thehope that the U.S.S.R. could still beatthe U.S. to the moon.

Just as the Soviet eÝort was gainingmomentum, disaster struck. In January1966 Korolev died unexpectedly during

simple surgery, robbing the Soviet spaceeÝort of its most eÝective and charis-matic leader. KorolevÕs successor, VasilyMishin, had neither KorolevÕs politicalstanding nor his ability to lead. Contin-uing struggles with various governmentministries and other design bureausslowed progress. Chelomei continuedto push an alternative lunar-landingscheme. To make matters worse, the re-vised N-1 launcher proved insuÛcient-ly powerful, so still more time was lostin another redesign.

Not until November 1966 did the Kel-dysh commission give a Þnal go-aheadto the lunar-landing project. A jointgovernment-party decree supportingthe project was issued the followingFebruary, but still the Soviet govern-ment allocated only limited resourcesto it. By then the date for an initial lu-nar-landing attempt had slipped intothe second half of 1969.

The U.S. was well aware of the Sovietdecision to proceed with the N-1 but forseveral years remained unsure of thekind of mission for which it would beused. In 1964 U.S. intelligence satellitesobserved the construction of a launch-pad for a large new booster and record-ed the building of a second such pad in1967. In a March 1967 national intelli-gence estimate (declassiÞed in 1992),the Central Intelligence Agency suggest-

ed that Òdepending upon their view ofthe Apollo timetable, the Soviets mayfeel that there is some prospect of theirgetting to the moon Þrst, and they maypress their program in the hopes of be-ing able to do so.Ó

After 10 successful launches of thetwo-man Gemini spacecraft during 1965and 1966, NASA seemed well preparedto move on to Apollo test ßights lead-ing to a lunar landing in 1968. Then, onJanuary 27, 1967, the program receiveda tragic setback. An electrical Þre brokeout in the Apollo 204 spacecraft (laterrenamed Apollo 1) during a countdownrehearsal on the launchpad. All threecrewmen perished. Although criticslashed out at NASA, the agency neverfaltered. With limited congressional andWhite House intervention, NASA swift-ly took the investigation into its ownhands and identiÞed and Þxed the prob-lems that had caused the Þre. By theend of 1967 the space agency had set anew schedule for Apollo that called foran initial attempt at a landing by mid-1969, approximately the same targetdate as that of the Soviet program.

The U.S. and U.S.S.R. were alsolocked into a second contest : tosee which country could Þrst

reach the vicinity of the moon. Afterthe end of the Khrushchev era, the new


Valentin Glushko,primary designer

of Soviet rocket engines

Sergei Korolev, Òchief designerÓ of rockets (right), with Gagarin

James Webb(left), withLyndonJohnson

Jerome Weisner

1962–1967, UNITED STATESAfter an intense dispute between Jerome

Weisner, the presidential science adviser, andNASA managers, the agency in 1962 finalizedits plan for the Apollo program to the moon.Under the guidance of NASA administratorJames Webb, and with the strong backing ofPresident Lyndon B. Johnson, the missionproceeded quickly. Meanwhile NASA contin-ued to lag in feats such as a space walk,which the Soviets had accomplished threemonths earlier. NASA received a serious blowin 1967, when a cabin fire during a count-down rehearsal killed three Apollo astronauts.

1962–1967, SOVIET UNIONPersonal conflicts hampered the Soviet lu-

nar-landing program. Sergei P. Korolev con-ceived of a huge rocket, the N-1, that wouldtransport cosmonauts to the moon. Korolev’splan was delayed by his clash with Valentin P.Glushko. After his death in 1966, Korolevwas replaced by Vasily P. Mishin, who keptthe beleaguered N-1 program alive. The Sovi-et space program also experienced technicalsetbacks, including a 1967 reentry mishapthat killed the cosmonaut on the first flight ofthe new Soyuz spacecraft.


Soviet leadership of Leonid I. Brezhnevand Alexei N. Kosygin asked Korolev todesign a circumlunar mission similarto that of the now canceled Chelomeiproject. The Soviets still hoped to carryout such a ßight in October 1967. Afternearly a year of often acrimonious ne-gotiations, Korolev and Chelomei inSeptember 1965 agreed on a plan thatwould use the Chelomei UR-500 boost-er, supplemented by a Korolev upperstage being developed for the N-1 rock-et and a two-cosmonaut version of thenew Soyuz spacecraft being designedby the Korolev bureau.

Although the Þrst few test ßights ofthe UR-500 booster in 1966 were suc-cessful, there were a series of seriousproblems with subsequent launches. Inaddition, the Þrst ßight of the Soyuz

spacecraft in April 1967 had a landingfailure that killed the cosmonaut onboard. Those setbacks made an Octo-ber 1967 ßight around the moon im-possible. Even so, tests during 1967and 1968 led to the successful Zond 5mission of September 1968, in whichthe UR-500 launched a modiÞed Soyuz

spacecraft carrying living organisms,including several turtles, on a coursethat took it around the moon and thensafely back to the earth. The ßight of aSoviet cosmonaut around the moonseemed imminent.

At the time of the Zond 5 mission, theU.S. had no oÛcially scheduled ßightto the lunar vicinity until well into 1969.The reality was rather diÝerent, howev-er. By mid-1968 development of the re-designed Apollo command-and-servicemodule, which would carry astronautsinto orbit around the moon and backto the earth, was on schedule for a Þrstorbital test ßight in October. But theseparate lunar landing module, intend-ed to place astronauts on the moonÕssurface, was months behind schedule.It seemed unlikely that the lunar mod-ule would be ready for an earth orbitaltest until February or March 1969.

George M. Low, deputy director ofNASAÕs Manned Spacecraft Center inHouston, recognized that the delay intesting the lunar module presented areal possibility that the U.S. might notmeet the end-of-the-decade deadlineoriginally set by Kennedy. On August 9,1968, Low therefore made a bold pro-posal : he suggested inserting an ad-ditional ßight into the Apollo launchschedule, one in which a Saturn Vwould send the command-and-servicemodule carrying a three-man crew intoorbit around the moon.

Such a mission obviously carried sub-stantial risks. It meant sending astro-nauts to the vicinity of the moon muchearlier than had been planned, and it

would be only the second ßight of theApollo spacecraft since its redesign af-ter the 1967 Þre. Moreover, the SaturnV had been launched only twice, andthe second launch had uncovered sev-eral major problems. But LowÕs strate-gy would allow NASA to gain the expe-rience of managing a mission at lunardistance many months earlier than hadbeen planned. The additional ßightwould greatly increase the probabili-ty of meeting the Apollo schedule. Itwould also improve the likelihood thatthe U.S. would reach the vicinity of themoon before the U.S.S.R. did.

LowÕs plan gained rapid acceptancewithin NASA, encountering only tempo-rary resistance from NASA administra-tor Webb and George Mueller, the headof NASAÕs Manned Spaceßight Program.In a little over a week the agency re-vised its entire Apollo schedule, creat-ing a new mission just four months be-fore it would lift oÝ. The dramatic na-ture of that ßight remained secret untilafter the October Apollo 7 mission, inwhich the command-and-service mod-ule performed ßawlessly. On November11, NASAÕs leaders formally sanctionedthe Apollo 8 ßight to the moon.

The Soviets, meanwhile, were strug-gling to keep up. In October 1968 a re-designed Soyuz spacecraft carrying onecosmonaut was successfully tested in


Vasily Mishin,KorolevÕs successor

Apollo 204 cabin after the fatal fire(January 27, 1967)

UR-500(Proton)

launch

Edward S. White II takesthe first American spacewalk (June 3, 1965)

Soyuz spacecraft


earth orbit. The Zond 6 mission, whichone month later sent a similar but un-manned spacecraft around the moon,did not fare so well. The spacecraft de-pressurized on reentry. If it had carrieda crew, they would have died.

Nevertheless, the Soviets made prep-arations for launching a circumlunarZond ßight carrying two cosmonauts inearly December. Both Mishin and thecrew agreed to take the substantial risksinvolved, because by then they knewthat the U.S. intended to send humansinto orbit around the moon later thatmonth. This launch presented the Sovi-ets with perhaps their Þnal opportuni-ty to beat the Americans to the moon,but they did not take advantage of thatchance. Just days before the scheduledtakeoÝ, the Soviet leadership canceledthe mission, presumably because theyjudged it too perilous.

During the Þnal weeks of training fortheir mission, the Apollo 8 crew mem-bers were well aware of when a Sovietcircumlunar mission could be launched.In a conversation with one of us (Logs-don), Mission Commander Frank Bor-man recalls breathing a sigh of relief asthe last possible date passed, and he re-alized that his own ßight to the moonhad not been preempted.

Apollo 8 entered lunar orbit on Christ-mas eve, 1968, all but ending the race

to the moon. Furthermore, its accom-plishments opened the way for the his-toric Apollo 11 mission seven monthslater, when Neil Armstrong planted theAmerican ßag in the lunar soil.

After the triumphs of Apollo 8 andApollo 11, the Soviet lunar program fad-ed into oblivion. But the Soviets did notgive up on the moon immediately. Twomore, unmanned Zond missions ßewaround the moon, one in 1969 and onein 1970. Shortly thereafter the Sovietleadership canceled the circumlunarprogram as it became clear that it hadbeen totally overshadowed by Apollo.

The Soviet lunar-landing programsuÝered a more ironic fate. The Þrst at-tempt to launch the N-1, in February1969, failed one minute into ßight. Thesecond launch attempt on July 3, just13 days before Apollo 11 lifted oÝ forthe moon, ended in an explosion onthe pad that destroyed much of theboosterÕs ground facilities and haltedthe Soviet lunar-landing program fortwo years. N-1 launches in July 1971and November 1972 also failed.

If they could not be Þrst, the Korolevdesign bureau leaders reasoned, theycould still be best. Led by Mishin, theyreorganized the program around theconcept of extended stays on the moonthat would be longer than the brief vis-its made by the crews of the six Apollo

missions. By early 1974 Mishin believedthat he and his associates had iden-tiÞed the sources of earlier problemsand were on the brink of success. Butin May 1974, Mishin was replaced ashead of the design bureau by Glushko,the man who more than a decade earli-er had fought with Korolev over thechoice of the N-1 propulsion system.

In one of his Þrst acts, Glushko ter-minated the N-1 program and destroyedthe 10 remaining N-1 boosters. Mishinargued that at least the two N-1s al-most ready for launch should be test-ed, but to no avail. Rather than contin-ue with the lunar program to which ithad devoted substantial resources formore than a decade, Glushko and hissuperiors chose the almost pathologi-cal response of destroying most of theevidence of its existence. The Soviet hu-man spaceßight program from the ear-ly 1970s on has concentrated entirelyon long-duration ßights in earth orbit.

Once astronauts had establishedan American presence on themoon, the U.S. lunar program

also soon wound down. The sixth andlast Apollo landing mission left themoon in December 1972. By then thelunar eÝort had clearly met the goalsthat Kennedy had set out in 1961.

Was the race to the moon worth win-


Lunarlander,

designed to fit atop

the N-1

N-1 rocket being readied for testing

1967–1972, UNITED STATESNASA recovered swiftly after the Apollo

fire. But George M. Low, director of the Apol-lo program, worried about delays affectingthe lunar lander. At his urging, NASA changedits launch schedule so that the first crew-car-rying test flight of the Saturn V rocket (Apol-lo 8 ) went into orbit around the moon onDecember 24, 1968. Then on July 20, 1969,the Apollo 11 lunar module made its historictouchdown on the surface, ending the raceto the moon. Five more Apollo landings fol-lowed before the U.S. lunar program taperedoff in 1972. George Low

1967–1974, SOVIET UNIONThe giant N-1 rocket never performed

properly. On its second test launch, the N-1exploded, wiping out its launch facilities.Glushko assumed control of N-1 develop-ment in 1974. He promptly canceled the pro-gram and dismantled the existing rockets.Pieces of the N-1 found ignominious duty asstorage sheds. Many associated pieces ofhardware, including a lunar lander and asemiflexible lunar space suit, were destroyedor placed into museums.

Earthrise over the moon, seen from Apollo 8(December 24, 1968)


ning? In our judgment, that questioncan be answered only in light of the cir-cumstances under which the competi-tion occurred. The moon race was acold war undertaking that should beevaluated primarily in foreign policyterms. On those grounds, it was an im-portant victory. The Apollo programundoubtedly aided AmericaÕs globalquest for political and military leader-ship during the 1960s. The lunar land-ing constituted a persuasive demon-stration of national will and technolog-ical capability for the U.S.

Likewise, the failure of the Soviet lu-nar program was more than a publicrelations defeat. In 1961, as the race tothe moon began, many people in theU.S. (and around the world) thoughtSoviet centralized planning and man-agement systems would allow the na-tion to pursue vigorously its long-rangegoals in space. The dissipation of theSoviet UnionÕs lead in space during the1960s tarnished the image of socialistcompetence and diminished Sovietstanding in world aÝairs.

Throughout his brief presidency,Kennedy was ambivalent about thecompetitive aspects of the space race.In his inaugural address, he suggestedto the Soviet Union that Òwe should ex-plore the stars together.Ó Shortly afterbeing sworn in, he asked NASA and the

state department to draw up proposalsfor enhanced U.S.-U.S.S.R. space coop-eration. Those proposals arrived at theWhite House on the day of GagarinÕsinitial orbital ßight, an event that con-vinced Kennedy that the U.S. had to as-sume leadership in space. Yet on Sep-tember 20, 1963, in an address to theGeneral Assembly of the United Na-tions, he still asked, ÒWhy should manÕsÞrst ßight to the moon be a matter ofnational competition?Ó

KennedyÕs dream of cooperation be-tween the two space superpowers is atlast on the verge of becoming a reality.On December 15 of last year Vice Pres-ident Al Gore and NASA administratorDaniel S. Goldin signed agreementswith their Russian counterparts for a

series of joint space activities. That col-laboration will culminate in an interna-tional space station, which will be builtaround U.S. and Russian capabilitiesbut will include contributions from Eu-rope, Japan and Canada. The stationwill begin operation soon after the turnof the century.

For 30 years, cold war rivalry was thelifeblood of both U.S. and Soviet pro-grams of human spaceßight. If the ad-venture of space exploration is to con-tinue into the 21st century, it will almost surely depend instead on wide-spread cooperation. The space stationmay serve as the harbinger of a newkind of foreign policy, one that bringsthe nations of the world together in thepeaceful conquest of space.


FURTHER READING

THE DECISION TO GO TO THE MOON:PROJECT APOLLO AND THE NATIONALINTEREST. J. Logsdon. MIT Press, 1970.

THE HEAVENS AND THE EARTH: A POLITI-CAL HISTORY OF THE SPACE AGE. Wal-ter A. McDougall. Basic Books, 1985.

THE SOVIET MANNED SPACE PROGRAM.Phillip Clark. Orion Books, 1988.

APOLLO: THE RACE TO THE MOON.Charles A. Murray and Catherine BlyCox. Simon & Schuster, 1989.

LÕASTRONAUTIQUE SOVIETIQUE. Christian

Lardier. Armand Colin, Paris, 1992.POURQUOI NOUS NE SOMMES PAS ALLES

SUR LA LUNE? V.-P. Michine (with M.Pouliquen). C�padu�s �ditions, Tou-louse, 1993.

SPACEFLIGHT. Periodical published byBritish Interplanetary Society, 27Ð29South Lambeth Road, London, EnglandSW8 1SZ.

QUEST. A quarterly magazine on the his-tory of spaceßight. P.O. Box 9331,Grand Rapids, MI 49509.

Soviet lunar space suit

N-1 pieces used asstorage sheds

Neil Armstrong on the moon (July 20, 1969)Apollo 11 crew (May 1969)

Astronaut and cosmonaut on boardApollo-Soyuz ( July 17, 1975)

1975–PRESENTIn 1975 the U.S. and the U.S.S.R. con-

ducted a rendezvous between a Soyuzand an Apollo spacecraft. That eventset a precedent for the current plan tocombine most U.S. and Russian humanspaceflight activities, leading to an in-ternational space station by 2002. Thestation could open a new chapter inthe collaborative exploration of space.

«


Throughout this century, physicshas made use of two quite diÝer-ent descriptions of nature. The

Þrst is classical physics, which accountsfor the motion of macroscopic objects,such as wheels and pulleys, planets andgalaxies. It describes the continuous,

usually predictable cause-and-eÝect re-lationships among colliding billiardballs or between the earth and orbitingsatellites. The second description isquantum physics, which encompassesthe microscopic world of atoms, mole-cules, nuclei and the fundamental par-ticles. Here the behavior of particles isdescribed by probabilistic laws that de-termine transitions between energy lev-els and govern tunneling through ener-

gy barriers. Because quantum mechan-ics is the fundamental theory of nature,it should also encompass classical phys-ics. That is, applied to macroscopic phe-nomena, quantum mechanics shouldreach a limit at which it becomes equiv-alent to classical mechanics.

Yet until recently, the exact nature ofthis transition had not been fully eluci-dated. Now that goal is within reach.Atomic systems have been created that

behaveÑfor a short periodÑaccordingto the laws of classical mechanics. Re-searchers fabricate such systems by ex-citing atoms so that they swell to about10,000 times their original size. On sucha scale the position of an electron canbe localized fairly closely; at least itsorbit no longer remains a hazy cloudthat represents only a probable loca-tion. In fact, as the electron circles thenucleus, it traces an elliptical path, justas the planets orbit the sun.

The importance of understanding theclassical limit of an atom takes on newmeaning in the light of modern technol-ogy, which has blurred the distinctionsbetween the macroscopic and micro-scopic worlds. The two domains hadremained largely separate; a scientistwould use classical mechanics to pre-dict, say, the next lunar eclipse and then


The Classical Limit of an Atom

By creating ultralarge atoms, physicists hope to study how the odd physics of the quantum world becomes

the classical mechanics of everyday experience

by Michael Nauenberg, Carlos Stroud and John Yeazell

MICHAEL NAUENBERG, CARLOS STROUD and JOHN YEAZELL combine theoretical andexperimental expertise in exploring the classical limit of the atom. Nauenberg, who re-ceived his Ph.D. in physics from Cornell University, directs the Institute of NonlinearScience at the University of California, Santa Cruz. Besides his focus on nonlinearphysics, he also studies the history of Western science and mathematics during the 17thcentury. Stroud received his physics doctorate from Washington University. Currently aprofessor of optics and physics at the University of Rochester, Stroud divides his timeformulating fundamental theories in quantum optics and then testing them in the labo-ratory. Yeazell received his Ph.D. in physics under StroudÕs tutelage Þve years ago. As afellow at the Max Planck Institute for Quantum Optics in Garching, Germany, he devoteshis time to the study of quantum chaosÑthat is, quantum systems whose classical ana-logue acts chaotically. This fall he will join the faculty of Pennsylvania State University.

CLASSICAL ORBITAL MOTION canemerge from a quantum-mechanical ob-ject called a wave packet, which deÞnesthe probable location of an electron.The series of plots shows how the local-ized wave packet traces an elliptical or-bit around the point where the nucleusresides (white dots ). Note that the wavepacket has begun to disperse after com-pleting one revolution.


switch to quantum calculations to in-vestigate radioactive decay. But engi-neers now routinely construct comput-er chips bearing transistors whose di-mensions are smaller than one micron.Such devices are comparable in size tolarge molecules. At the same time, anew generation of microscopes can seeand even manipulate single atoms. Find-ing the best way to exploit these tech-nologies will be aided by the under-standing we obtain from studies of theclassical limit.

The profound diÝerences between thequantum world and the classical do-main emerged around the turn of thecentury. Experiments by such great sci-entists as Ernest Rutherford, the NewZealandÐborn physicist who worked atthe University of Cambridge, establishedthat the atom consists of a pointlikepositive charge that holds negativelycharged electrons. To early investiga-tors, this arrangement mirrored the so-lar system. Indeed, the force that holdsthe electrons to the nucleusÑcalled theCoulomb forceÑvaries with the inversesquare of the distance, as does the grav-itational force.

This simple planetary model did notprove satisfactory. According to classi-cal electromagnetic theory, any electriccharge moving in a closed orbit mustradiate energy. Thus, the electron in an

elliptical orbit should quickly expendits energy and spiral into the nucleus.All matter would therefore be unstable.Furthermore, the radiation that an elec-tron emitted as it plunged into the nu-cleus would have a continuous spec-trum. But experiments indicated thatelectrons emit radiation in ßashes, yield-ing a spectrum of discrete lines.

The Danish physicist Niels Bohr re-solved some of these diÛculties byaugmenting the classical physics of theplanetary model of the atom with a setof constraints. They were based on a the-ory about the nature of radiation Þrstintroduced by the German physicistMax Planck, who found that radiationis emitted in discrete units (the energy

of which depends on the fundamentalparameter now known as PlanckÕs con-stant, h ).. Bohr retained the notion ofclassical orbits but assumed that onlycertain discrete values of energy andangular momentum were permitted.An integer, called a principal quantumnumber, characterized each energystate that an electron could occupywhile associated with a nucleus. For ex-ample, the ground state was numberedone, the Þrst excited state numberedtwo, and so on. Other quantum num-bers describe a particleÕs angular mo-mentum, which according to BohrÕs the-ory would occur only in integer multi-ples of PlanckÕs fundamental constant.Electrons could make transitions be-tween orbits in the form of Òquantumjumps.Ó Each jump gave oÝ a distinctfrequency of light, which equaled thediÝerence in energy between the twoorbits divided by PlanckÕs constant. Thefrequencies predicted in this way agreedcompletely with the observed discretespectra of light emitted by hydrogen.

Bohr also postulated a rule that iden-tiÞed the classical limit of his quantumtheory. This rule is named the corre-spondence principle. It states that forlarge quantum numbers, quantum the-ory should merge into classical mechan-ics. This limit corresponds to physicalsituations in which the classical actionis much larger than PlanckÕs constant.Therefore, it has become customary torefer to the classical limit as the scale at

which PlanckÕs constant vanishes. BohrÕscorrespondence principle has remainedas a basic guideline for the classical lim-it of quantum mechanics, but as we shallsee, this principle, while necessary, is notsuÛcient to obtain classical behavior.

The Bohr-Rutherford model success-fully explained the characteristics of hy-drogen. But it produced diÛculties andinconsistencies when applied to the be-havior of more complicated atoms andto the properties of molecules. The Ger-man physicist Werner Heisenberg sur-mised that to make further progress, a

quantum theory of atoms should bebased only on directly observable quan-tities, such as the well-known spectrallines mentioned above. He believed cer-tain concepts of classical physics, suchas the electronic orbits that Rutherfordand Bohr used, had to be completelydiscarded. He wrote to his Austrian col-league Wolfgang Pauli that these orbitsdo not have the slightest physical signif-icance. Indeed, his matrix formulationof quantum mechanics did away withelectron orbits entirely. Heisenberg ac-counted for the frequency and magni-tude of the discrete spectral lines interms of PlanckÕs constant and otherfundamental values in nature. Indepen-dently, the Austrian physicist ErwinSchr�dinger derived an alternative butequivalent formulation. Following ideasof the French physicist Louis de Broglie,he represented physical systems with awave equation. Solutions to this equa-tion assigned probabilities to the possi-ble outcomes of a systemÕs evolution.

Whereas Heisenberg felt thatclassical orbits had no place inquantum theory and should

be abandoned, Schr�dinger was of adiÝerent mind. From the start he wasconcerned with the relation of the mi-croscopic to the macroscopic world.Classical dynamics, he believed, shouldemerge from his wave equation. As aÞrst step, Schr�dinger investigated avery simple kind of system, called theharmonic oscillator. This system is notexactly that of an orbiting body; it cor-responds to the up-and-down motion

of a block hanging from the end of aspring. The harmonic oscillator sharesa crucial feature of an orbit around aCoulomb or gravitational potential : pe-riodicity. Such an orbiting body repeatsits motion once each cycleÑthe periodof the earthÕs orbit is just a year. Thesuspended block also has a cycle: itcompletes one up-down action oversome unit of time.

Schr�dinger managed to extract clas-


sical behavior from his theory for a har-monic oscillator. He did so by construct-ing a solution for his equation that wasa sum of solutions that had discrete en-ergy values. Graphically, these solutionsresemble sinusoidal waves of diÝerentfrequencies. Superposing such wavesproduced a ÒGaussian wave packet,Ówhich looks like a bell-shaped curve.The remarkable property of this wavepacket was that it remained localizedaround a center that executed classical,periodic behavior. Schr�dinger, howev-er, failed to derive similar classical mo-tion for more complex cases, such asthe movement of an electron in the hy-drogen atom.

On the face of it, formulating a classi-cal wave-packet description for an elec-tron associated with an atom would notseem to be diÛcult. One would similar-ly choose appropriate atomic energystates, Þnd their wave solutions andthen superpose them. The problem liesin the way energy states are actuallyseparated. A theorem developed by theFrench mathematician Jean-BaptisteFourier indicates that only energy lev-els that are equally spaced with respectto one another can be combined to forma coherent state that moves periodical-ly. But in an atom, the adjacent energystates are not equally spaced. For exam-ple, the energy separating the groundstate from the Þrst excited state is ex-

tremely large compared with the ener-gy gaps at high quantum numbers: theÞrst gap is one million times greaterthan that separating energy states whosequantum numbers are 100 and 101. Awave packet made up of a superpositionof states near the ground state there-fore disperses shortly after its creation.Obviously, a classical atom cannot beconstructed from such states.

As Bohr noted, the key to achievingclassical correspondence is to work withhigh-energy states, which have largequantum numbers. The energy separat-ing these adjacent states is proportion-al to the inverse cube of the principalquantum number. That means, for largequantum numbers, the energy spacingsbetween adjacent states are almostequal. In this limit, the spatial localiza-tion should persist for some time, per-mitting the center of the wave packetto evolve in a classical manner. Thus,the bigger the quantum numbers used,the easier it should be to produce a rel-atively stable, classical atom.

Until recently, no experimental deviceexisted by which researchers could testthe proposition by creating a superpo-sition of excited atomic states in thelaboratory. The development of lasersthat can deliver short, powerful pulsesof light proved to be the answer. Bymeans of such devices, researchersformed the Þrst localized wave packets

in atoms during the late 1980s. Amongthe successful groups were ours at theUniversity of Rochester, Ben van Lin-den van den Heuvell and his colleaguesat the FOM Institute of Atomic and Mo-lecular Physics in Amsterdam and PaulEwart and his co-workers at the Univer-sity of Oxford.

In a typical experiment, a brief, ultra-violet pulse of laser light lasting a mere20 picoseconds (20 trillionths of a sec-ond) intersects a beam of potassiumatoms in an evacuated chamber. Potas-sium is used because it readily absorbsthe energy from our lasers, and, likehydrogen, it has one electron availablefor bonding. Each pulse excites an elec-tron from a single ground state to manyvery high states. The result is a wavepacket localized at a distance of aboutone micron from the nucleus.

Laser pulses of picosecond durationare essential because short bursts havea broad spectrum of frequencies. Thespectral width of such a coherent pulseis proportional to the reciprocal of itsduration, so that a pulse with a spec-trum wide enough to overlap many lev-els must be extremely short. Tradition-al spectroscopy relies on long pulses,which contain a narrow band of fre-quencies and so excite only one or a fewstates. In our experiments, the averagequantum number excited was 85, andabout Þve states were superposed.


We probed the characteristics of ourwave packet by measuring how it ab-sorbs energy from a second laser pulse,Þred shortly after the Þrst one. At theperigee of its orbit, the wave packet ab-sorbs the most energy. In fact, theamount of energy absorbed is suÛ-cient to tear the electron away from theatom. Thus, to map out the electronÕsorbit, we simply measured the numberof atoms that were ionized as we variedthe delay between the two laser pulses.The ionization signals correspond to theexpected oscillation of the wave packetas it periodically moves through theperigee of its orbit.

This method excites orbits of a fairlywell deÞned energy and angular mo-mentum. It does not select the orienta-tion of the orbits. Instead the wave-packet state resides in the form of astatistical ensemble of classical orbits.Each member of the ensemble possess-es the same radius and eccentricity butoccupies all possible orientations inspace. This superposition is well local-ized only in the radial dimensionÑthatis, at a particular time, its distance fromthe nucleus is about as well determinedas HeisenbergÕs uncertainty principlepermits. Hence, investigators have chris-tened this object a radial wave packet.

The motion of the radial wave packetcontains many classical elements. The

wave packet evolves from the nucleustoward the edge of a classical orbit andthen returns. The period of this oscilla-tion is just that of an electron follow-ing a classical elliptical orbit about thenucleus. Moreover, the wave packetmoves most slowly at the apogee of itscircuit and most quickly at the perigee,just as do comets and other orbitingbodies in their paths around the sun.

In forming a radial wave packet, wecreated a state that exhibits strongclassical characteristics. Our goal,

however, was to form a classical atom.In that regard, the radial wave packethas a shortcoming. Despite the classi-cal orbital period of its oscillations, thepacket follows a planetary trajectoryonly in a statistical sense. An electronin such a wave packet traces orbits thatare oriented at all angles in space. In ef-fect, the particles move about in a spher-ical shell wrapped around the nucleus[see upper left illustration on next page].Obviously, this picture is not equivalentto that of a planetary system, in whichthe major axis of the ellipse describingthe motion of a planet is (approximate-ly) Þxed in space. Furthermore, thewave packet spreads as it propagatesradially, a behavior comparable in clas-sical physics to a planet breaking up asit moves in its orbit.

Jean Claude Gay, Dominique Delandeand Antoine Bommier of the �cole Nor-male Sup�rieure in Paris and one of us(Nauenberg) recently propounded a de-tailed theory that shows how to con-struct a wave packet that is oriented ina particular direction in space. We foundthat, for large quantum numbers, a sta-tionary solution of Schr�dingerÕs equa-tion exists that amounts to an Òellipti-cal stationary state.Ó This state is unusu-al. A conventional atomic state has adiscrete energy value and a range ofangular momenta [see ÒHighly ExcitedAtoms,Ó by Daniel Kleppner, Michael G.Littman and Myron L. Zimmerman; SCI-ENTIFIC AMERICAN, May 1981]. The el-liptical stationary state, however, con-sists of a well-deÞned linear superposi-tion of these ordinary atomic statesthat centers within a spread of angularmomenta. The eccentricity of the cor-responding elliptical orbit determinesthe spread. The square of the magni-tude of the wave function gives theprobability for Þnding the electron at aparticular position. Graphically, thisprobability appears as a bump on theorbit, representing the maximum valueof the wave function [see upper right il-

lustration on next page].Classical arguments explain the pres-

ence of the bump. The quantum-me-chanical state is analogous to an en-


REACHING THE CLASSICAL LIMIT demands the excitation ofatoms by brief pulses of laser light. A green laser beamemerges from behind the right side of the partition. It ÒpumpsÓa dye laser, which then produces yellow pulses ( it appearsfaint green in the photograph on the opposite page). The non-linear crystal converts the yellow light into ultraviolet (invis-ible in photograph ). A beam splitter separates each ultravioletpulse into two parts that move along diÝerent paths. A com-

puter-controlled motor can alter the length of one path byshifting a mirror. Such adjustments allow one pulse to lag be-hind the other : a 0.3-millimeter increase produces a one-pico-second delay. The beams are recombined and directed atatoms in an evacuated chamber. The first pulse excites theatoms; the second pulse probes the result. The red and or-ange beams, used to maintain mirror alignments, and somecomponents have been omitted from the diagram for clarity.

NONLINEARCRYSTAL

TO CONVERTPULSES INTOULTRAVIOLET

COMPUTER-CONTROLLED

MOTOR

BEAM SPLITTERPUMPBEAM

EVACUATEDCHAMBER

TILT-ADJUSTMENTKNOB

ULTRAVIOLETPULSES

MIRRORS

CAVITY OFDYE LASER

CELL TOHOLD DYE

DYE LASER BEAM


semble of electrons traveling on classi-cal orbits. Because their velocity is at aminimum at apogee, the electrons willtend to bunch up there. The bunchingyields the bump on a graphical repre-sentation of the elliptical state. It repre-sents the region in which the electronis most likely to be found. Making theelliptical stationary state in the labora-tory is substantially more complicatedthan forging a radial wave packet. Ashort pulse of laser light that excites anatom is not enough. The set of statesneeded to form the elliptical state turnsout to require a superposition of manyangular momentum states rather thanmany energy states. The laser beam can-not directly excite such a superposition.An additional Þeld must be applied si-multaneously with the laser pulses. Sev-eral solutions have been proposed. Twoof us (Stroud and Yeazell) have excitedsuch a state by employing a strong ra-dio-frequency Þeld in conjunction witha short optical pulse.

Although this elliptical state incorpo-rates a deÞnite angular orientation, it isstationary. It does not evolve in time.

The Þnal step in producing a classicalstate of the atom consists of makingthe wave packet move along the ellipti-cal path [see illustrations on pages 44

and 45 ]. Although we have createdsuch a wave packet as a solution of theSchr�dinger equation on the computer,to date no one has succeeded in pro-ducing this state in the laboratory.

The theoretical wave packet we con-structed is the most nearly classicalstate we know how to make. It showsstriking classical properties but alsomaintains an underlying quantum-me-chanical nature. As the wave packetmoves around the elliptical path, itmanifests one of its most obvious quan-tum properties. On each successive or-bit, the wave packet spreads, a behav-ior akin to a classical group of elec-trons in which each particle moves at adiÝerent speed. Such a group wouldcontinue to spread indeÞnitely. But forthe wave packet, a phenomenon quitedistinct from classical behavior appears:quantum interference. This eÝect hap-pens once the wave packetÕs head catch-es up to its tail and begins to interfere

with it. Then, surprisingly, at a later,well-deÞned time, the wave packet re-constitutes itself, a behavior that doesnot have any classical analogue what-soever. In between these full revivals,the state of the electron cannot be de-scribed as a single, spatially localizedwave packet.

Indeed, windows in time exist inwhich the wave packet is localized inmore complex structures. They consti-tute miniature replicas of the originalwave packet that move classically asthey maintain uniformly spaced posi-tions on the orbit. These moments havebeen characterized as fractional, or par-tial, revivals. At a stage called the one-half revival, the wave packet has splitinto two smaller ones. Likewise, at theone-third revival, it has broken up intothree packets, and so on. A classicalparticle by deÞnition cannot sponta-neously fracture and revive in this way,but a quantum particle canÑand does.

A classical analogy can explain manyfeatures of the quantum revivals. Inparticular, they can be likened to thebunching of runners on a racetrack.The runners represent the ensemble ofelectrons we use to imitate the quan-tum state. The racetrack contains a setof discrete classical orbits that satisfyBohrÕs quantum conditions. At the be-ginning of the race, the runners line upat the startÑthat is, they are well local-ized. Each one runs in one of the quan-tized Bohr orbits. During the initial laps,the runners remain closely bunched. Butafter a few circuits, the runners havebegun to spread around the track. It isnot the quantum constraints or dis-creteness that causes this initial spread-ing. It is simply that the wave packetconsists of a collection of waves ofvarying frequenciesÑa group of run-ners moving at diÝerent speeds.

The quantum features begin to ap-pear when the racers start to clumpÑthat is, when the fastest runner catchesup to the slowest runner. Further intothe race, the quicker runners continueto pass the slower runners. Occasional-


ENSEMBLE OF CLASSICAL ORBITS (left ) is one way to describe a radial wave pack-et. The packet consists of a superposition of several energy levels; in eÝect, anelectron moves simultaneously in many orbits that surround the nucleus. A moreplanetlike behavior would have the orbits lie in one plane. Such a state, called theelliptical stationary state, has been created (right ). The bump on the left side rep-resents the most likely location of the electron.

ONE-HALF REVIVAL of a wave packet (left )Ñthat is, the for-mation of two smaller packets after the original has dis-persedÑtakes place after about 15 orbits. It is indicated by

ionization signals that appear twice as frequently (right ). Af-ter about 30 orbits, the ionization signal returns to its origi-nal value, showing that the wave packet has fully revived.

NUMBER OF ORBITS

ION

IZA

TIO

N S

IGN

AL

0 10 20 30 40

ONE-HALFREVIVAL

FULL REVIVAL


ly several runners may form a clump.Because of the particular distributionof speeds allowed by the quantum con-straints, there is a moment when therunners form two clumps on oppositesides of the track. This clumping corre-sponds to the one-half fractional revival.Quantum constraints sort the runnersinto groups, so that one pack containsall the odd-numbered runners and theother all the even-numbered runners.

As the race continues, the runnersspread out and eventually clump again,but into three groups. Finally, after manycircuits, each runner has run a full lapfarther than the next slower runner, soa full revival occurs. The number ofsuch fractional revivals depends on thenumber of runners in the race. It re-quires at least two runners to form aclump. Similarly, in the atom the num-ber of fractional revivals depends onthe number of levels in the superposi-tion. Neither the fractional nor full re-vivals would appear in this classicalmodel without the imposition of thequantum constraints that place the run-ners into discrete orbits.

Investigations into this realm ofphysics have shown that despiteHeisenbergÕs attempt to banish

them, classical orbits remain a part ofmodern quantum mechanics. But theirrole is far more subtle than even Bohrrealized. Wave packets that travel onclassical trajectories are not producedby simply letting the quantum numbersof the system become large. Rather theformation of a special coherent super-position of states that have large quan-tum numbers is necessary for a wavepacket to demonstrate two hallmarkclassical features: spatial localizationand motion along an orbital path. Theseclassical actions persist for only a limit-ed period. For longer times, the under-lying quantum dynamics manifests it-self in previously unexpected wave phe-nomena that have no classical analogy.

Such results may best be understoodin terms of theories that incorporateclassical dynamics into quantum me-chanics. Such semiclassical techniquesare invaluable because conventionalquantum-mechanical calculations arediÛcult and time-consuming, even onthe largest supercomputers. Moreover,by themselves the resulting numerical

solutions often cannot be understoodor interpreted physically.

Although semiclassical methods havebeen used for a long time, especially indescriptions of a quantum systemÕs en-ergy, they have only recently been ex-tended successfully to the time domain.They can now predict quantum behav-ior, even under nonlinear, or chaotic, cir-cumstances. For example, Eric J. Hellerof Harvard University and Steven Tom-sovic of the University of Washingtonstudied the motions of a wave packettrapped inside a Òbox.Ó They showedthat semiclassical methods describethe packetÕs chaotic motions as well asquantum calculations do. Such schemesalso promise to illuminate other topicsassociated with quantum chaos thathave received much attention lately.Among them are the microwave ioniza-tion of atoms and the behavior ofatoms in strong electromagnetic Þelds.

Of course, short intense laser pulsescan excite systems other than atoms.When a molecule is excited this way, itsatoms can form wave packets. Presum-ably an appropriate tailoring of the la-ser pulse could control the internal dy-namics of the molecule [see ÒThe Birthof Molecules,Ó by Ahmed H. Zewail;SCIENTIFIC AMERICAN, December 1990].

These techniques have also been usedto form wave packets of electrons, oreven positively charged holes, in semi-conductor quantum wells. The coher-ent oscillations of the wave packets canthen produce novel devices that cannotbe made with more conventional meansof excitation. Such devices would be bo-nuses that come packaged with the fun-damental information we seek at theclassical limit of quantum mechanics.


FURTHER READING

COHERENCE AND DECAY OF RYDBERGWAVE PACKETS. Jonathan Parker and C.R. Stroud, Jr., in Physical Review Let-ters, Vol. 56, No. 7, pages 716Ð719;February 17, 1986.

QUANTUM WAVE PACKETS ON KEPLERELLIPTIC ORBITS. Michael Nauenberg inPhysical Review A, Vol. 40, No. 2, pages1133Ð1136; July 15, 1989.

LASER EXCITATION OF ELECTRONIC WAVEPACKETS IN RYDBERG ATOMS. G. Alberand P. Zoller in Physics Reports, Vol. 199,

No. 5, pages 231Ð280; January 1991.OBSERVATION OF FRACTIONAL REVIVALSIN THE EVOLUTION OF A RYDBERGATOMIC WAVE PACKET. John A. Yeazelland C. R. Stroud, Jr., in Physical ReviewA, Vol. 43, No. 9, pages 5153Ð5156;May 1, 1991.

SEMICLASSICAL THEORY OF QUANTUMPROPAGATION: THE COULOMB POTEN-TIAL. I. M. Su�rez Barnes et al. in Physi-cal Review Letters, Vol. 71, No. 13,pages 1961Ð1964; September 27, 1993.

RUNNERS ON A TRACK can portray the wave-packet revivals. At the start (1), therunners are bunched together, representing a well-localized wave packet. Duringthe course of the race, the faster runners pull ahead (2 ); soon they begin to lap theslower competitors (3 ). Eventually two clumps of runners form (4), correspondingto a one-half revival. After many more circuits, they clump back into a singlegroup (5 ). A problem with this model is that the full revival actually takes place onthe side of the track opposite from the location of the clumped runners.

1

2

3

4

5


Despite millennia of preoccupa-tion with every facet of humanemotion, we are still far from

explaining in a rigorous physiologicalsense this part of our mental experi-ence. Neuroscientists have, in moderntimes, been especially concerned withthe neural basis of cognitive processessuch as perception and memory. Theyhave for the most part ignored thebrainÕs role in emotion.

Yet in recent years, interest in thismysterious mental terrain has surged.Catalyzed by breakthroughs in under-standing the neural basis of cognitionand by an increasingly sophisticatedknowledge of the anatomical organiza-tion and physiology of the brain, inves-tigators have begun to tackle the prob-lem of emotion. One quite rewardingarea of research has been the inquiryinto the relation between memory andemotion. Much of this examination hasinvolved studies of one particular emo-tionÑfearÑand the manner in whichspeciÞc events or stimuli come, throughindividual learning experiences, toevoke this state. Scientists, myself in-cluded, have been able to determinethe way in which the brain shapes howwe form memories about this basic, butsigniÞcant, emotional event. We call thisprocess Òemotional memory.Ó

By uncovering the neural pathwaysthrough which a situation causes a crea-ture to learn about fear, we hope to elu-cidate the general mechanisms of thisform of memory. Because many humanmental disordersÑincluding anxiety,phobia, post-traumatic stress syndromeand panic attackÑinvolve malfunctionsin the brainÕs ability to control fear,studies of the neural basis of this emo-tion may help us further understandand treat these disturbances.

Most of our knowledge abouthow the brain links memoryand emotion has been gleaned

through the study of so-called classicalfear conditioning. In this process thesubject, usually a rat, hears a noise orsees a ßashing light that is paired witha brief, mild electric shock to its feet.After a few such experiences, the rat re-sponds automatically to the sound orlight even in the absence of the shock.Its reactions are typical to any threat-ening situation: the animal freezes, itsblood pressure and heart rate increase,and it startles easily. In the language ofsuch experiments, the noise or ßash isa conditioned stimulus, the foot shockis an unconditioned stimulus and theratÕs reaction is a conditioned response,which consists of readily measured be-havioral and physiological changes.

Conditioning of this kind happensquickly in ratsÑindeed, it takes placeas rapidly as it does in humans. A sin-gle pairing of the shock to the soundor sight can bring on the conditionedeÝect. Once established, the fearful re-action is relatively permanent. If thenoise or light is administered manytimes without an accompanying elec-tric shock, the ratÕs response diminish-es. This change is called extinction. Butconsiderable evidence suggests thatthis behavioral alteration is the resultof the brainÕs controlling the fear re-

sponse rather than the elimination ofthe emotional memory. For example, anapparently extinguished fear responsecan recover spontaneously or can bereinstated by an irrelevant stressful ex-perience. Similarly, stress can cause thereappearance of phobias in people whohave been successfully treated. Thisresurrection demonstrates that theemotional memory underlying the pho-bia was rendered dormant rather thanerased by treatment.

Fear conditioning has proved an ide-al starting point for studies of emotion-al memory for several reasons. First, itoccurs in nearly every animal group inwhich it has been examined: fruit ßies,snails, birds, lizards, Þsh, rabbits, rats,monkeys and people. Although no oneclaims that the mechanisms are pre-cisely the same in all these creatures, itseems clear from studies to date thatthe pathways are very similar in mam-mals and possibly in all vertebrates. Wetherefore are conÞdent in believingthat many of the Þndings in animalsapply to humans. In addition, the kindsof stimuli most commonly used in thistype of conditioning are not signalsthat ratsÑor humans, for that matterÑencounter in their daily lives. The nov-elty and irrelevance of these lights andsounds help to ensure that the animalshave not already developed strong emo-tional reactions to them. So researchersare clearly observing learning and mem-ory at work. At the same time, suchcues do not require complicated cogni-tive processing from the brain. Conse-quently, the stimuli permit us to studyemotional mechanisms relatively di-rectly. Finally, our extensive knowledgeof the neural pathways involved in pro-cessing acoustic and visual informationserves as an excellent starting point forexamining the neurological founda-tions of fear elicited by such stimuli.

My work has focused on the cerebral

Emotion, Memory and the Brain

The neural routes underlying the formation of memories about primitive emotional

experiences, such as fear, have been traced

by Joseph E. LeDoux


JOSEPH E. LEDOUX is interested in theneural foundation of memory and emo-tion. He studies the anatomy, physiologyand behavioral organization of these aspects of mental functioning. LeDoux,who is a professor of neural science andpsychology at New York University, isthe recipient of two National Institute ofMental Health distinctions: a Merit Awardand a Research Scientist DevelopmentAward. He has also received an Estab-lished Investigator Award from the Amer-ican Heart Association.


roots of learning fear, speciÞcally fearthat has been induced in the rat by as-sociating sounds with foot shock. Asdo most other investigators in the Þeld,I assume that fear conditioning occursbecause the shock modiÞes the way inwhich neurons in certain important re-gions of the brain interpret the soundstimulus. These critical neurons arethought to be located in the neuralpathway through which the sound elic-its the conditioned response.

During the past 10 years, researchersin my laboratory, as well as in others,have identiÞed major components ofthis system. Our study began when mycolleagues at Cornell University Med-ical College, where I worked severalyears ago, and I asked a simple ques-tion: Is the auditory cortex required forauditory fear conditioning? In the audi-tory pathway, as in other sensory sys-tems, the cortex is the highest level of


ANATOMY OF EMOTION includes sev-eral regions of the brain. Shown here inthe rat (above), the amygdala, the thala-mus and parts of the cortex interact tocreate memories about fearful experi-ences associated, in this case, withsound. Recent work has located preciseareas where fear is learned and remem-bered: certain parts of the thalamus(light pink at top right ) communicatewith areas in the amygdala (light yellowat bottom right ) that process the fear-causing sound stimuli. Because theseneural mechanisms are thought to besimilar in humans, the study of emo-tional memory in rodents may illumi-nate aspects of fear disorders in people.

HIPPOCAMPUS

AUDITORYTHALAMUS

AMYGDALA

AUDITORYCORTEX

HIPPO-CAMPUS


processing; it is the culmination of a se-quence of neural steps that starts withthe peripheral sensory receptors locat-ed, in this case, in the ear. If lesions in,or surgical removal of, parts of the au-ditory cortex interfered with fear condi-tioning, we could conclude that the re-gion is indeed necessary for this activi-ty. We could also deduce that the nextstep in the conditioning pathway wouldbe an output from the auditory cortex.But our lesion experiments conÞrmedwhat a series of other studies had al-ready suggested: the auditory cortex is not needed in order to learn manythings about simple acoustic stimuli.

We then went on to make lesions inthe auditory thalamus and the auditorymidbrain, sites lying immediately belowthe auditory cortex. Both these areasprocess auditory signals: the midbrainprovides the major input to the thala-mus; the thalamus supplies the majorinput to the cortex. Lesions in both re-gions completely eliminated the ratÕssusceptibility to conditioning. This dis-covery suggested that a sound stimu-lus is transmitted through the auditory

system to the level of the auditory tha-lamus but that it does not have to reachthe cortex in order for fear conditioningto occur.

This possibility was somewhat puz-zling. We knew that the primary nerveÞbers that carry signals from the audi-tory thalamus extend to the auditorycortex. So David A. Ruggiero, Donald J.Reis and I looked again and found that,in fact, cells in some regions of the au-ditory thalamus also give rise to Þbersthat reach several subcortical locations.Could these neural projections be theconnections through which the stimu-lus elicits the response we identify withfear? We tested this hypothesis by mak-ing lesions in each one of the subcorti-cal regions with which these Þbers con-nect. The damage had an eÝect in onlyone area: the amygdala.

That observation suddenly createda place for our Þndings in an al-ready accepted picture of emo-

tional processing. For a long time, theamygdala has been considered an im-portant brain region in various forms

of emotional behavior. In 1979 Bruce S.Kapp and his colleagues at the Univer-sity of Vermont reported that lesionsin the amygdalaÕs central nucleus inter-fered with a rabbitÕs conditioned heartrate response once the animal had beengiven a shock paired with a sound. Thecentral nucleus connects with areas inthe brain stem involved in the controlof heart rate, respiration and vasodila-tion. KappÕs work suggested that thecentral nucleus was a crucial part of thesystem through which autonomic con-ditioned responses are expressed.

In a similar vein, we found that le-sions of this nucleus prevented a ratÕsblood pressure from rising and limitedits ability to freeze in the presence of afear-causing stimulus. We also demon-strated, in turn, that lesions of areas towhich the central nucleus connectseliminated one or the other of the two responses. Michael Davis and his asso-ciates at Yale University determined thatlesions of the central nucleus, as well aslesions of another brain stem area towhich the central nucleus projects, di-minished yet another conditioned re-


CLASSICAL FEAR CONDITIONING can be brought about bypairing a sound and a mild electric shock to the foot of a rat.In one set of experiments, the rat hears a sound (left ), whichhas little eÝect on the animalÕs blood pressure or patterns ofmovement. Next, the rat hears the same sound, coupled with

a foot shock (center ). After several such pairings, the ratÕsblood pressure rises at the same time that the animal holdsstill for an extended period when it hears the sound. The rathas been fear-conditioned (right ): sound alone achieves thesame physiological changes as did sound and shock together.

BLO

OD

PR

ES

SU

RE

(M

ILLI

ME

TE

RS

OF

ME

RC

UR

Y)

20

15

10

5

0

10

00 100 10 0 10 0 10 0 10 0 10

CE

SS

AT

ION

OF

MO

VE

ME

NT

(C

UM

ULA

TIV

E S

EC

ON

DS

)

BLO

OD

PR

ES

SU

RE

20

15

10

5

0

BLO

OD

PR

ES

SU

RE

20

15

10

5

0

10

0

CE

SS

AT

ION

OF

MO

VE

ME

NT

10

0

CE

SS

AT

ION

OF

MO

VE

ME

NT

2

4

6

8

2

4

6

8

2

4

6

8

TIME (SECONDS)

SOUND

ELECTRICITY


sponse: the increased startle reactionthat occurs when an animal is afraid.

The Þndings from various laborato-ries studying diÝerent species and mea-suring fear in diÝerent ways all impli-cated the central nucleus as a pivotalcomponent of fear-conditioning circuit-ry. It provides connections to the vari-ous brain stem areas involved in thecontrol of a spectrum of responses.

Despite our deeper understanding ofthis site in the amygdala, many detailsof the pathway remained hidden. Doessound, for example, reach the centralnucleus directly from the auditory tha-lamus? We found that it does not. Thecentral nucleus receives projectionsfrom thalamic areas next to, but not in,the auditory part of the thalamus. In-deed, an entirely diÝerent area of theamygdala, the lateral nucleus, receivesinputs from the auditory thalamus. Le-sions of the lateral nucleus preventedfear conditioning. Because this site getsinformation directly from the sensorysystem, we have come to think of it asthe sensory interface of the amygdalain fear conditioning. In contrast, thecentral nucleus appears to be the inter-face with the systems that control responses.

These Þndings seemed to place us onthe threshold of being able to map theentire stimulus response pathway. Butwe still did not know how information

received by the lateral nucleus arrivedat the central nucleus. Earlier studieshad suggested that the lateral nucleusprojects directly to the central nucleus,but the connections were fairly sparse.Working with monkeys, David Amaraland Asla Pitkanen of the Salk Institutefor Biological Studies in San Diego dem-onstrated that the lateral nucleus ex-tends directly to an adjacent site, calledthe basal or basolateral nucleus, which,in turn, projects to the central nucleus.

Collaborating with Lisa Stefanacciand other members of the Salk team,Claudia R. Farb and C. Genevieve Go inmy laboratory at New York Universityfound the same connections in the rat.We then showed that these connec-tions form synaptic contactsÑin otherwords, they communicate directly, neu-ron to neuron. Such contacts indicatethat information reaching the lateralnucleus can inßuence the central nucle-us via the basolateral nucleus. The lat-eral nucleus can also inßuence the cen-tral nucleus by way of the accessorybasal or basomedial nucleus. Clearly,ample opportunities exist for the later-al nucleus to communicate with thecentral nucleus once a stimulus hasbeen received.

The emotional signiÞcance of such astimulus is determined not only by thesound itself but by the environment inwhich it occurs. Rats must therefore

learn not only that a sound or visualcue is dangerous, but under what con-ditions it is so. Russell G. Phillips and Iexamined the response of rats to thechamber, or context, in which they hadbeen conditioned. We found that lesionsof the amygdala interfered with the an-imalsÕ response to both the tone andthe chamber. But lesions of the hippo-campusÑa region of the brain involvedin declarative memoryÑinterfered onlywith response to the chamber, not thetone. (Declarative memory involves ex-plicit, consciously accessible informa-tion, as well as spatial memory.) Atabout the same time, Michael S. Fanse-low and Jeansok J. Kim of the Universi-ty of California at Los Angeles discov-ered that hippocampal lesions madeafter fear conditioning had taken placealso prevented the expression of re-sponses to the surroundings.

These Þndings were consistent withthe generally accepted view that thehippocampus plays an important rolein processing complex information,such as details about the spatial envi-ronment where activity is taking place.Phillips and I also demonstrated thatthe subiculum, a region of the hippo-campus that projects to other areas ofthe brain, communicated with the lat-eral nucleus of the amygdala. This con-nection suggests that contextual infor-mation may acquire emotional signi-


BRAIN LESIONS have been crucial to pinpointing the sites in-volved in experiencing and learning about fear. When a soundis processed by the rat brain, it follows a pathway from earto midbrain to thalamus to cortex (left ). Lesions can be madein various sites in the auditory pathway to determine which

areas are necessary for fear conditioning (center ). Only dam-age to the cortex does not disrupt the fear response, whichsuggests that some other areas of the brain receive the out-put of the thalamus and are involved in establishing memo-ries about experiences that stimulate fear (right ).

AUDITORYTHALAMUS

AUDITORYMIDBRAIN

AUDITORYCORTEX

EFFECTS OF LESIONSON FEAR CONDITIONING

IMPLICATION OF LESIONS FOR FEAR-CONDITIONING PATHWAY

AUDITORY PATHWAYIN THE RAT BRAIN

CONDITIONINGDISRUPTED

CONDITIONINGDISRUPTED

EARAUDITORYNERVE

SOUND

LESION

NO EFFECT

UNKNOWN LOCATION


Þcance in the same way that otherevents doÑvia transmission to the lat-eral nucleus.

Although our experiments had iden-tiÞed a subcortical sensory pathwaythat gave rise to fear conditioning, wedid not dismiss the importance of thecortex. The interaction of subcorticaland cortical mechanisms in emotionremains a hotly debated topic. Someresearchers believe cognition is a vitalprecursor to emotional experience; oth-ers think that cognitionÑwhich is pre-sumably a cortical functionÑis neces-sary to initiate emotion or that emo-tional processing is a type of cognitiveprocessing. Still others question wheth-er cognition is necessary for emotionalprocessing.

It became apparent to us that the au-ditory cortex is involved in, though notcrucial to, establishing the fear re-sponse, at least when simple auditorystimuli are applied. Norman M. Wein-berger and his colleagues at the Univer-sity of California at Irvine have per-formed elegant studies showing thatneurons in the auditory cortex undergospeciÞc physiological changes in theirreaction to sounds as a result of condi-tioning. This Þnding indicates that thecortex is establishing its own record ofthe event.

Experiments by Lizabeth M. Roman-ski in my laboratory have determinedthat in the absence of the auditory cor-tex, rats can learn to respond fearfully

to a single tone. If, however, projec-tions from the thalamus to the amyg-dala are removed, projections from thethalamus to the cortex and then to theamygdala are suÛcient. Romanski wenton to establish that the lateral nucleuscan receive input from both the thala-mus and the cortex. Her anatomicalwork in the rat complements earlier re-search in primates.

Theodore W. Jarrell and otherworkers in Neil SchneidermanÕslaboratory at the University of Mi-

ami have shown that lesions in the au-ditory cortex disrupt fear conditioningto one of two stimuli that was pairedwith foot shock. Rabbits expressed fearresponses only to the sound that hadbeen coupled with the shock. After re-ceiving auditory cortex lesions, howev-er, the animals responded to both tones.When the auditory cortex was absentand animals had to rely solely on thethalamus and the amygdala for learn-ing, the two stimuli were indistinguish-able. This work suggests that the cortexis not needed to establish simple fearconditioning; instead it serves to inter-pret stimuli when they become moreintricate. SchneidermanÕs Þndings aresupported by research in primatesshowing that projections to the amyg-dala from sensory regions of the cortexare important in processing the emo-tional signiÞcance of complex stimuli.

Some of this work has been chal-

lenged by the intriguing studies of Da-vis and his team. They reported thatdamage to a region of the perirhinalcortexÑa transitional region betweenthe older and newer cortexÑpreventsthe expression of a previously learnedfear response. Davis argues, therefore,that the cortex is the preferred path-way to the amygdala and that thalamicprojections are not normally used dur-ing learning, unless the cortex is dam-aged at the time of learning. Our gener-al understanding of the eÝect of le-sions administered after learning hastaken place is that they interfere withlong-term memory storage or retrieval.This interpretation seems applicable toDavisÕs work as well and is suggestedby recent studies by Keith P. Corodi-mas in my laboratory. He showed thatat least part of the deÞcit can be elimi-nated by providing reminder cues.

Once we had a clear understandingof the mechanism through which fearconditioning is learned, we attemptedto Þnd out how emotional memoriesare established and stored on a molec-ular level. Farb and I showed that theexcitatory amino acid transmitter glu-tamate is present in the thalamic cellsthat reach the lateral nucleus. Togetherwith Chiye J. Aoki, we showed that it isalso present at synapses in the lateralnucleus. Because glutamate transmis-sion is implicated in memory formation,we seemed to be on the right track.

Glutamate has been observed in aprocess called long-term potentiation,or LTP, that has emerged as a model forthe creation of memories. This process,which is most frequently studied in thehippocampus, involves a change in theeÛciency of synaptic transmission alonga neural pathwayÑin other words, sig-nals travel more readily along this path-way once LTP has taken place. Themechanism seems to involve glutamatetransmission and a class of postsynap-tic excitatory amino acid receptorsknown as NMDA receptors [see ÒTheBiological Basis of Learning and Indi-viduality,Ó by Eric R. Kandel and RobertD. Hawkins; SCIENTIFIC AMERICAN, Sep-tember 1992].

Various studies have found LTP in thefear-conditioning pathway. Marie-Chris-tine Clugnet and I noted that LTP couldbe induced in the thalamo-amygdalapathway. Thomas H. Brown and PaulChapman and their colleagues at Yalediscovered LTP in a cortical projectionto the amygdala. Other researchers, in-cluding Davis and Fanselow, have beenable to block fear conditioning by block-ing NMDA receptors in the amygdala.And Michael T. Rogan in my laboratoryfound that the processing of sounds bythe thalamo-amygdala pathway is am-


Structure of the Amygdalaýhe amygdala's role in emotional behavior has long been considered important. Experiments in rodents have elucidated the structures of

various regions of the amygdala and their role in learning about and remem- bering fear. The lateral nucleus receives inputs from sensory regions of the brain and transmits these signals to the basolateral, the accessory basal and the central nuclei. The central nucleus connects to the brain stem, bringing about physiological changes.

Tý

AMYGDALASTIMULUS

HIPPOCAMPUSTHALAMUSBASO-

LATERALNUCLEUS

ACCESSORYBASALNUCLEUS

LATERALNUCLEUS

CENTRALNUCLEUS

CORTEX


pliÞed after LTP has been induced. Thefact that LTP can be demonstrated in aconditioning pathway oÝers new hopefor understanding how LTP might re-late to emotional memory.

In addition, recent studies by FabioBordi, also in my laboratory, have sug-gested hypotheses about what could begoing on in the neurons of the lateralnucleus during learning. Bordi moni-tored the electrical state of individualneurons in this area when a rat was lis-tening to the sound and receiving theshock. He and Romanski found that es-sentially every cell responding to theauditory stimuli also responded to theshock. The basic ingredient of condi-tioning is thus present in the lateralnucleus.

Bordi was able to divide the acousti-cally stimulated cells into two classes:habituating and consistently responsive.Habituating cells eventually stoppedresponding to the repeated sound, sug-gesting that they might serve to detectany sound that was unusual or diÝer-ent. They could permit the amygdala toignore a stimulus once it became famil-iar. Sound and shock pairing at thesecells might reduce habituation, therebyallowing the cells to respond to, ratherthan ignore, signiÞcant stimuli.

The consistently responsive cells hadhigh-intensity thresholds: only loudsounds could activate them. That Þnd-ing is interesting because of the roleloudness plays in judging distance.Nearby sources of sound are presum-ably more dangerous than those thatare far away. Sound coupled with shockmight act on these cells to lower theirthreshold, increasing the cellsÕ sensitivi-ty to the same stimulus. Consistently re-sponsive cells were also broadly tuned.The joining of a sound and a shockcould make the cells responsive to anarrower range of frequencies, or itcould shift the tuning toward the fre-quency of the stimulus. In fact, Wein-berger has recently shown that cells inthe auditory system do alter their tun-ing to approximate the conditionedstimulus. Bordi and I have detected thiseÝect in lateral nucleus cells as well.

The apparent permanence of thesememories raises an important clinicalquestion: Can emotional learning beeliminated, and, if not, how can it betoned down? As noted earlier, it is ac-tually quite diÛcult to get rid of emo-tional memories, and at best we canhope only to keep them under wraps.Studies by Maria A. Morgan in my labo-ratory have begun to illuminate how thebrain regulates emotional expressions.Morgan has shown that when part ofthe prefrontal cortex is damaged, emo-tional memory is very hard to extin-

guish. This discovery indicates that theprefrontal areasÑpossibly by way ofthe amygdalaÑnormally control ex-pression of emotional memory andprevent emotional responses once theyare no longer useful. A similar conclu-sion was proposed by Edmund T. Rollsand his colleagues at the University ofOxford during studies of primates. Theresearchers studied the electrical activ-ity of neurons in the frontal cortex ofthe animals.

Functional variation in the pathwaybetween this region of the cortex andthe amygdala may make it more diÛ-cult for some people to change theiremotional behavior. Davis and his col-leagues have found that blocking NMDAreceptors in the amygdala interfereswith extinction. Those results hint thatextinction is an active learning process.At the same time, such learning couldbe situated in connections between theprefrontal cortex and the amygdala.More experiments should disclose theanswer.

Placing a basic emotional memoryprocess in the amygdalic path-way yields obvious beneÞts. The

amygdala is a critical site of learningbecause of its central location betweeninput and output stations. Each routethat leads to the amygdalaÑsensorythalamus, sensory cortex and hippo-campusÑdelivers unique informationto the organ. Pathways originating inthe sensory thalamus provide only acrude perception of the external world,but because they involve only one neu-ral link, they are quite fast. In contrast,pathways from the cortex oÝer detailedand accurate representations, allowingus to recognize an object by sight orsound. But these pathways, which runfrom the thalamus to the sensory cor-tex to the amygdala, involve severalneural links. And each link in the chainadds time.

Conserving time may be the reasonthere are two routesÑone cortical andone subcorticalÑfor emotional learning.


MEMORY FORMATION has been linkedto the establishment of long-term po-tentiation, or LTP. In this model of mem-ory the neurotransmitter glutamate andits receptors, called NMDA receptors(top), bring about strengthened neuraltransmission. Once LTP is established,the same neural signals produce largerresponses (top middle). Emotional mem-ories may also involve LTP in the amyg-dala. Glutamate (red circle in top photo-graph) and NMDA receptors (red circlein bottom photograph) have been foundin the region of the amygdala wherefear conditioning takes place.

PRESYNAPTICNEURON

POSTSYNAPTICNEURON

NMDARECEPTOR

GLUTAMATE

NE

UR

AL

IMP

ULS

E (

MIL

LIV

OLT

S)

TIME (MILLISECONDS)0 10 20

AFTER LTP

BEFORE LTP


Animals, and humans, need a quick-and-dirty reaction mechanism. The thal-amus activates the amygdala at aboutthe same time as it activates the cortex.The arrangement may enable emotion-al responses to begin in the amygdalabefore we completely recognize what itis we are reacting to or what we arefeeling.

The thalamic pathway may be partic-ularly useful in situations requiring arapid response. Failing to respond todanger is more costly than respondinginappropriately to a benign stimulus.For instance, the sound of rustlingleaves is enough to alert us when weare walking in the woods without ourhaving Þrst to identify what is causingthe sound. Similarly, the sight of a slen-der curved shape lying ßat on the path

ahead of us is suÛcient to elicit defen-sive fear responses. We do not need togo through a detailed analysis of wheth-er or not what we are seeing is a snake.Nor do we need to think about the factthat snakes are reptiles and that theirskins can be used to make belts andboots. All these details are irrelevantand, in fact, detrimental to an eÛcient,speedy and potentially lifesaving reac-tion. The brain simply needs to be ableto store primitive cues and detect them.Later, coordination of this basic infor-mation with the cortex permits veriÞ-cation (yes, this is a snake) or bringsthe response (screaming, hyperventilat-ing or sprinting) to a stop.

Although the amygdala stores primi-tive information, we should not consid-er it the only learning center. The estab-

lishment of memories is a function ofthe entire network, not just of one com-ponent. The amygdala is certainly cru-cial, but we must not lose sight of thefact that its functions exist only by vir-tue of the system to which it belongs.

Memory is generally thought to bethe process by which we bring back tomind some earlier conscious experi-ence. The original learning and the re-membering, in this case, are both con-scious events. Workers have determinedthat declarative memory is mediatedby the hippocampus and the cortex. Butremoval of the hippocampus has little


CORTICAL AND SUBCORTICAL PATHWAYS in the brainÑgeneralized from our knowledge of the auditory systemÑmaybring about a fearful response to a snake on a hikerÕs path.Visual stimuli are Þrst processed by the thalamus, whichpasses rough, almost archetypal, information directly to theamygdala (red ). This quick transmission allows the brain tostart to respond to the possible danger (green ). Meanwhile

the visual cortex also receives information from the thalamusand, with more perceptual sophistication and more time, de-termines that there is a snake on the path (blue). This informa-tion is relayed to the amygdala, causing heart rate and bloodpressure to increase and muscles to contract. If, however, thecortex had determined that the object was not a snake, themessage to the amygdala would quell the fear response.

VISUAL THALAMUSVISUALCORTEX

AMYGDALA

BLOOD PRESSURE MUSCLE

HEART RATE


eÝect on fear conditioningÑexcept con-ditioning to context.

In contrast, emotional learning thatcomes about through fear conditioningis not declarative learning. Rather it ismediated by a diÝerent system, whichin all likelihood operates independent-ly of our conscious awareness. Emo-tional information may be stored with-in declarative memory, but it is keptthere as a cold declarative fact. For ex-ample, if a person is injured in an auto-mobile accident in which the horn getsstuck in the on position, he or she maylater have a reaction when hearing theblare of car horns. The person may re-member the details of the accident,such as where and when it occurred,who else was involved and how awful itwas. These are declarative memoriesthat are dependent on the hippocam-pus. The individual may also becometense, anxious and depressed, as theemotional memory is reactivatedthrough the amygdalic system. The de-clarative system has stored the emo-tional content of the experience, but ithas done so as a fact.

Emotional and declarative memoriesare stored and retrieved in parallel,and their activities are joined seamless-ly in our conscious experience. Thatdoes not mean that we have direct con-scious access to emotional memory; itmeans instead that we have access tothe consequencesÑsuch as the way webehave, the way our bodies feel. Theseconsequences combine with current de-clarative memory to form a new declar-ative memory. Emotion is not just un-

conscious memory: it exerts a powerfulinßuence on declarative memory andother thought processes. As James L.McGaugh and his colleagues at the Uni-versity of California at Irvine have con-vincingly shown, the amygdala playsan essential part in modulating thestorage and strength of memories.

The distinction between declarativememory and emotional memory is animportant one. W. J. Jacobs of the Uni-versity of British Columbia and LynnNadel of the University of Arizona haveargued that we are unable to remembertraumatic events that take place earlyin life because the hippocampus hasnot yet matured to the point of form-ing consciously accessible memories.The emotional memory system, whichmay develop earlier, clearly forms andstores its unconscious memories ofthese events. And for this reason, thetrauma may affect mental and behavior-al functions in later life, albeit throughprocesses that remain inaccessible toconsciousness.

Because pairing a tone and a shockcan bring about conditioned re-sponses in animals throughout

the phyla, it is clear that fear condition-ing cannot be dependent on conscious-ness. Fruit ßies and snails, for example,are not creatures known for their con-scious mental processes. My way of in-terpreting this phenomenon is to con-sider fear a subjective state of aware-ness brought about when brain systemsreact to danger. Only if the organismpossesses a suÛciently advanced neu-

ral mechanism does conscious fear ac-company bodily response. This is not tosay that only humans experience fearbut, rather, that consciousness is a pre-requisite to subjective emotional states.

Thus, emotions or feelings are con-scious products of unconscious pro-cesses. It is crucial to remember thatthe subjective experiences we call feel-ings are not the primary business ofthe system that generates them. Emo-tional experiences are the result of trig-gering systems of behavioral adaptationthat have been preserved by evolution.Subjective experience of any variety ischallenging turf for scientists. We have,however, gone a long way toward un-derstanding the neural system that un-derlies fear responses, and this samesystem may in fact give rise to subjec-tive feelings of fear. If so, studies of theneural control of emotional responsesmay hold the key to understandingsubjective emotion as well.


FURTHER READING

THE AMYGDALA: NEUROBIOLOGICAL AS-PECTS OF EMOTION, MEMORY AND MEN-TAL DYSFUNCTION. Edited by John P.Aggleton. Wiley-Liss, 1992.

BRAIN MECHANISMS OF EMOTION ANDEMOTIONAL LEARNING. J. E. LeDoux inCurrent Opinion in Neurobiology. Vol.2, No. 2, pages 191Ð197; April 1992.

THE ROLE OF THE AMYGDALA IN FEARAND ANXIETY. M. Davis in Annual Re-view of Neuroscience, Vol. 15, pages353Ð375; 1992.

Fruit fly

Macaque

Baboon

Rabbit

Cat

Dog

Marine snail

Fish

Rat

Pigeon

Some Species That ExhibitFear Conditioning

Emotional memories brought about by fear-conditioning experiments have been ob-

served in many animal groups. It appears thatonce a fearful memory has been established, it isrelatively permanent: changes in behavior can bebrought about by controlling the fearful responserather than by eliminating the emotional memoryitself. This continuity between findings in diversespecies suggests that brain pathways for this form of learningare similar. A fuller understanding of these mechanisms in ani-mals may lead researchers to new treatments for fear disorders,such as panic attack or phobia, in humans.

Human

Lizard


Atmospheric turbulence, which causes stars to twinkle and dis-tant objects to shimmer, has

frustrated astronomers ever since tele-scopes were invented. ÒThe only Reme-dy is a most serene and quiet Air,Ówrote Sir Isaac Newton in 1704, Òsuchas may perhaps be found on the topsof the highest Mountains above thegrosser Clouds.Ó Astronomers have fol-lowed this advice, which Newton of-fered in his Opticks, but even on thehighest peaks atmospheric turbulenceseverely limits the power of big tele-scopes such as the 200-inch Hale tele-scope at Mount Palomar in California.The launch of the Hubble Space Tele-

scope showed to what heights as-tronomers are willing to go to circum-vent turbulence.

My colleagues and I at Litton Itek Op-tical Systems in Lexington, Mass., as wellas workers at other institutions, havebeen pursuing another, earthbound, so-lution to the problem of atmosphericturbulence. Our approach, called adap-tive optics, also has its roots in the de-velopment of space technology, butnow, somewhat ironically, it is beingapplied to ground-based astronomicaltelescopes. Adaptive optics uses a de-formable mirror or similar device tocompensate, or correct, for the distor-

tion of light caused by atmospheric tur-bulence. Adaptive optics technology isimproving the ability of the next gener-ation of earthbound telescopes to re-solve Þne detail and to detect extreme-ly faint objects in the sky.

The challenge in building astronomi-cal telescopes is to obtain the clearestpossible image of a distant star, whichshould appear as a single point. Extend-ed objects such as galaxies and planetscan be regarded as collections of points.A distant star produces a sphericalwavefront that travels vast distancesthrough space until it reaches the earthÕsatmosphere, where air turbulence dis-torts it. Temperature variations associ-ated with the turbulence generate chang-es in the air density, with the result thatparts of the wavefront are slowed bydiÝerent amounts, distorting the im-age. An adaptive optics telescope seeksto reverse this eÝect by restoring thespherical shape of the wavefront.

The Þrst step is to determine howmuch each component of the wave-front is out of phase with the others.One way to that end is to divide the tele-scopeÕs mirror into a number of zonesand then measure the tilt of the wave-front in each zone. After processing byhigh-speed electronic circuits, this in-formation is used to control actuatorsthat determine the position of individ-ual areas of the mirrorÕs surface. Themirror is thereby deformed in such away that any wave component arrivinglater than another actually travels ashorter distance to the focal point. Thisprocess of measurement and adjust-mentÑa classic feedback setupÑhap-pens several hundred times a second.When the adaptive optics is workingproperly, all the components shouldarrive at the focal point in phase, tocreate a perfectly sharp image.

Radar engineers were the Þrst to de-velop the notion of breaking a wave-front into parts and then bringing theparts into correct phase. The mathemat-ical principles for compensating for dis-tortion in a wavefront are virtually the

same for optical images as for radar. Inthe early 1950s radar engineers beganto divide antennas into segments sothat the phase of the signal from eacharea could be independently adjusted.By phase shifting the wave componentsin this way, they were able to track mov-ing objects with a Þxed antenna or tofocus the beam on objects at diÝerentdistances.

The idea of applying adaptive princi-ples to optical systems was Þrst sug-gested in 1953 by Horace W. Babcock.He proposed that an electron beam beapplied to control the thickness of aliquid Þlm on a rigid mirror to compen-sate for errors in the phase of the in-coming wavefront. The components ofthe wavefront that had phases preced-ing the others were delayed by passingthem through a thicker Þlm of liquid.

BabcockÕs ingenious concept wouldhave required considerable develop-ment, and because the problem was ofconcern only to a fairly small communi-ty of astronomers, it was not pursuedfurther. The simpler idea of stabilizingimage motion with a tilting ßat platewas used on a spectrograph of the Haletelescope in 1956. In an article in Scien-

tiÞc American in June 1956, Robert B.Leighton described the use of a tip-tiltmirror to obtain high-quality photo-graphs of the planets.

Full correction of atmospheric tur-bulence, however, remained anunattained goal until the 1970s,

when the U.S. military looked into thesubject. Its interest stemmed from twosources. Pentagon scientists workingon antiballistic missile defense neededa way to focus a laser beam on a distant


JOHN W. HARDY began working onadaptive optics in 1972 and, during thenext two decades, developed the tech-nology for applications in defense andastronomy. He was awarded a bachelorÕsdegree in electrical engineering fromLondon University in 1946 and has spe-cialized in electro-optical technology. Forhis contributions to adaptive optics, Har-dy won the Goddard Award of the Soci-ety of Photo-Optical Instrumentation En-gineers in 1989 and the Michelson Med-al of the Franklin Institute in 1992. Heretired from Litton Itek Optical Systemsin 1990 but is still active as a consultantand is now developing a low-cost adap-tive optical device for small telescopes.

Adaptive OpticsTechnology developed during the cold war is giving new capabilities to ground-based

astronomical telescopes

by John W. Hardy

TELESCOPE equipped with adaptive op-tics is being tended by Robert Q. Fugateof the U.S. Air ForceÕs Phillips Laborato-ry. Adaptive optical systems sharpenthe images collected by ground-basedtelescopes by eÝectively erasing theblurring eÝects of the atmosphere.

Copyright 1994 Scientific American, Inc.Copyright 1994 Scientific American, Inc.

target while protecting the ray fromdegradation in the atmosphere. The sec-ond objective was at the time even moreurgent: the Soviet Union was launchinggreat numbers of military satellites. TheDefense Advanced Research ProjectsAgency (now ARPA) was searching forbetter methods of identifying thosespacecraft. Photographs taken withground-based satellite-tracking tele-scopes were too blurred by the atmo-sphere to yield useful images, evenwhen digitally enhanced.

I was part of a team at Litton Itek Op-tical Systems in 1972 that won a con-tract with ARPA to develop a more ef-fective approach. We decided to useadaptive optics to ÒundoÓ the distortionbefore the image was recordedÑthat is,to build a real-time atmospheric com-pensation system (RTAC).

Although the principle had beenproved in radar applications, the com-ponents of an adaptive optical systemhad yet to be built. To create such asystem, a key question had to be ad-dressed: How Þnely must the incomingwavefront be divided to achieve a satis-factory reconstruction of the originalimage? The answer determines howmany independently controlled actua-

tors are required for the deformablemirror, which in turn determines thecost and complexity of the system. For-tunately, David L. Fried, then at NorthAmerican Aviation, Inc., had provided away to Þnd the answer in 1966. Friedfound that the optical eÝects of air tur-bulence, which at Þrst appear complexand random, can be described in termsof simple wavefront shapes such as tilt,defocus and astigmatism (spherical andcylindrical curvature), which are famil-iar to all workers in optics. Furthermore,the strength of the turbulence can berepresented as a single quantity, r0. Forconventional telescopes, r0 is the diam-eter of the largest aperture that can beused before turbulence starts to de-grade the image quality. As the turbu-lence gets stronger, r0 becomes small-er. For earthbound observatories, it ty-pically ranges between Þve and 15centimeters at visible wavelengths, withan average value of 10 centimeters.

Most of the time, therefore, large tele-scopes cannot resolve objects such asdouble stars any better than can ama-teursÕ small instruments. (Astronomersuse large telescopes to collect enoughlight to enable them to record very faintobjects. There are also periods when

turbulence is quite low, enabling largetelescopes to give good resolution.)

In adaptive optics, r0 deÞnes the sizeof each zone that must be adjusted torestore the image. To achieve good com-pensation at visible wavelengths, a four-meter telescope needs a deformablemirror controlled by about 500 actua-tors. The value of r0 also depends onthe wavelength of the incident light. Inthe infrared band, at two microns, anaverage value of r0 is about 50 centi-meters, so the number of actuators re-quired by a four-meter telescope dropsto about 60. We wanted to build a pro-totypical instrument equipped with anumber of actuators suÛcient to testthe concept without the taskÕs becom-ing too complex. So we settled, ratherarbitrarily, on 21 actuators.

The only wavefront correctors avail-able in 1972 were segmented mirrorsthat had been designed for remedyingdistortion in infrared laser beams.These devices were too slow and impre-cise for our purposes. At Þrst, a crystalof bismuth silicon oxide seemed apromising alternative. We found thatwe could adjust the phase shift of lightpassing through it by applying a volt-age. But the crystal transmitted an in-


suÛcient amount of light, and its phase-correction capability was too small foratmospheric turbulence.

We considered using a ßexible mir-ror made from a thin aluminized plate,which would reßect light eÛciently andbend easily, but we struggled with theproblem of stability. Although the sur-face of a deformable mirror moves lessthan 10 microns (one hundredth of amillimeter), it must be controlled withhigh accuracyÑto a tolerance of as lit-tle as one Þftieth of a micron. My co-workers Julius Feinleib, Steven G. Lipsonand Peter F. Cone, then at Itek, foundthat by mounting a thin glass mirror ona block of piezoelectric material Þttedwith electrodes, they could control de-formations in hundreds of zones of the

mirror to the required accuracy andspeed. We called the device a monolith-ic piezoelectric mirror.

Next we addressed the problem ofmeasuring distortion in the wavefront.At that time, the standard technique forprecise measurement of optical wave-fronts was a leisurely process in whichphotographs from a laser interferome-ter were manually scanned and digi-tized. With luck, information about thewavefronts was available the next day,a far cry from the one thousandth of asecond response time needed for adap-tive optics.

Fortunately, a new method of mea-suring wavefronts, called shearing in-terferometry, was under development.Interferometers are commonly used in

optics to measure the phase of onewavefront by superimposing it on asecond wavefront of known character-istics, thus producing an interferencepattern. For adaptive optics, we need toknow only the relative phase of eachzone of the aperture with respect to itsneighbors to determine the extent towhich atmospheric turbulence has dis-turbed the wavefrontÕs shape. Shearinginterferometers accomplish this taskby displacing (ÒshearingÓ) two copiesof the same wavefront by a known dis-tance and then superimposing them.The intensity of the resulting interfer-ence pattern is proportional to the gra-dient, or slope, of the wavefront.

Conventional shearing interferome-ters, however, worked only with mono-chromatic light and produced only aÞxed interference pattern. For adaptiveoptics, we needed to make rapid wave-front measurements using broadbandwhite light from sunlit satellites. My col-league James Wyant was able to build aÒwhite lightÕÕ shearing interferometer,using a moving diÝraction grating thatproduced an interference pattern inwhich the intensity varied sinusoidally.An array of photodetectors picked upthe signal. The phase shift of this elec-trical signal when compared with aÞxed reference was exactly proportion-al to the optical wavefront slope in thecorresponding area of the aperture.This type of shearing interferometer isoptically stable and reliable and needslittle calibration. Later improvementsincreased the speed of this device sothat it could measure 10,000 completeoptical wavefronts per second, a speedsuÛcient to measure the worst atmo-spheric turbulence.

We needed one more element to com-plete the system: a fast method for syn-thesizing the individual wavefront mea-surements from each zone into a mapof the continuous wavefront across theentire optical aperture. This wavefrontreconstruction process is essential fordetermining the adjustment of the in-dividual actuators. Serial computation,the most obvious method available tous, was problematic, given the small dig-ital computers of that era, so we re-verted to analog technology. Our groupbuilt a simple electrical network ar-ranged in the same pattern as the actu-ators behind the deformable mirror.Electric currents representing the mea-sured wavefront values were applied tothe nodes in the network, which gener-ated the exact voltages needed to ad-just the corresponding actuators. Thisparallel network was extremely fast andcould be expanded to manage a largenumber of actuators without losingspeedÑa vast advantage over tech-


APPEARANCE OF STARS viewed from a great distance depends on the integrity ofthe spherical wavefronts of light they produce. If all the components of the wave-front can be focused, a star looks like a perfect point source of light (left ). Atmo-spheric turbulence, however, randomly disrupts the wavefrontÕs shape, whichcauses the components to arrive at a focal point out of phase (right ).

SPHERICALWAVE-FRONT

PERFECTIMAGE

ASTRONOMICALTELESCOPE

DISTORTEDIMAGE

TURBULENT LAYER

DISTORTED WAVEFRONT


niques that utilized serial computation.As the December 1973 test date for

our real-time atmospheric compensa-tion system approached, we grew in-creasingly concerned about whether themachine would operate stably. Each ofthe 21 actuators had its own feedbackloop, but there was unavoidable cross-coupling between loops through thedeformable mirror. In other words, awavefront correction in one zone had asmall eÝect on all the others. Our calcu-lations showed that the new systemshould have been stable, but there wasalways the possibility of some unfore-seen problem. We were concerned thatif the RTAC burst into oscillation, ascan happen with complex feedback sys-tems, it might destroy the monolithicpiezoelectric mirror that we had sopainstakingly designed. When it wasÞrst tested, we were relieved to Þndthat the system functioned perfectlyand was as stable as a rock.

The RTAC demonstrated for the Þrsttime that adaptive optics could com-pensate extended images degraded byturbulence, and its basic architecturehas been used in many subsequent sys-tems. Still, it had too few actuators tobe used on a large telescope. So in 1976,again with ARPA sponsorship, we start-ed to build a much larger machine,called the compensated image system(CIS), which employed 168 actuators.My colleagues J. Kent Bowker, RichardA. Hutchin and Edward P. Wallnerplayed important roles in the design ofthis pioneering system. We installed itin 1980 on the 1.6-meter telescope thenoperated by ARPA at Mount Haleakalain Maui, Hawaii. Once again we worriedabout stability : some members of theteam predicted that large sections ofthe mirror would lock and render theentire system useless. When it was Þrsttested in the spring of 1982 on brightstars, CIS proved to be perfectly stable.Since then, even larger adaptive opticalsystems have been built and operatedwithout stability problems.

The CIS gave us the Þrst real proof ofhow impressively adaptive optics canenhance the performance of ground-based telescopes. The results, particu-larly for double stars, were striking. Thebrighter star, with which the wavefrontsensor measures turbulence, was easilyvisible before correction, but its com-panion was just a fuzzy, dim patch.Both stars exhibited a shimmering mo-tion over several arc seconds. When thecompensation loop was switched on,each star sprang into sharp focus andremained steady on the viewing moni-tor. The increase in brightness of theimages was even more impressive thanthe improvement in sharpness.

Since these early eÝorts, new typesof deformable mirrors with more than1,000 actuators have been made. Someof these are segmented mirrors, whichconsist of many ßat plates, each mount-ed on three multiple-layer piezoelectricactuators. Segmented mirrors have thelargest capacity to compensate for se-vere turbulence. Each segment is physi-cally separate from the others, givinggreat freedom of movement. Unfortu-

nately, the individual facets require fre-quent calibration. Because of disconti-nuities from one segment to the next,segmented mirrors tend to diÝractsome of the light, which aÝects an im-ageÕs clarity. Consequently, astrono-mers generally prefer continuous face-plate mirrors. Such a mirror is made ofa ßexible aluminized glass faceplatemounted on actuators made of multi-ple layers of piezoelectric or electro-


Adaptive optics can compen-sate for distortions in the

wavefront of starlight. The op-tics first concentrates the lightentering a telescope into a nar-row beam. That beam reflectsoff a deformable mirror and asecond mirror that corrects formotion of the image. The beamis also split so that a wavefrontsensor can measure the degreeto which each component of thewavefront is distorted; that in-formation directs the compen-sating movements of the mir-rors. The compensated beam isfinally focused into a camera,which records the corrected im-age. Shown are corrected anduncorrected images of a starand of the Hubble Space Tele-scope, as seen from the earth.

ASTRONOMICALTELESCOPE

DATAPROCESSOR

WAVEFRONTSENSOR

CAMERA

COMPENSATEDIMAGE

BEAMSPLITTER

IMAGE-MOTIONCORRECTOR

DEFORMABLEMIRROR

DISTORTEDWAVEFRONT

CORRECTEDWAVEFRONT

ADAPTIVE OPTICAL SYSTEM

STAR 4968COMPENSATED

STAR 4968UNCOMPENSATED

HUBBLECOMPENSATED

HUBBLEUNCOMPENSATED

How AdaptiveOptics Works


strictive material, which expand orcontract in response to a controllingvoltage. The actuators are mounted ona rigid base plate. A continuous face-plate mirror tends to have better di-mensional stability, requires less main-tenance and provides a smoother cor-rection over the telescope aperture.Another type of deformable mirror be-ing developed is the bimorph mirror,consisting of ßat piezoelectric elementsglued to the back of a thin face sheet,which becomes bent when a voltage isapplied. An adaptive optical systemthat employs a bimorph mirror is beingengineered at the University of Hawaii.This system uses a wavefront sensorthat directly measures wavefront curva-ture, simplifying the computation need-ed to control the actuators.

At infrared wavelengths, for whichthe eÝects of atmospheric turbulenceare less severe, Òmodal correctorsÕÕ po-tentially oÝer an elegant way of com-pensating for wavefront distortion.These devices are eÝective for correct-ing tilt, defocus, astigmatism and othermore systematic distortions.

Other avenues currently under inves-tigation include the application of neu-ral networks to interpret signals fromthe optical sensor and to control thedeformable mirror. Some researchersbelieve a neural network consisting ofmany interconnected decision-makingcells can be trained to interpret theseinputs better than the algorithmicallybased networks now in use.

Although the initial application ofadaptive optics has been very success-ful, the widespread adoption of thistechnology in observational astronomyis impeded by two fundamental prob-lems. First, faint objects (a categorythat includes most of the objects of in-terest to astronomers) can be observedonly when a bright star lies in closeproximity to them. The need to makereal-time measurements of turbulenceat least as quickly as the atmosphere

changes dictates how bright the guidestar needs to be: enough photons mustbe collected within each small zone of the telescope aperture to make an accurate wavefront measurement. For an adaptive optical system working at visible wavelengths under average con-ditions, this means that at least 100photons must be detected within each 10- by 10-centimeter zone in each onehundredth of a second. To meet thatrequirement, a guide star must be ofthe 10th visual magnitude or brighter.On average, only three stars of thismagnitude are found in each degree-square patch of sky.

This limitation would be acceptablewere it not for a second fundamentalproblem: adaptive compensation is ef-fective only over an extremely small an-gle of sky, called the isoplanatic angle,which at visible wavelengths is usuallyless than five arc seconds across. Overa larger area the turbulence varies toowidely from that measured by the wave-front sensor to get a consistently clearimage. Thus, only the center would becorrectly compensated, and the imagewould become increasingly blurred to-ward the edges. Because no more thana tiny area of the sky surrounding eachguide star can be compensated, mostof the sky is unavailable to adaptive op-tics using natural guide stars.

Researchers are working on twoways of circumventing these lim-itations. The Þrst is to operate at

longer (infrared) wavelengths, at whichthe optical eÝects of turbulence aremuch less severe. Because the value ofr0 at infrared wavelengths is Þve to 12times greater than at visible wave-lengths, each correction zone can bemade correspondingly larger. Also, be-

cause disturbances in the wavefronttake longer to change in large zones,more time is available for collectinglight. As a result, dimmer stars can beused as guide stars. Furthermore, theisoplanatic angle is larger at longerwavelengths. Consequently, the areaover which adaptive compensation iseÝective also increases. Taken togeth-er, these factors allow a visible guidestar to be used to sharpen infrared ob-servations over a much larger fractionof the sky than is possible at visiblewavelengths.

The Þrst infrared system, known asCome-On, was developed jointly by theEuropean Southern Observatory andresearchers in France in the late 1980s.It has been tested on the observatoryÕs3.6-meter telescope at La Silla, Chile,with excellent results.

The second approach is to use lasersto generate artiÞcial beacons, or guidestars. Researchers at the MassachusettsInstitute of TechnologyÕs Lincoln Labo-ratory and the U.S. Air ForceÕs PhillipsLaboratory fortuitously provided uswith this powerful new way to measureatmospheric turbulence during theirwork for the Strategic Defense Initia-tive. In the 1980s they were studyinghow to Þre a laser weapon so that itcould deliver as much energy as possi-ble to a target above the atmosphere.Because laser beams suÝer the samekind of distortion at visible wavelengthsthat aÜicts light from a distant star,the principles of adaptive optics can beapplied. In 1982 M.I.T. researchers be-gan employing a 69-actuator version ofthe CIS, known as the atmospheric com-pensation experiment (ACE), to correctdistortion in a laser beam projectedfrom the ground into space. In one ex-periment, the space shuttle Discovery


TO VIEW DIM OBJECTS, astronomersuse brighter stars to gauge atmosphericturbulence (a ). This technique, howev-er, works only if a bright star is in thesame part of the sky as the objectviewed; if they are far apart, their lightexperiences diÝerent degrees of turbu-lence (b ). Because few stars are brightenough to serve as references, thistechnique is useful only in a small partof the sky. One solution is to employ alaser beam directed through the atmo-sphere as an artificial reference star (c ).With an array of laser beacons, as-tronomers can illuminate an entire fieldof view (d ). A nearby star is still re-quired to point the telescope.

TURBULENTLAYER

aBRIGHTSTAR

SHAREDREGION OFTURBULENCE

b

DIFFERENTREGIONS OF

TURBULENCE

DIMOBJECT


carried a retroreßector to bounce a laserbeam back to the earth, where research-ers used it to measure atmospheric dis-tortion. In later tests, retroreßectors in-stalled on rockets reached altitudes of600 kilometers. By feeding this infor-mation into a deformable mirror, theworkers were able to ÒpredistortÓ a sec-ond beam so that it passed through theatmosphere and focused on a smalltarget on the rocket. The ACE adaptiveoptics equipment has since been usedsuccessfully for astronomical work onthe 60-inch telescope at Mount WilsonObservatory in California.

For astronomical telescopes, laserscreate artiÞcial guide stars in theupper atmosphere, either by pro-

ducing backscatter from air moleculesat altitudes of 10 to 40 kilometers,which is known as Rayleigh scatter, orby stimulating ßuorescence in a natu-rally occurring layer of sodium vaporthat lies at about 90 kilometers. Be-cause a laser beacon is much closer tothe telescope than is a natural star, thedevice generates a cone-shaped beam(rather than a virtually cylindricalbeam), which passes through only afraction of the turbulent atmosphericlayer before it reaches the telescopeaperture. This eÝect is more seriouswith the lower-altitude Rayleigh bea-cons and requires the use of more thanone laser beacon.

In 1983 Robert Q. Fugate of PhillipsLaboratory demonstrated the feasibili-ty of using laser guide stars for wave-front measurement. Researchers at theM.I.T. Lincoln Laboratory created theÞrst complete adaptive optical systemdepending on laser guide stars, knownas SWAT (for short wavelength adap-tive techniques). Between 1988 and

1990 at the U.S. Air ForceÕs Maui Opti-cal Site at Mount Haleakala, they usedpulsed dye lasers operating at a wave-length of 0.512 micron to generate la-ser beacons at altitudes of four to eightkilometers. Compensation of turbu-lence was proved by comparing imagesof natural stars with and without theadaptive correction, and the experimentshowed that two laser beacons yieldedbetter results than one.

A diÝerent type of wavefront sensor,the Shack-Hartmann sensor, is general-ly used with laser beacons because itcan handle either continuous or pulsedlight sources. First used by Roland V.Shack of the University of Arizona in1971, it employs an array of small lens-es covering the optical beam, each pro-ducing an image of the guide star.Wavefront gradients are determined bymeasuring the image displacement ineach zone.

Laser beacons should in principlepermit the use of adaptive compensa-tion on any celestial object, howeverfaint, at any wavelength that can passthrough the atmosphere. But the eÝec-tiveness of laser beacons is still limitedby the need for a natural star by whichto point the telescope. Laser beaconsare useless for pointing because theyare not Þxed in the sky but vary in ab-solute position according to the eÝectof turbulence on the laser beam. Be-cause of the need for a pointing star,adaptive optics can cover only roughly30 percent of the sky at visible wave-lengths. At infrared wavelengths, thesky coverage approaches 100 percent.Several organizations, including PhillipsLaboratory, the University of Chicagoand Lawrence Livermore National Labo-ratory, are pursuing the developmentof adaptive optics using laser beacons.

One of the unsolved problems ofadaptive optics is how to create sharpimages across large Þelds of view. Forinstance, no one has yet been able toobtain a fully compensated image ofJupiterÕs disk. The problem is that thedisk is about 40 arc seconds across,which encompasses about 50 diÝerentisoplanatic patches, or areas in whichatmospheric turbulence diÝers signiÞ-cantly. One often discussed approachis to employ multiple deformable mir-rors in conjunction with an array oflaser guide stars. Each mirror would ineÝect serve as a three-dimensional cor-rector by compensating for turbulenceover a range of altitudes in the atmo-sphere. Multiple wavefront measure-ments covering the Þeld of view wouldbe made with the guide stars.

After 25 years of development, most-ly for defense purposes, adaptive op-tics is now Þnding wider scientiÞc uses,primarily in ground-based astronomy.Most large astronomical telescopes cur-rently being planned or constructed in-clude adaptive optics in their design.

A frequently raised question is whywe continue to build large ground-basedtelescopes when we can escape atmo-spheric limitations completely by goinginto space. The answer is that space tele-scopes are far more expensive to buildand maintain than those on the earth,even when the cost of adaptive opticsis included.

Space- and ground-based astronomyshould be considered complementaryrather than competitive. Space tele-scopes have the advantage of beingable to perform at wavelengths that arecut oÝ by atmospheric absorption, de-tecting ultraviolet and x-ray radiation,for example. Ground-based telescopes,which at present have much largerapertures and baselines, are better suit-ed for work at longer wavelengths, forwhich turbulence eÝects are more easi-ly compensated. With each technologydoing what it does best, our view of theuniverse will become clearer than ever.


FURTHER READING

ACTIVE OPTICS: A NEW TECHNOLOGYFOR THE CONTROL OF LIGHT. J. W. Har-dy in Proceedings of the IEEE, Vol. 66,No. 6, pages 651Ð697; June 1978.

ADAPTIVE OPTICS REVISITED. Horace W.Babcock in Science, Vol. 249, pages253Ð257; July 20, 1990.

PRINCIPLES OF ADAPTIVE OPTICS. RobertK. Tyson. Academic Press, 1991.

ADAPTIVE OPTICS FOR ASTRONOMY:PRINCIPLES, PERFORMANCE AND APPLI-CATIONS. J. M. Beckers in Annual Re-view of Astronomy and Astrophysics,Vol. 31, pages 13Ð62; 1993.

c

LASERBEAM

ARTIFICIALSTAR

d

MULTIPLELASERBEAMS


For more than half a century, ar-chaeologists have believed thatÒtrueÓ civilization in Peru began

with the Chavin Culture of what theycall the Early Horizon, which extendedfrom 1100 to 250 B.C. Yet on trianglesof desert plain 350 kilometers north ofLima, nestled between coastward-point-ing Þngers of the Andes Mountains, liethe remains of cities at least 700 yearsolder. Since 1985 we have been uncov-ering through a series of excavationstwo of these sites. They lie in the Cas-ma Valley and testify to the existenceof a complex civilization: there are tem-

ple-topped mounds as much as 30 me-ters high, administrative enclosures,large-scale irrigation systems and hun-dreds of houses for both rich and poor.

Not until we sent wood and charcoalsamples from these sites for radiocar-bon dating in 1980 did it become clearthat the sites had been occupied dur-ing the Initial Period between 2200 and1100 B.C., the time when ceramics, weav-ing and large-scale irrigation agricultureÞrst appeared on the Peruvian coast.(This period corresponds roughly to thetime of the Middle Kingdom in Egypt, afew hundred years after the building of

the great pyramids.) Increasing knowl-edge about this civilization has yield-ed a new appreciation of its critical rolein shaping subsequent cultures in the region.

Our understanding of the culture atthese sites is largely the result of thepeculiar climate conditions that prevailthere. SigniÞcant amounts of rain lastfell in Casma in 1983 and, before that,in 1925 and 1891. (Indeed, many of thestructures we excavated had no roofs,as the climate makes them unneces-sary.) As a result, the preservation ofartifacts is nothing short of incredible:buried textiles, wood and even fragileleaves are often found in pristine con-dition. Furthermore, knowing that cul-tivators practiced irrigationÑlarge-scaleagriculture would have been impossi-ble otherwiseÑgives archaeologists ad-ditional clues in piecing together thevanished cultureÕs social structure.

While it ßourished, Pampa de las Lla-mas-Moxeke, located along the south-ern branch of the Casma, was a bustling


Early Andean CitiesSome 3,800 years ago Pampa de las Llamas-Moxeke

and Taukachi-Konkan were carefully laid-out urban centers that housed many hundreds of people

by Shelia Pozorski and Thomas Pozorski

SHELIA POZORSKI and THOMAS POZORSKI have carried out archaeological Þeldworkin Peru since 1970. Both are associate professors of anthropology at the University ofTexas Pan American in Edinburg, Tex. Shelia Pozorski is especially concerned with pre-historic subsistence and its relation to early societal development. She received her B.A.from Harvard University and her Ph.D. in anthropology from the University of Texas atAustin in 1976. Thomas Pozorski also earned his B.A. from Harvard and his Ph.D. fromthe University of Texas at Austin in 1976. His interests lie in the development of com-plex societies within the Andean area.


center that housed about 2,500 peoplewithin an area of two square kilome-ters. Two massive mounds dominatethe site: Moxeke on the south and Hua-ca A on the north. The two structuresface each other across a series of sym-metrical plazas, forming a central axisthat establishes the orientation for al-most all other public architecture.

Flanking this axis are more than 110intermediate-size administrative build-ings arranged in parallel rows. Most faceinward toward the siteÕs center, andmost are small mounds, like miniature,simple versions of Huaca A. There isalso evidence of Òurban renewal.Ó Low-status houses appear to have beenrazed to gain space for new rows of intermediate-size mounds, several ofwhich were never Þnished.

Moxeke, a U-shaped mound about160 meters wide, 170 meters long and30 meters high, probably served as atemple from which ceremonies wereperformed for large crowds in the pla-zas surrounding it. Huaca A (119 me-ters wide, 136 meters long and about15 meters high) is reached via much

smaller plazas at either end. Because ofthe downward ground slope from northto south, however, the tops of the twomounds appear to be about the samelevel. To the northeast of Huaca A is asunken circular plaza as well. This ar-chitectural feature is common to manysites of the period.

Planning is evident in all aspectsof the siteÕs architecture. All thepublic buildings exhibit internal

symmetryÑthe two halves of the Mox-eke mound are mirror images, for ex-ample. In keeping with its accessibilityfrom plazas at either end, Huaca A hasa four-way internal symmetry.

Huaca A is replete with what seem tobe control devices: gates, barriers andwalled compounds. These structuresare probably related to its function as ahuge warehouse. Pampa de las Llamas-Moxeke administered much of the sur-rounding area, and smaller settlementssupplied labor and goods to its inhabi-tants. The moundÕs rooms appear tohave held at least 4,400 cubic meters ofagricultural products and luxury goods.

The rows of smaller mounds, mean-while, appear to have served as lower-level bureaucratic centers. They proba-bly contained oÛces and storage areasused by mid-level oÛcials involved in


HUACA A mound at Pampa de las Llamas-Moxeke would have glistened white inthe desert sun 3,800 years ago. The structure stood 15 meters high and stored atleast 4,400 cubic meters of valuables and commodities. It was part of a city, builtaccording to plan (right ), that housed more than 2,000 people. The political entityto which Pampa de las Llamas-Moxeke belonged contained more than half a dozencities scattered over 1,000 square kilometers. Civilization in the region evolvedover a period of roughly 4,000 years (below ).

YEARS BEFORE PRESENTCITY OCCUPIED

5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0

COTTONPRECERAMIC

PERIOD

EARLYHORIZON

EARLYINTERMEDIATE

PERIOD

MIDDLEHORIZON

LATEINTER-

MEDIATEPERIOD

LATEHORIZON

INITIALPERIOD

MOXEKE

HUACA A

MODERNFARMED AREA

ADMINISTRATIVEMOUNDS

LOW-STATUSHOUSING

METERS

0 100 200


the awesome task of acquiring, moni-toring and redistributing the vast quan-tity of commodities stored and protect-ed within Huaca A.

Residential architecture occupies asurprisingly small percentage of the to-tal area of Pampa de las Llamas-Mox-eke, but we have found both low- andhigh-status houses. Public oÛcials prob-ably occupied the Þner homes, whichhad plastered stone walls as well as in-ternal storage rooms and niches fortheir personal wealth. In contrast, mostof the populationÑmost likely farmers,laborers and artisansÑlived in small,crowded, irregular clusters of roomswith mud-coated cane or wood walls.These dwellings left few remains exceptfor low, stone-wall footings.

For more than 40 years after theCasma Valley sites were found,no one realized just how old they

were. Investigations of the area beganin 1937, when renowned Peruvian ar-chaeologist Julio C. Tello excavated themound of Moxeke. He uncovered sever-al huge adobe friezesÑhigh-relief sculp-tures depicting human Þgures and mas-sive headsÑon the front of the moundand extending slightly around to thesides. These friezes had been brilliantlypainted in red, blue, green, black andwhite, and their discovery brought thevalley to prominence in studies of earlyAndean civilization. Tello attributed thespectacular sculpture to the highlandChavin Culture, which characterized theEarly Horizon.

In 1980 we examined the site, alongwith Þve others, as part of our eÝort toconstruct a precise chronography of themany ÒearlyÓ Casma Valley sites. We

also needed to decide where to excavateon a larger scale. From the start, Pampade las Llamas-Moxeke was full of sur-prises. Although the Initial Period ismarked by the widespread introductionof pottery, we found relatively smallamounts of very simple ceramics. Mostwere jars without necks, probably mod-eled after the gourd containers used be-fore the invention of Þred clay vessels.Simple woven textiles reßected the be-ginnings of weaving, but alongside themlay elaborate twined textiles (twisted to-gether by hand rather than on a loom)characteristic of prepottery cultures.The remains of cultivated plants wereabundant, but none of our more than50 test pits yielded maize, a staple oflater Andean populations.

All these data suggested that Pampade las Llamas-Moxeke was a very earlysite; nevertheless, we were unpreparedfor the results of carbon 14 dating ofcharcoal and wood samples. Nine ini-tial samples (and 15 collected duringsubsequent research) yielded calibrat-ed dates between 2000 and 1500 B.C.Ñhundreds of years before the beginningof the Early Horizon.

In 1985 we returned to start excava-tions in earnest. We concentrated onthe northern two thirds of the site be-cause the southern third has been af-fected by both later prehistoric reoccu-pation and modern agriculture. Wesampled large mounds, intermediate-size mounds and enclosures, and smalland irregular house structures, seekingclues to the activities carried out withinand around them and to their role inthe site as a whole.

Tello had already worked extensivelyat Moxeke, so we chose Huaca A. The

mound looked easy before we got un-der way: its four-way symmetry, readi-ly visible from the surface, meant thatexcavation of only one quadrant wouldelucidate the structure of the entire edi-Þce. Even before excavation, the wallswere fairly straightforward to trace onthe surface.

Our Peruvian workmen, 15 to 25 innumber, began by moving huge fallenwall boulders weighing up to 150 kilo-grams. They used the wheelbarrow wefurnished to shade the water jug, pre-ferring to carry the stones atop a pad-ded cushion on their backs in much thesame manner as the original buildersof Huaca A.

More than two meters down, weÞnally encountered a ßoor.The volume of the fallen rub-

ble implies that the walls of the largercentral rooms were originally at leastÞve meters high. In the wallsÕ upperreaches, we could see bare stones stillheld in place by mud mortar, their ßatfaces forming the wall face. The verytop courses of the walls were probablymade of cone-shaped adobe bricks,whose lighter weight would have easedconstruction.

The granite stones making up thewalls were covered with a thick layer ofcoarse mud plaster, replete with thelong Þnger strokes characteristic of pre-historic masons. The Þne Þnishing plas-ter, rich in clay, had ßaked oÝ the up-per parts of the walls, but lower downit was well preserved, as was the Þnelyplastered ßoor. Traces of pigment revealthat the walls and ßoors were paintedwhite. The entire mound would haveglistened in the desert sun.

Ultimately we were able to deÞne thewalls for the north quadrant of HuacaA; from this, we reconstructed the con-Þguration of the entire mound summit.The whole ediÞce consists of repetitions(in various sizes) of a basic constructionunit: a square room with rounded exte-rior corners and two rounded and twosquare interior corners. The entrance ofeach room has a raised threshold; theupper walls contain large niches 1.25 totwo meters above the ßoor. Twined andwoven sedge mats carpeted the ßoorsof these square rooms; their impres-sions remain in the clay-rich plaster ofthe ßoors.


PERUVIAN COASTAL DESERT is ill suit-ed for agriculture but almost perfect forarchaeologists. Arid conditions have pre-served perishable items for more than3,000 years. The sites investigated bythe authors lie in one of more than 50river valleys that cut through the desert.

0 5 10KILOMETERS

PAN-AMERICANHIGHWAY

PAMPA DE LASLLAMAS-MOXEKE

MOUNTAINS

ARCHAEOLOGICAL SITEMODERN CULTIVATION

PERU

SECHINRIV

ER

CASMA RIVER

SECHINALTO

COMPLEX

MODERN CITY

CASMA


The largest units lie along the centralnortheast-southwest axis of the mound;rooms become increasingly smaller asone moves toward the periphery. Spac-es between the square-room units seemto have served mainly as access corri-dors. They generally contained neithermatting nor niches.

To Þnd out how people approachedHuaca A and moved about on itssummit, we cleared both the

northeast and southwest atria and ev-ery entrance within the north quadrant.We uncovered evidence that access tothe mound was so carefully restrictedand monitored that the associated bu-reaucratic system boggles the mind.Yet the physical controls we found aremore suggestive of a powerful underly-ing authority than of insurmountablephysical barriers.

Intimidation (or inspiration) of a vis-itor began as he or she ascended eithercentral staircase to enter one of the atriaof Huaca A. The rear walls of both atriacontained magniÞcent friezes on eitherside of the central entrance, clearly de-signed to impress the approaching vis-itor. The frieze within the northeastatrium is better preserved: there, sculpt-ed in low relief in a medium of clay-richsilt, are the remains of two immensefelines, probably jaguars.

These colossal Þgures, once more thansix meters high and 10 meters long, faceeach other across the northeast entranceinto the mound. The curving walls of theatrium enhance the three-dimensionaleÝect of the giant cats. Traces of red

pigment suggest that they were oncebrightly colored. After 3,800 years, onlythe feet and serpent-shaped tails re-main. ProÞle felines are frequently rep-resented in Andean art, however, andtherefore it is an easy task to recon-struct the body and head.

Along with the felines, the space im-mediately adjacent to the gateway con-tains circular reliefs whose central cir-cles hold four evenly spaced rectangles.This design occurs on the ends of victo-rious warriorsÕ clubs at the nearby siteof Cerro Sechin, leading us to suspectthat it is a symbol of authority.

In the southwest atrium, we foundan elaborately carved stone almost 50centimeters long. A double-bodied ser-pent design is incised into one face. Anadjacent surface of the stone containsa surprisingly realistic carving of theimpression of a right hand [see illus-

tration on page 72 ]. This stone wasaccompanied by two much longer onesthat probably supported it to form asmall arch with the serpent facing out-ward and the hand downwardÑpossi-bly a shrine or an altar. The west cor-ner of the atrium also housed a pair oflong stones. We believe that they werepart of a second arch, especially be-cause abundant food remainsÑpos-

sibly offeringsÑlay scattered nearby.Each entrance within the mound was

narrowed by opposing pilasters, sup-ported internally by small brancheslashed into a post-size bundle. Largerpieces of wood, a rare commodity in thedesert, in all likelihood served as lintelsabove some entrances and over mostof the wall niches within the rooms. Agate of large logs controlled accessthrough the main northeast entranceof the mound. The log gate, located onthe opposite side of the wall containingthe frieze, consisted of four uprightlogs about 3.5 meters long mounted oneither side of the entrance. Six addi-tional logs, each almost exactly 4.1 me-ters long, Þtted into the gap betweenthe vertical logs and the walls. Theirlarge and tapered ends alternated tocreate a level entrance barrier.

Other entrances on Huaca AÑ93 inallÑwere protected by one or more bar-closure mechanisms, devices never be-fore described at early Peruvian sites.While Þrst clearing these entrances, wenoticed small square stone niches intheir side walls. These niches always oc-


WALLS OF HUACA A were built mostly of granite stones cov-ered with layers of coarse mud plaster and a Þne clay-richÞnishing coat. Niches held luxury goods and staple commodi-

ties. Room entrances were ßanked by columns made ofwoody plants tied into bundles and coated with plaster. Con-ical adobe bricks (left ) made up the top courses of the walls.


curred in opposing pairs about half ameter above the entrance ßoor, and onemember of each pair was consistentlymany times deeper than the other.

The purpose of the niches becameclear as a result of simultaneous exca-vations in a nearby compound. An emp-ty niche 35 centimeters deep was op-posed by a niche that still had a hori-zontal wooden pole extending from itscenter. The pole was Þrmly embeddedin rubble, so we cleared along the walltop and discovered that it was housedin a well-constructed stone chambermore than two meters long. Freed ofdebris, the pole was pulled across theopening to rest in the opposing niche,barring access to the compound as ithad 3,800 years ago.

Why were all of the entrances barred?There is no evidence that anyone livedon Huaca A: neither cooking hearthsnor refuse. The mound was neither aheavily guarded palace nor royal livingquarters. Instead Huaca A was appar-ently used to store commodities andother valuable objects. We found only afew artifacts during the excavationÑtobe expected because valuables wouldhave been removed when the site wasabandoned. What remained were tex-tile fragments in the wall niches, eitherfrom stored cloth or cloth containers.In addition, soil samples from the nich-es yielded pollen from cotton as wellas from food plants, including beans,potatoes, sweet potatoes and peanuts,indicating that such commodities werestored there. (Contrary to commonlyheld beliefs that corn was an essentialstaple for New World civilizations, it was

not among the agricultural productswe found.) Thousands of rodent bonesconÞrm the evidence of food storage.

There are tantalizing clues about thepeople who used Huaca A and moni-tored traÛc within it. The Þnest textilesfrom the site were found on the mound,as were many of the turquoise beadsand a unique wooden Þgurine. Theseitems suggest that only the wealthierelite inhabitants of the city had readyaccess to Huaca A; if the lower classeshad frequented the mound, we wouldhave found evidence of their posses-sions there as well.

The square-room unit that domi-nates Huaca A is duplicated inthe rows of intermediate-size

public buildings that parallel the mainsite axis. Each of the compounds nearthe main mound consists of one suchunit, and single square-room units formthe central core of each of the smallermounds in the longer rows. These build-ings were apparently occupied by bu-reaucrats who monitored the ßow ofcommodities into and out of the mound.

Behind several of the small moundsstand the remains of upper-class hous-es. Like the mounds and enclosures,these structures are aligned with thesite axis; they are well made, with highwalls of stone with mud plaster. Thereis also ample evidence of domestic ac-tivities. One room has a large, square,stone-lined hearth with a central circu-lar depression for the Þre. Small roomscontaining large jars, rooms with wallniches and subßoor chambers wereused for storage within the houses of

the elite. The residences lack the bilat-eral symmetry of the public architec-ture, however, and household debrissurrounds them.

Both Huaca A and the small moundsand enclosures were kept free of house-hold refuse, and so this garbage yieldsthe best information we have about thelives of the bureaucrats. Fragments ofsolid pottery Þgurines are commonwithin and near the elite houses. Wealso found many fragments of stonemortars and pestles. These stones beartraces of red pigment that were used topaint friezes, walls and possibly thebodies of citizens.

Most interesting, in light of the siteÕsdeveloped bureaucracy, are what ap-pear to be stamp seals and cylinderseals made of pottery. Both types arerare in the Andean area. Although laterAndean cultures used stamp seals toimpress designs on pottery, there is noevidence that they served this functionat Pampa de las Llamas-Moxeke. Insteadtwo examples contain red pigment res-idues, indicating that they were used toapply colored decoration on surfacessuch as textiles or human skin. Wethink it is possible the seals served assymbols of power and authority, muchas they did in the centers of early civi-lization in the Mediterranean and theNear East, where they are more fre-quently found.

Low-status housing at Pampa de lasLlamas-Moxeke shares some character-istics with the elite domestic architec-ture; after all, both classes performedmany of the same basic householdchores. Refuse from food preparationand other domestic activities is abun-dant, and square, stone-lined hearthsare common but much smaller. Manylow-status residents cooked insteadover cloverleaf-shaped hearths in whichthree upright stones supported round-bottomed pots. Moreover, the construc-tion of low-status houses diÝered sub-stantially from that of high-status resi-dences: the rooms are smaller, irregularand clustered, and the stone parts ofthe walls (still extant) are but a fewstones high. The upper walls were madeof perishable cane or wood coveredwith mud plaster.

Although our excavations have re-vealed a society of unsuspected complexity for its time, continu-

ing Þeld research has made us aware ofa much larger, integrated political enti-ty. In the northern branch of the CasmaValley lies the Sechin Alto ComplexÑfour sites that collectively cover 10.5square kilometers, more than Þve timesthe area of Pampa de las Llamas-Mox-eke. Before 1992, detailed scientiÞc ex-


ANCIENT FRIEZES are still visible at the bottom of the walls that ßanked the mainentrance to the Huaca A mound. Their feline images once stood more than Þve me-ters high and were painted with red and white pigment.


cavations had been carried out only atthe smallest site, Cerro Sechin. Sincethen, we have uncovered a wealth ofevidence that links all the sites in thecomplex and ties them to Pampa de lasLlamas-Moxeke.

Our excavations in 1992 and 1993concentrated on the best-preserved Se-chin Alto site, Taukachi-Konkan. Imme-diately we became aware that the peo-ple of Taukachi-Konkan used the samekinds of pottery and architectural ele-ments we had seen Þve kilometers tothe south. The large mounds containedthe same square-room units, wall nich-es and restricted entrances as did theintermediate-size ones. The siteÕs cen-ter is open, formed by several large rect-angular plazas lined by intermediate-size mounds. Two large mounds opentoward sunken circular plazas similarto the one adjacent to Huaca A. Radio-carbon dates of six samples range fromabout 2000 to 1300 B.C.

The sites are very similar in someways, but their diÝerences have provedeven more signiÞcant, leading us to re-alize that the early society in the CasmaValley was more sophisticated than wehad imagined. Taukachi-KonkanÕs larg-est mound, the Mound of the Columns,is unlike any of the others that havebeen excavated. It is located on thewestern edge of the site, and its summitis divided into two large central courtswith numerous square-room units ar-ranged symmetrically around them. Onthe front half of the summit we foundthe remains of about 100 round col-umns made of bundled cane, rope andstones covered with mud mortar andplaster. These columns supported aroofÑa unique feature that presum-ably shaded members of the elite andscreened them from view.

We believe the Mound of the Columnsis probably the residence of an extreme-ly important person, possibly the rulerof the Sechin Alto complex or the en-tire Casma Valley system. In additionto the roofed court, it contains a fewroofed rooms lined with high nichesthat were most likely storerooms tohouse valuable goods both received andgiven. Other rooms have low nichesmore appropriate for seating and werepossibly meeting or audience areas.Rooms on the rear half of the summitare irregular in layout and less accessi-ble, leading us to believe they mighthave been living quarters. Two roundsweathouse structures (enclosed, heat-ed rooms intended for small religiousceremonies) are also located there.

The placement of this palace at Tau-kachi-Konkan is logical because the Se-chin Alto complex probably served asthe capital of a valleywide political en-tity. Although both sites are large andimportant centers, we now know thateach formed a vital part of an evenlarger, more complex society that waswell developed by 1800 B.C.

After enduring for the better part of a millennium, these cities were abandoned. The many buildings

left unÞnished at both sites suggest thatlocal development stopped suddenly atits height, rather than declining gradu-ally. Researchers Þrst speculated thatinvaders from highland cultures madewar on the coastal sites and disruptedthe civilization. Evidence of the demiseof the Initial Period polity coincides withthe abrupt appearance of maize, domes-ticated animals, new kinds of artifactsand diÝerent building styles. We haverecently uncovered evidence, however,that internecine conßict weakened the

Casma Valley culture and helped theinvadersÕ cause.

Stone carvings in Cerro Sechin depict,at life size, victorious warriors and theirmutilated victimsÑa clear representa-tion of battle. The warriorsÕ facial deco-rations and the rectangle-in-circle sym-bols on their war clubs (like those onthe Huaca A frieze) argue that the vic-tors were local. Yet the clothing of thevictims matches that of humanlike Þg-ures that decorate the front of the Mox-eke temple mound.

We can only guess which factionswarred against one another. Cerro Se-chin may commemorate the victory ofone segment of the elite over another;perhaps bureaucrats associated withthe more secular activities of Huaca Adeposed the religious hierarchy ofMoxeke in a decisive conßict betweenchurch and state. Then again, the carv-ings may represent a rebellion by thegeneral population against the prolifer-ating bureaucracy. As these plannedcities grew and their elite populationexpanded, demands on the laboringpopulace could have become excessive.

Although archaeologists may neverknow precisely how these cities roseand fell, their extreme age makes themcrucial to the understanding of Peru-vian prehistory. Pampa de las Llamas-Moxeke and Taukachi-Konkan do notrepresent an unknown or undiscoveredcivilization. Rather their new place-ment so far back in time demonstratesthat the Initial Period was not merely aprelude to ÒtrueÓ civilization. Insteadthis time span witnessed the develop-ment of fundamental social, political,economic and religious principles thatshaped Peruvian culture for the next3,000 years.


CARVED STONE (possibly an altarpiece) found in the atrium of Huaca A was in-cised on the front and bottom faces. In use, it was supported by two rough stonesand probably surrounded by oÝerings of food.

FURTHER READING

RECENT EXCAVATIONS AT PAMPA DE LASLLAMAS-MOXEKE, A COMPLEX INITIALPERIOD SITE IN PERU. Shelia Pozorskiand Thomas Pozorski in Journal ofField Archaeology, Vol. 13, No. 4, pages381Ð401; Winter 1986.

EARLY SETTLEMENT AND SUBSISTENCE INTHE CASMA VALLEY, PERU. Shelia Po-zorski and Thomas Pozorski. Universi-ty of Iowa Press, 1987.

THE ORIGINS AND DEVELOPMENT OF THEANDEAN STATE. Edited by JonathanHaas, Shelia Pozorski and Thomas Po-zorski. Cambridge University Press,1987.

CHAVIN AND THE ORIGINS OF ANDEANCIVILIZATION. Richard L. Burger. Thamesand Hudson, 1992.

EARLY CIVILIZATION IN THE CASMA VAL-LEY, PERU. Shelia Pozorski and ThomasPozorski in Antiquity, Vol. 66, No. 253,pages 845Ð870; December 1992.


For many centuries, naturalistshave observed that honeybees telltheir nestmates about discoveries

they make beyond the hive. Neverthe-less, the system of communication thatthe insects use remained a mystery un-til the 1940s, when Karl von Frisch ofthe University of Munich in GermanyÞrst discovered the signiÞcance of beesÕdances. In the hive the steps and wag-gles of a successful forager correlateclosely with the exact distance and di-rection from the nest to the resourceshe has discovered. For the next twodecades, most scientists believed beesrelied primarily on these silent move-ments to communicate.

In the 1960s this view was challengedin two ways. The Þrst challengers wereAdrian M. Wenner, then a graduate stu-dent at the University of Michigan, nowat the University of California at SantaBarbara, and Harald E. Esch of the Uni-versity of Munich, now at Notre DameUniversity. Working independently, thetwo researchers discovered that thedances were not silent after all. As thebees dance, they emit faint low-frequen-cy sounds, and Wenner and Esch bothsuggested that the sounds might play a critical role in the beesÕ communica-tion. The use of sounds, they reasoned,might account for the beesÕ ability tocommunicate eÝectively in the com-plete darkness that prevails inside theirnests. At the time, however, many sci-entists believed bees were deaf, and sothe issue remained open.

Wenner later raised the second chal-lenge to von FrischÕs description of thedance language, rethinking his Þrst hy-pothesis at the same time. Bees, he ar-gued, use none of the information inthe dances or the sounds. Instead heproposed that the insects rely on odorsto Þnd the new resource advertised bythe dancer.

Now both of these debates have been

resolved. Bees, it turns out, can hear,and their ears are well suited for detect-ing the sounds associated with the danc-es. Observation of how the insects re-spond to a robot that dances and singslike a live forager shows that both sound

and dance are needed to communicateinformation about the location of food.Silent dances, the experiment demon-strates, communicate nothing, andsound without dance also fails. Odorstoo are involved but appear to lack the


The Sensory Basis of theHoneybeeÕs Dance Language

Novel experiments, such as training bees to respond to sounds and recruiting them using a robot, have ended

several debates surrounding the dance language

by Wolfgang H. Kirchner and William F. Towne

ROBOTIC HONEYBEE, directed by a se-ries of step motors, can dance and emitsounds similar to those a live foragermakes (above). Experiments using thisrobot have enabled researchers to ex-plore how sounds are involved in thedance language. A syringe ending nearthe artiÞcial dancerÕs head delivers sug-ar water to its followers, much in thesame way that a live forager oÝers sam-ples of food to her nestmates (right ).


importance that Wenner ascribes tothem. Beyond the resolution of theseissues, we have also recently learnedmuch more about the nature of thedance sounds, the beesÕ sense of hear-ing and the aspects of the dance that aremost essential in the communication.

A certain pleasure comes from the so-lution of such a long-standing mystery.Aristotle himself documented the hon-eybeeÕs ability to recruit her nestmatesto a good food source but did not spec-ulate on how the communication tookplace. He and other naturalists did, how-ever, observe that a bee that Þnds a newfood source returns to the nest anddances for her sisters, rather than feast-ing alone. Pliny reportedly constructeda hive that had a window made fromtransparent horn, through which hecould monitor dancing bees. By study-ing bees kept in a glass-walled hive, vonFrisch and his followers in the 20th cen-tury were able to recognize a pattern inthe dance: The forager walks across thevertical sheets of comb hanging in the

hive and traces out the shape of a Þg-ure eight; she pauses in each loop toshake her body from side to side. A fewpotential recruits chase after the dancerattentively for some time and then ßyout on their own toward the target.Provided they like what they Þnd, theserecruits return to the nest and dance aswell, sending even more bees from thehive to investigate the site. Eventually,the best food sources inspire the mostdances and so attract the most bees.

In the 1920s von Frisch Þrst proposedthat the foragerÕs dance somehow gavethe other bees information about thefood source. But it was not until 1943

that von Frisch discovered that the di-rection in which the dancer faced dur-ing her waggling run pointed towardthe food site in relation to the sun: Ifshe waggled while facing straight up-ward, toward the 12 on a clockface,then the food could be found in the di-rection of the sun; if she waggled 60degrees to the left of 12, facing the 10on a clockface, then the food lay 60 de-grees to the left of the sun. In addition,he noticed that how fast the dancercompleted her circuits corresponded tothe distance between the hive and thefeeding site: the closer the food, themore frenzied her pace. Von Frisch and


WOLFGANG H. KIRCHNER and WILLIAM F. TOWNE began their collaboration in 1987and jointly discovered the auditory sense of the honeybee. Kirchner has studied themechanisms and evolution of dance communication in honeybees for the past 10 years.He received a Ph.D. in 1987 from the University of W�rzburg in Germany, where he cur-rently holds a faculty position. Towne has studied the dance communication of honey-bees since 1980. He received a Ph.D. in biology from Princeton University in 1985 and isan associate professor of biology at Kutztown University in Pennsylvania.


his colleagues made detailed accountsof the dance language. With a stopwatchclose at hand, they could observe thedance, decipher its meaning and thenlocate the food supply of which itspoke. The accurate translations were astunning accomplishment.

Nevertheless, several seriousquestions remained unanswered..First, investigators could not

prove that the dance truly functionedas a language. Perhaps the correlationbetween the excited foragerÕs move-ments and her Þnds in the Þeld weremerely coincidental and of no real sig-niÞcance to the bees themselves. More-over, researchers had yet to Þgure outhow the bees perceived and interpretedthe dances. The scientists could clearlysee the dancer through the glass walland clock her movements, but the beesthemselves, normally enclosed in darknests, certainly did not have this bene-Þt. What were the dances like from thebeesÕ point of view?

Years after von Frisch interpreted thesymbolism of the dances, Wenner andEsch independently found that dancing

bees make sounds during their wag-gling run. Both men suggested that thesounds might help the dancer attractan audience in the dark nest. Many re-searchers doubted this premise becausethey thought bees could not hear air-borne sounds. Still, the notion was notignored altogether. Many insects, in-cluding bees, are quite sensitive to vi-brations. Hence, some investigatorsspeculated that the sounds the foragersproduced could vibrate the combs un-der their feet as they danced. The combvibrations might then advertise thedance to those bees who could not oth-erwise see the forager.

One of us (Kirchner), together withAxel Michelsen of Odense University inDenmark, answered part of this ques-tion several years ago. In their experi-ments, Michelsen and Kirchner aimed alaser beam at the comb near a dancingbee to determine whether or not thedance sounds generated vibrations inthe comb. Surface vibrations, if any oc-curred, would cause minute changes inthe light reßected from the comb. In thisway, it was possible to measure the vi-brations without touching the comb and

possibly triggering additional tremors.These measurements revealed that danc-ing bees do not rattle the comb butthat their audience does.. The dance at-tenders sometimes emit a short squeakby pressing their thoraxes against thecomb. This action vibrates the combenough so that the dancing bee stopsher movements. She then doles outsmall samples of the food she has col-lected so that her audience knows notonly the direction and distance to thefeeding site but how the food smellsand tastes as well.

Although this result was intriguing, we still did not know whether hon-eybees could hear the sounds a

forager made during her waggling run.Because the sounds seemed likely to beinvolved in the dance communication,we have analyzed them thoroughly. Ini-tially, we discovered that the foragermade the dance sounds by beating hertiny wings and that the follower beesstood quite close to her when she didso. Therefore, our measurements of thesounds needed to be made in the sameproximity.

In further collaboration with Michel-sen, we gauged both the pressure chang-es and the air-particle movements nearthe dancer. Some insectsÕ ears do notrespond to sound pressures as humanears do but instead react to back-and-forth oscillations of nearby air mole-cules. In fact, it turned out that the pres-sure changes a dancing bee produceswhen she beats her wings are relativelysmallÑwhich no doubt explains whythe pressure-sensitive ears of scientistsdid not detect them for so long. But wefound signiÞcant air movements withina few millimeters of the dancerÕs vibrat-ing wings. We concluded that whereasthe dance followers produce sounds byvibrating the comb, the forager trans-mits her own sound signals exclusivelythrough the air. As a result, the noisesfrom the foragerÕs instructions and herfollowersÕ requests do not drown oneanother out.

Two research teams, using diÝerentexperimental approaches, then testedthe hypothesis that the forager trans-mits dance information acoustically. Inthe Þrst line of inquiry, Kirchner andKathrin Sommer, a graduate student atthe University of W�rzburg in Germany,changed the dance sounds simply byshortening the dancerÕs wings slightly.Clipped wings have a smaller vibratingsurface and so produce sounds thathave a higher pitch and a smaller am-plitude. Sommer found that bees whosewings were experimentally shortenedcontinued to forage and dance normal-ly. The insects could not, however, re-


RETURNING FORAGER reports about her Þnds beyond the hive by dancing andbeating her wings, thereby producing low-frequency airborne sounds. Her nest-mates crowd around her and hear the sounds by using organs in the second jointof their antennae. They monitor the dance attentively for a short while and thenßy from the nest to search out the feeding site alone.


cruit nestmates. Similarly, bees from amutant strain, called Òdiminutive wings,Óthat have congenitally small wings can-not recruit at allÑeven though they ßyand dance normally. Sommer studied acolony of honeybees, half of whomwere normal bees and half diminutive-wing bees. She found that the normaldancing bees recruited both normaland short-winged bees equally well; incontrast, the mutant dancing bees re-cruited both strains poorly.

In the second series of experiments,Michelsen and Kirchner used a modelbee that could perform the honeybeedances. Others had employed roboticbees in their research before, but noneof the devices could make the correctsounds while it danced. Thus, a new at-tempt seemed worthwhile. Michelsenand his co-workers in Odense con-structed the computer-controlled mod-el bee, which danced in W�rzburg forÞve summers.

MichelsenÕs team fashioned the mod-el bee from brass and coated it with athin layer of beeswax. The brass bee isslightly larger than a worker honeybee.Workers cut a razor blade to the appro-priate size to make a set of wings. Anoperator can vibrate these metal wingson command by tweaking a stiÝ wirethat connects them to an electromag-

net. He or she can rotate the model inplace by turning a thin rod attached toits back. A step motor at the far end ofthe rod steers the modelÕs rotations au-tomatically during its Þgure-eight dance.This same motor makes the model wag-gle from side to side. An x-y plotter,wired to a sliding metal sleeve placedaround the rod, moves the model back-ward, forward, or left or right as need-ed. Finally, a thin plastic tube that endsnear the modelÕs head delivers foodsamples (sucrose solution) from a sy-ringe. Yet another step motor regulatesthis mechanism. A desktop computercontrols all the motors, which in turndirect the dance.

Each experimental session typicallylasted for three hours. First Michelsenand Kirchner scented the model and itssamples of sugar water with a faint ßo-ral fragrance, then placed baits in theÞeld that gave oÝ trace amounts of thesame odor. At each of the baits, an ob-server recorded the approach of search-ing bees. The results showed repeated-ly that the mechanical dancer could in-deed recruit live bees. Most of the beesinvariably went to the bait in the direc-tion indicated by the modelÕs dancesteps.

A number of additional experimentsusing the model bee followed. These

trials were designed to determine howimportant various aspects of the danc-es are to the dance followers. For ex-ample, in some experiments, the modeldelivered food samples but did notdance. In this case, far fewer recruitsventured to the targeted bait. Also, themodel failed completely to recruit livebees when its metal wings did not vi-brate; these silent dances did not work,demonstrating that the sounds are tru-ly an essential part of the honeybeeÕsdance language.

The model beeÕs ability to recruit itsnestmates showed that von Frisch wascorrect about the communicatory func-tion of the dances. Earlier, however, de-spite what many researchers felt wascompelling evidence supporting vonFrischÕs theory, some of our colleaguesdoubted that the bees used the distanceand direction data encoded in the danc-es. Wenner and Patrick H. Wells of Oc-cidental College and others have arguedpersistently that the coordinates givenin the dance represent correlations onlyand are not signals. They believe re-cruits depend solely on odors to Þndfeeding sites.

Von Frisch himself Þrst attempted totest the signiÞcance of the dance move-ments. He found that recruits could nolonger Þnd the correct food source


WAGGLING DANCE, performed on the vertical sheets of combhanging in the nest, follows the path of a Þgure eight. In eachloop the dancer waggles only when she is facing the direc-tion of the food source in relation to the sun. If the food canbe found by ßying toward the sun, the dancer waggles facing

straight up. If the food lies 135 degrees to the right of thesun, she waggles facing 135 degrees to the right; if the foodcan be found by ßying away from the sun, then she wagglesfacing down. Her pace reveals the distance between the foodand the hive: the faster the dance, the closer the food.


when he laid their hive horizontally onits side. The maneuver prevented thedancer from using gravity to orient thedirection of her waggling run. As a re-sult, the dance attenders could not in-terpret her actions correctly.

James L. Gould of Princeton Universi-ty later punctured the odor hypothesis.He showed that a forager can dispatch

her nestmates to a site she has nevervisited. Such a feat would be impossi-ble if the recruits relied on odors aloneto track down a feeding site. The eventcould occur, however, if the searchingbees gave priority to the informationthey received from the dance. In theseexperiments, Gould placed a brightlight in the hive, which the dance fol-

lowers mistook for the sun. In doing so,they interpreted the dances erroneous-ly. These misdirected bees most oftensearched in the Þeld using the mis-aligned dance information and seemedto ignore other cues such as odor. Gouldconcluded that they evidently pre-ferred the message given in the danceto the other signals. Finally, the experi-


Aforager emits dance sounds that are dis-

tinguishable from those adance follower makes inthe hive. The forager beatsher wings and generatessounds that travel exclu-sively through the air.These sounds convey in-formation about the loca-tion of food outside thehive. In contrast, adance follower pro-duces sounds bypressing her thoraxagainst the comb;this action vibratesthe comb and causesthe forager to stopher dance and pro-ceed to dole out foodsamples.

DANCING ROBOT (left ) successfully recruited its nestmatesto food away from the hive. The experimenters placed eightbaits around the hive and programmed the robot to dance

concerning one site. Observers in the Þeld recorded the ap-proach of searching bees. Most of the robotÕs recruits went tothe bait indicated by its dance (two trials shown at right ).

The Sounds of the Dance Language

TRIAL ONE

TRIAL TWO

TARGETEDBAIT

DANCE SIGNAL (AIRBORNE SOUND) STOP SIGNAL (SUBSTRATE VIBRATION)


ments using the model bee conÞrmedvon FrischÕs hypothesis; the dances doindeed represent a sophisticated formof communication.

The model bee has helped us answerother questions raised by von FrischÕsobservations, such as determining whichcomponents of the dance language rep-resent what kinds of instructions. Forexample, when the model danced sothat her waggling run appeared on theoutside of the Þgure-eight path, the re-cruits followed the direction indicatedin the waggling run, rather than thatgiven by the orientation of the Þgureeight as a whole. Thus, the wagglingrun alone, during which the sounds areproduced, tells the recruits the direc-tion in which they can Þnd the food.

Although the experiments using both the model bee and the di-minutive-wing bees conÞrmed

that the dance sounds are an impor-tant part of the dance language, a cru-cial element of the picture was absent:the identity of the structures throughwhich bees hear airborne sounds. Sev-eral previous attempts made to resolvethis issue had shown that bees seem-ingly could not hear at all. Yet becausewe now knew that the dance soundstraveled exclusively through the air, wefelt inspired to reinvestigate the ques-tion. In these renewed trials, we testedthe bees using sounds very similar tothose that dancing bees make.

In our Þrst series of experiments, wetrained the bees to associate a sound,lasting for Þve seconds, with a verymild electric shock, arriving four sec-onds after the sound had started. Wegenerated the sound at the open end ofa narrow glass tube. The shock alone, ifdelivered while the bee was feeding,would drive the bee from the feeder fora few seconds; shortly thereafter shewould return and continue feeding. Wethen posed the following question: If abee repeatedly experiences a tone fol-lowed by an adverse stimulus, will sheeventually learn to withdraw from thefeeder within the Þrst four seconds ofthe sound, before the shock? If so, thenshe can hearÑand of course learn. Wefound that bees can indeed be trainedto respond to airborne sounds, althoughthey learned to do so very slowly.

More recently, we have employed adiÝerent training technique. In theseexperiments, a bee entered a very sim-ple Y-shaped maze. We played a soundat one end of this two-sided feeder.The side from which the sound camechanged unpredictably from one trialto next. If the bee turned toward thesound, she received a reward of sugarwater; if she went away from the sound,

she received nothing. We observed thatthe bees learned quickly to turn towardthe sound. Claudia Dreller, a graduatestudent at W�rzburg, used the proce-dure to explore the frequency and am-plitude range in which the bees couldhear. DrellerÕs work showed that honey-bees sense only low frequencies, thosebelow 500 hertz. They hear these toneswith suÛcient sensitivity to pick up thesounds of a dancing nestmate, whichrange from 250 to 300 hertz. They alsoshow some ability to discriminate be-tween frequencies in this range; theycan discriminate between low- (20hertz), medium- (100 hertz) and high-pitched (320 hertz) sounds. We do notyet know for what purpose the beesmight use this latter ability.

The same training technique enabledus to Þnd out through what sensorystructures the bees detect the sounds.We altered some of our trained bees byremoving one antenna, Þxing a certainantennal joint or removing sensoryhairs from their head. We found thatbees use a structure called the John-stonÕs organ, a chordotonal organ madeup of nerve cells in the second joint ofa beeÕs antennae, to pick up airbornesounds. Some ßies and mosquitoes relyon the same structure to perceivesounds.

If we now put the pieces together, wecan see how the dance language works.The dancer emits sound signals thathelp the dance followers determinewhere the dancer is and how she ismoving, which in turn oÝers them criti-cal information regarding the directionand distance to the feeding site. Thedance attenders receive these signalsthrough the JohnstonÕs organs locatedin their antennae, which are always

held near the dancer. Because these or-gans are bilateralÑone on the left andone on the rightÑthe dance followerscan use them to judge their positionwith respect to the dancer and thereforeunderstand the direction to the food.At the same time, the followers emitsounds that vibrate the comb. The for-ager stops her dance when she receivesthese signals and delivers samples ofthe food she has collected. These appe-tizers give the dance followers addi-tional hints about the taste, smell andquality of the food source. The bees at-tend the dancing for a while and thenßy out to Þnd the food source on theirown. If they are lucky, they will Þnd thefood. If they fail, they will return to thenest and try again.

This dance language is clearly avery complicated, highly devel-oped system of recruitment. To

understand how such a system evolved,we and other scientists have examinedthe recruiting techniques of related spe-cies. The genus Apis, to which all hon-eybees belong, has no close living rela-tives. Bumblebees and stingless beesare their nearest kin, and, unfortunate-ly, it seems that bumblebees do not re-cruit at all. Many species of stinglessbees, on the other hand, do recruit. But,as far as we know, none has developeda symbolic language similar to thedance language of Apis. All four spe-cies of honeybees studied so far (threeof which live in Asia) speak some vari-ant of the dance language. As MartinLindauer discovered in the 1950s, whenhe was at the University of Munich, allspecies use similar distance and direc-tion codes, even though there are somediÝerences.


FREQUENCY RANGE of the sounds a bee can detect extends well below the rangeheard by human ears. The graph shows how fast the air particles near a dancerÕswings must travel to generate audible signals. Within this range, the bees show anability to diÝerentiate between sounds having varying frequencies.

2,000

1,500

1,000

500

00 100 200 300 400 500

FREQUENCY (HERTZ)

PA

RT

ICLE

VE

LOC

ITY

(M

ILLI

ME

TE

RS

PE

R S

EC

ON

D)


Although we have looked for dancesounds in four species of honeybees,we have found that only three producethem. These three species hold some-thing else in common: they all must oc-casionally dance in the dark. Two ofthem, the familiar western bee, A. melli-

fera, and the Asian bee, A. cerana, nestin lightless, enclosed areas such as hol-low trees or other similar cavities. Thethird sound producer, the giant bee, A.dorsata, nests in the open on singlesheets of comb, hanging under rockoutcrops or thick branches of trees.Fred C. Dyer of Michigan State Univer-sity Þrst showed that A. dorsata some-times dances at night, and Kirchnerdiscovered only recently that this beeproduces sounds. A. dorsataÕs signalswere very diÛcult to detect because

these sounds are particular-ly low in pitch.

The single species thatdances silently, the dwarf bee, A.ßorea, dances in the open like A. dorsa-

ta, but only during the day. Dancers ofthis species make gestures that may, indaylight, serve as visual signals to at-tract dance followers in the same waythat sounds assist those bees whodance in the dark. Because some indi-cations suggest that A. ßoreaÕs habitsare the most primitive, we assume thatthe complicated acoustical communica-tion system of the other three speciesmost likely evolved from a visual dis-play when these bees developed habi-tations that cut them oÝ from light.

Now that we can Þnally listen to thebeesÕ language and even speak it a lit-tle, we face a host of new questions. Forexample, we have yet to learn for whatpurpose the bees possess the ability to

distinguish between diÝerent pitchedsounds. In addition, perhaps they usesimilar airborne sounds in ways thatwe do not even suspect at this time. Inthe hopes of revealing their communi-cation system in full, we will continueto eavesdrop on their conversations.


NESTING HABITS may have inßuenced the evolution of the dance lan-guage, as it is spoken by diÝerent species of honeybee. Apis ßorea(bottom left ), considered the most primitive in behavior, does not pro-duce dance sounds. These bees nest in the open and dance only duringthe day. A. dorsata (above) also nests in the open but sometimes dancesat night. A. dorsata makes dance sounds, as do A. cerana (top left ) andthe familiar western bee, which usually dance in dark nests.

FURTHER READINGTHE DANCE LANGUAGE AND ORIENTA-TION OF BEES. K. von Frisch. HarvardUniversity Press, 1967.

COMMUNICATION AMONG SOCIAL BEES.Martin Lindauer. Harvard UniversityPress, 1971.

THE HONEY BEE. James L. Gould andCarol G. Gould. Scientific American Li-brary, 1988.

ACOUSTICAL COMMUNICATION IN HON-EYBEES. W. H. Kirchner in Apidologie,Vol. 24, No. 3, pages 297Ð307; July1993.


In 1785 William Withering, a Britishphysician, reported that ingestionof dried leaves from the foxglove

plant eased dropsy, an accumulation ofßuid now known to be caused by theheartÕs failure to pump adequately. With-ering credited an unexpected source forhis information. ÒI was told,Ó he wrote,that this use of foxglove (a member ofthe genus Digitalis) Òhad long been kepta secret by an old woman in Shrop-shire, who had sometimes made curesafter the more regular practitionershad failed.Ó

Digitalis has been helping cardiac pa-tients ever since. Today two of its com-ponentsÑthe glycosides digoxin anddigitoxinÑare prescribed to hundredsof thousands of people throughout theworld every year. Indeed, these glyco-sides currently serve as the treatmentsof choice for rapid atrial Þbrillation, adangerous cardiac irregularity. Giventhe importance of Digitalis, we have nodoubt that many readers of ScientiÞc

American are alive in 1994 becauseWithering investigated the secret reme-dy of Òan old woman in Shropshire.Ó

A decade ago this story would proba-bly have been regarded as nothing morethan a historical anecdote, of little rele-vance to contemporary drug discovery.By the mid-1980s most pharmaceuticalmanufacturers had abandoned explor-ing folk uses of plants in their searchfor new drugs. Now, however, the pen-dulum is beginning to swing back to-ward an appreciation that plants usedin traditional medicine can serve as asource of novel therapeutic agents.

Such appreciation has emerged inpart because of recent discoveries madeby a small but growing group of ethno-botanistsÑresearchers who study therelationships between plants and peo-ple. Fieldwork exploring the medicinaluses of plants by indigenous peoples inremote parts of the world, coupled withthe introduction of sophisticated assaysable to determine whether plants exerta biological eÝect, has facilitated thediscovery of bioactive molecules madeby medicinal plants. Some of these mol-ecules show promise as possible thera-pies for a range of diseases, includingAIDS and cancer.

In the U.S. the drug developmentprocess is a long and arduous one, de-signed to ensure that therapies releasedto market are eÝective and safe. It canthus easily take many years for a sub-stance to become commercially avail-able as a drug. There seems little doubt,however, that within a decade severalnew agents derived from ethnobotani-cal research will be introduced. We can-not claim that most drugs of the futurewill be found in this way. Yet the strat-egy, as will be seen, has many merits.

Until the 1950s, almost all pharma-ceutical research relied heavily on vas-cular plants as sources of medicines.Flowering plants and ferns (as opposedto microscopic organisms and fungi)have given rise to about 120 commer-cially sold drugs and account for some

25 percent of all prescriptions issuedevery year in North America. Many ofthese agents are now synthesized in thelaboratory, but others are still isolatedfrom plants. Most were discovered bystudying indigenous uses of plants.

For instance, the drug reserpine,which is still occasionally pre-scribed in the U.S. for hyperten-

sion, was isolated from the root of theclimbing shrub RauvolÞa serpentina (In-dian snakeroot) after scientists begananalyzing Ayurvedic remediesÑthe tra-ditional treatments used by the peoplesof India. Other examples include aspi-rin, opiates, pilocarpine (prescribed forglaucoma and dry-mouth syndrome)and two cancer treatmentsÑvincristineand vinblastine. Vincristine and vinblas-tine, both of which are still extractedfrom Catharanthus roseus (rosy peri-winkle), have been prescribed for pedi-atric leukemia and HodgkinÕs disease,respectively, since the 1960s.

Plants have been a rich source ofmedicines because they produce a hostof bioactive molecules, most of whichprobably evolved as chemical defensesagainst predation or infection. Never-theless, several forces conspired by theclose of the 1970s to cause plants tolose much of their appeal as drug sourc-es for the pharmaceutical industry. Mi-croorganisms and fungi that inhabitsoil, which are easy to collect and cul-ture, had provided a dazzling array ofantibiotics. Advances in synthetic chem-istry and molecular biology promised tosupply new means for designing drugsin the laboratory. And few major dis-coveries of plant-derived drugs had fol-lowed the identiÞcation of vincristineand vinblastine. Given these conditions,many pharmaceutical Þrms simplystopped searching for therapeutic com-pounds in higher plants.

In spite of this discouraging state


The Ethnobotanical Approach to Drug Discovery

Medicinal plants discovered by traditional societies are proving to be an important source

of potentially therapeutic drugs

by Paul Alan Cox and Michael J. Balick

PAUL ALAN COX and MICHAEL J. BALICK met in the late 1970s, when theywere doctoral students at Harvard Uni-versity. Cox is dean of general educationand honors and professor of botany atBrigham Young University. He is alsopresident of the Society for EconomicBotany. Balick is Philecology Curator ofEconomic Botany and director of the In-stitute for Economic Botany of the NewYork Botanical Garden in Bronx, N.Y. Heis also past president of the Society forEconomic Botany. Cox and Balick, whoare collaborating on a book about ethno-botany for W. H. Freeman and Company,serve as advisers to a variety of govern-mental, academic and industrial researchgroups and foundations.


of aÝairs, in the late 1970s and early1980s several groups of scientists inde-pendently set oÝ to diÝerent cornersof the world for the express purpose ofÞnding innovative drugs through eth-nobotanical research. Some of the re-searchers, such as Gunnar Samuelssonand Lars Bohlin of the University ofUppsala, were former students of thedistinguished Swedish ethnobotanistFinn Sandberg. A number of Americanresearchers trace much of their inspi-ration to another leading ethnobota-nist, Richard Evans Schultes of HarvardUniversity.

The two of us are among that lattergroup. When we were graduate studentsat Harvard in the late 1970s, Schultesencouraged us to continue our studiesof traditional uses of plants in remoteregions of the earth. Balick, in his doc-toral work, extended SchultesÕs pioneer-ing ethnobotanical research in the Ama-zon. Meanwhile Cox, who had been amissionary in Samoa for the Mormonchurch during his undergraduate years,returned to the South PaciÞc to studythe ecology of the rain forest and theuses of plants by the Samoan people.After receiving our degrees, we contin-ued to investigate herbal medicine intwo culturally and geographically dis-tinct areas: Central America (Balick) andPolynesia (Cox). Both of us have stud-ied intensively with indigenous healers.

The ethnobotanical approach is ac-tually one of several methods thatcan be applied in choosing plants

for pharmacological studies. It is esti-mated that 265,000 ßowering speciesgrace the earth. Of these, less than halfof 1 percent have been studied exhaus-tively for their chemical compositionand medicinal value. In a world withlimited Þnancial resources, it is impos-sible to screen each of the remainingspecies for biological activity. Somekind of collection strategy is needed.

Investigators, for example, can gath-er vegetation randomly in an area sup-porting rich biological diversity. Unfor-tunately, random searches yield rela-tively few new drug possibilities. Onenotable exception is taxol. In 1992 the

Food and Drug Administration ap-proved taxol, derived from Taxus brev-

ifolia (the PaciÞc yew tree), as a treat-ment for ovarian cancer and in 1994 ap-proved it for treating metastatic breastcancer unresponsive to other therapies.Taxol was found in the course of a ran-dom-screening program conducted bythe National Cancer Institute (NCI),which has maintained a plant-screening

program, with varying degrees of ener-gy, since 1960. (Today the NCI employsseveral plant-gathering strategies. It alsoexamines specimens not only for theiranticancer eÝects but also for their abil-ity to impede the functioning of the hu-man immunodeÞciency virus, or HIV,which causes AIDS.)

Other plant-collecting methods, in-cluding the ethnobotanical approach,

ANDREW RAMCHARAN, a healer livingin Belize, Central America, is collectingthe leaves and flowers of the plant Cor-nutia pyramidata to include in a mixtureused for skin rashes. Ethnobotanists in-terested in drug discovery often rely onhealers to identify plants that are likelyto contain potent bioactive chemicals.There is some urgency to this work:many healers are elderly and lack ap-prentices. As they die, much of theirknowledge of local vegetation dies, too.


are more targeted. In phylogenetic sur-veys, researchers choose close relativesof plants known to produce useful com-pounds. In ecological surveys, they se-lect plants that live in particular habi-tats or that display characteristics indi-cating they produce molecules capableof exerting an eÝect on animals. Collec-tors might, for example, focus on spec-imens that seem to be immune frompredation by insects. The absence ofpredation suggests a plant may pro-duce toxic chemicals. Many chemicalsthat are toxic to insects also show bio-activity in humans, which means theymight be capable of achieving sometherapeutic eÝect.

Finally, the ethnobotanical ap-proach assumes that the indig-enous uses of plants can oÝer

strong clues to the biological activitiesof those plants. For instance, if our col-leagues at the NCI or at pharmaceuticalÞrms asked us to focus on collectingplants that might serve as therapies forHIV, we would pay special attention toplants a society uses against diseaseswe know are caused by viruses. Plantsexploited for their poisonous eÝectsare also of interest. Blowgun poisonssuch as curare have in the past beenfound to contain compounds able toserve as anesthetics.

The history of drug discovery im-

plies that the ethnobotanical approachis the most productive of the plant-sur-veying methods, and recent ÞndingsconÞrm that impression. Some of thedata were collected by Cox with his stu-dent Rebecca Sperry and with Bohlinand Mervi Tuominen of the Universityof Uppsala. This group found that 86percent of the plants used by Samoanhealers display signiÞcant biologicalactivity in a variety of assays. Balick andRosita Arvigo of the Ix Chel Tropical Re-search Foundation in Belize discoveredthat crude extracts of the plants thatone healer (Don Elijio Panti, in Belize)considered to be his most powerfulgave rise to four times as many positiveresults in a preliminary laboratory testfor activity against HIV than did speci-mens collected randomly.

We should note, however, that few ofthe compounds exhibiting activity inlaboratory tests will become new drugs.Some will turn out to be identical to, orless potent than, existing agents; oth-ers will prove too toxic for commercialuse. Nevertheless, demonstrating activ-ity in a bioassay is a necessary Þrst stepin the drug development process.

How do ethnobotanists choose thesocieties they study? Researchers diÝerin their criteria. Because not all culturesare equally likely to make use of plantshaving signiÞcant pharmaceutical ac-tivity, the two of us focus our eÝorts

on those possessing three characteris-tics. First, the societies should be locat-ed in a ßoristically diverse area, suchas a tropical rain forest. Such diversitydramatically increases the number ofplants available; it thus enhances thelikelihood that plants with pharmaco-logically active molecules will be pressedinto service.

Second, the societies should have re-mained in the region for many genera-tions. Groups who have resided in oneplace for a long time presumably havehad ample opportunity to explore andexperiment with local vegetation. Ac-cording to this requirement, aboriginalpeoples who have populated Australiafor many thousands of years would bea better choice for study than Europeansettlers.

Third, the cultures must have a tra-dition in which healers transmit theirplant knowledge from generation togeneration, usually through apprentic-es. Consistent application of a givenspecies for an ailment over millenniagenerates information rather analogousto that produced by large-scale clinicaltrials. Such repetitive, long-term use ofbotanical species can be expected tohave identiÞed both the most eÝectivemedicinal plants and those that are tootoxic for use.

As these criteria suggest, we are es-pecially interested in recording the spe-


Drugs Discovered from Ethnobotanical Leads

The well-established drugs listed below are among doz-ens that were developed after scientists began to analyzethe chemical constituents of plants used by traditionalpeoples for medicinal or other biological effects. For in-

stance, Western researchers isolated reserpine in 1952from the climbing shrub Rauvolfia serpentina (photo-graph ), which has been employed in India for many cen-turies to treat snakebite and mental illness.

DRUG MEDICAL USE PLANT SOURCEAspirin Reduces pain and inflammation Filipendula ulmariaCodeine Eases pain; suppresses coughing Papaver somniferumIpecac Induces vomiting Psychotria ipecacuanhaPilocarpine Reduces pressure in the eye Pilocarpus jaborandiPseudoephedrine Reduces nasal congestion Ephedra sinicaQuinine Combats malaria Cinchona pubescensReserpine Lowers blood pressure Rauvolfia serpentinaScopolamine Eases motion sickness Datura stramoniumTheophylline Opens bronchial passages Camellia sinensisVinblastine Combats Hodgkin’s disease Catharanthus roseus

Rauvolfia serpentinaINDIAN SNAKEROOT

Filipendula ulmariaMEADOWSWEET

Papaver somniferumOPIUM POPPY

Datura stramoniumJIMSON WEED

Cinchona pubescensFEVER TREE

Catharanthus roseusROSY PERIWINKLE


cialized information possessed by heal-ers; that is, knowledge not necessarilyshared by most other individuals in thecommunity. Many members of indige-nous societies can identify some com-mon plants that have medicinal uses,much as many Europeans realize thatAloe vera is helpful for burns. But onlya healer might know the best plant tochoose for treating serious or uncom-mon diseases, such as yellow fever.

Other ethnobotanistsÑnotably, BrentBerlin of the University of California atBerkeley and Walter H. Lewis and Mem-ory Elvin-Lewis of Washington Universi-tyÑprefer to study plants used by abroad array of villagers. This approach,called the consensus technique, is basedon the supposition that awareness ofhighly eÝective medicinal plants willspread rapidly throughout a culture. Weconcentrate on seeking esoteric infor-mation because many healers no longerhave apprentices. As healers die, theirarcane wisdom is likely to be extin-guished forever. Botanical insights thatare widely held should be accessiblelonger.

Interestingly, healers often practiceboth as generalists and specialists. InBelize, caregivers provide some level ofprimary health service to residents oftheir villages and surrounding regions,yet they often also have unique areas ofexpertise. Hortense Robinson of Lady-ville specializes in midwifery and otherhealth care concerns of women andchildren. Andrew Ramcharan of Ran-chito focuses on treating snakebite inan area that harbors many venomoussnakes. The knowledge such healerspossess can be quite remarkable. In theislands of Samoa, most healers are fe-male and specialize in herbal medicine;only bone setters are commonly male.The herbalists, or taulasea, typicallymake use of more than 100 species ofßowering plants and ferns.

We Þnd healers mainly by askingpeople to identify the individuals towhom they turn when they are sick.Having located healers, we take painsto explain our mission to them and totribal chiefs. This process is analogousto Òinformed consentÓ in clinical set-tings. The interactions go most smooth-ly if Western researchers are trained inboth botany and anthropology andhave a good understanding of the soci-eties they visit. If the healers are amen-able to teaching us, we spend manyweeks, months or even years with themand take careful notes on the proper-ties and uses of the plants they showus. During such Þeldwork, we attemptto live in accordance with local cus-toms, and we usually learn the locallanguage.

When, with the guidance of the heal-ers, we have decided on the plants wewant to collect, we gather a kilogram(about two pounds) of each species.(Prior to collection, we always obtainpermission from the healers, villagechiefs, landowners and governments.)Whenever possible, we collaborate withlocal scientists or students. We savefour or Þve samples of a plant to serveas Òvoucher,Ó or reference, specimens.These specimens are pressed ßat, driedand labeled carefully. The labels indicatespecies name and indigenous name andapplication, as well as any directionsneeded for locating the plant again.Eventually the specimens are depositedin herbaria (collections of preservedplants) in several parts of the world forconsultation by other botanists.

We then prepare the rest of thematerial for transport to ourlaboratories. Typically we dry

the plants or preserve them in aqueousalcohol. Later, in the laboratory, we orour colleagues usually extract diÝerentkinds of molecules from the plants byimmersing them in various solvents.Next, we separate out the solvent anddeliver the resulting solvent-free ex-tracts (or sometimes the plant samplesthemselves) to collaborators at otheruniversities, governmental agencies ordrugmakers for screening.

In the 1960s standard screening as-says for many potential drugs consist-ed of injecting test material into a ro-dent and waiting to see if the animalfell ill, got well or displayed some oth-er change in behavior or health. Screen-

ing was thus a time-consuming, costlyand imprecise endeavor. Thirty yearslater bioassays are faster (usually beingautomated) and are signiÞcantly morespeciÞc. At the NCI, tiny amounts ofmaterial can be screened rapidly againstan array of up to 60 distinct human tu-mor cell lines. Many other assays assessthe ability of an extract to inßuence theactivity of a single enzyme involved inthe biochemical interactions that un-derlie a disease. For example, scientistslooking for an AIDS treatment mightevaluate whether the extracts inhibitthe activity of the enzyme reverse tran-scriptase in cells. Reverse transcriptaseis coded for by the genetic material ofHIV and is needed for replication of theviral particles.

When an extract displays signiÞcantactivity in a bioassay, we return to theÞeld with our local collaborators andretrieve a bulk sample of the originalplant, typically 50 to 100 kilogramsÕworth. Chemists need these extra sup-plies in order to isolate the molecule re-sponsible for the observed activity andto determine (by spectroscopic tech-niques) its structure. Newly emergingtechniques that require very little plantmaterial may render such return collec-tion trips unnecessary.

Once the molecular structure is iden-tiÞed, researchers compare it with thatof known chemicals. If the isolatedmolecule is novel, or has already beenfound but has not been studied as apossible drug, it may be analyzed fur-ther as is. In other cases, a syntheticversion will be examined insteadÑthatis, if workers can determine how to con-


ETHNOBOTANIST MICHAEL J. BALICK, one of the authors, is drying plants thatwere collected by his colleagues in Belize. Drying is one of the techniques bywhich investigators preserve their specimens in the field.


struct the substance and can do so at areasonable cost. Still other times, themolecule will serve as raw material thatis altered to produce some desired ac-tivity. Even if an isolated molecule or asynthetic version clearly cannot serveas a drug for some reason, its discov-ery might suggest previously unconsid-ered avenues for attacking a disease.

After investigators settle on a testmolecule having signiÞcant activityÑalso known as a lead compoundÑeval-uation proceeds as it would for any oth-er potential drug. Researchers study thecompound for such characteristics asits strength of binding to a particulartarget molecule and its toxicity to cells.If it passes these tests, a company orgovernmental agency may decide to in-vest the money needed (an estimated$200 million) to bring it to market. Atthat point, it becomes a drug candidate.To make its way to pharmacy shelves,the drug candidate must prove its worthin additional laboratory analyses andin clinical trials.

So far the rigorous ethnobotanicalÞeld searches that have been con-ducted since the early 1980s have

generated many lead compounds. Someof these have been identiÞed as drugcandidates, and a few may eventuallyreach pharmacy shelves. We cannot besure of the exact number of leads ordrug candidates, because most pharma-ceutical makers will not discuss prod-ucts in the early stages of development.We do know, however, that many of thelead compounds derived from ethno-botanical research exhibit potent anti-viral, antifungal or anticancer eÝectsÑproperties that are sorely needed in thepharmaceutical arsenal.

Two drug candidates that havereached clinical trials are being devel-oped by Shaman Pharmaceuticals inSouth San Francisco, Calif. The activecomponent in both drugs is derivedfrom a plant that grows in Central andSouth America. One formulation showspromise against respiratory viruses.The second version may be adminis-tered topically for treating infectionscaused by the herpes simplex virus.

Another exciting agent in the drugpipeline is a powerful antiviral com-pound called prostratin. The story ofprostratinÕs identiÞcation serves as agood paradigm of the drug discoveryprocess. When Cox was in Samoa in1984, several healers told him that theytreat yellow fever by giving patientsbrews made by steeping in water pul-verized wood of the rain-forest tree Ho-malanthus nutans. Cox collected sam-ples and brought them to the U.S.,where he prepared freeze-dried solvent

extracts. Because of the prevailing atti-tudes at the time, no drug companieswere inclined to analyze his specimens.

Fortunately, Michael R. Boyd and Gor-don M. Cragg of the NCI agreed to testthem. The extracts exhibited strong ac-tivity against HIV in the test tube.Chemists John H. Cardellina II and JohnA. Beutler, also at the NCI, then discov-ered that the active component was achemical called prostratin. The chemi-cal was known but was not being stud-ied as a drug, because it was a phorbol,a class of chemicals that promotes thedevelopment of tumors.

Worry over tumor promotion initiallydampened interest in prostratin at theNCI. But Cox argued that if this partic-ular phorbol stimulated the growth ofcancer, its carcinogenic propertieswould have become evident in Samoa;its continued use by healers suggestedit was safe. When Peter M. Blumberg, acell biologist at the NCI, examined thedrugÕs carcinogenic eÝect in mice, hefound, to his co-workersÕ surprise, thatit did not promote tumor formation,even though it activated an enzyme(protein kinase C) that participates inmalignancy. The NCI is now planning toaccept bids from drug companies forthe right to investigate prostratin fur-ther as a possible drug.

An additional compound identiÞed through CoxÕs ethnobotanical re-

search highlights the strikingprecision of the knowledge possessedby many healers. Samoan taulasea ap-ply the bark of a local tree, Erythrina

variegata, to the skin to treat inßam-mation. In describing this tree, they in-sisted that only one of the two types ofE. variegata in the region was helpful.Sure enough, when a team lead by Vin-od R. Hegde and Mahesh G. Patel ofSchering-Plough Corporation in Kenil-worth, N.J., evaluated the ability of barksamples to inhibit an enzyme involvedin inßammation (phospholipase A2),they found the healers were absolutelycorrect. Bark from only one variety of E.

variegata displayed anti-inßammatoryactivity. From that bark, the team iso-lated the active componentÑa chemi-cal known as a ßavanone. The chemicalis now under development as a topicalanti-inßammatory at Schering and atPhyton Catalytic in Ithaca, N.Y.

In Thailand a team led by Hans T.Beck of the New York Botanical Gardenand Weerachai Nanakorn, then of theRoyal Forest Department in Bangkok,learned from healers that the roots ofCurcuma comosa, a relative of the gin-ger plant, eased stomach pains and oth-er gastrointestinal disorders. Tannis Jur-gens and colleagues at Merck & Co., in

Rahway, N.J., then isolated from theroots a novel compound that kills para-sitic worms in the stomach. The com-pound is now under investigation as apotential treatment for parasitic dis-eases. In Peru, Lewis and Elvin-Lewis ob-tained samples of a tree sap that JivaroIndians and others apply for hasteningthe healing of wounds. The active in-gredient in the sap, called taspine, isbeing tested for that same purpose inanimals. If those tests are successful,human trials are likely to follow.

The probability that drug compa-nies, universities and Western sci-entists will proÞt Þnancially from

information provided by healers in tra-ditional societies raises a serious ques-tion: What is being done to protect theinterests of healers and their communi-ties? We are well aware that the healerswith whom we work provide signiÞcantintellectual guidance. Indeed, we referto them as Òcolleagues,Ó ÒguidesÓ andÒteachers,Ó instead of as Òinformants.ÓGiven the signiÞcant participation of in-digenous peoples in our research pro-grams, we believe they are entitled tothe same intellectual property rightsenjoyed by other investigators. Hence,we do all we can to see that those rightsare protected.

For instance, in the case of prostratin,the NCI and Brigham Young University,where Cox is a professor, have guaran-teed that a signiÞcant part of any royal-ties earned from the drug will be re-turned to the Samoan people. Virtuallyall ethnobotanists active in drug re-search are involved in making similararrangements for the cultures theystudy. In fact, now that ethnobotanicalinquiry is expanding, formal guidelinesare being devised.

To many cultures, protecting the for-ests around them is more importantthan receiving money. Hence, many re-searchers are devoting eÝort to protect-ing the rain forests in the regions wherethey work. For example, Thomas Eisnerof Cornell University helped to convinceMerck & Co. to invest $1 million in aproject designed to inventory and ulti-mately preserve a part of the Costa Ri-can rain forest. In return, Merck canstudy chemicals extracted from livingorganisms in Costa RicaÕs national parksand elsewhere.

Four village-owned and village-man-aged reserves encompassing 64,000acres have now been formed in Samoawith funds raised through the Seacolo-gy Foundation of Provo, Utah (an orga-nization Cox helped to create). Theseprotected territories include the Falea-lupo Rain Forest Reserve, the site wherethe plant that produces prostratin was


Þrst collected. In the U.S. territory ofAmerican Samoa, Cox participated inestablishing the 50th national park ofthis nation. By mandate of Congress,the park will allow and encourage Sa-moan healers to continue to use medic-inal plants in a sustainable wayÑthatis, in a way that guarantees their ongo-ing availability.

Balick and Arvigo and their colleaguesGregory Shropshire of the Ix Chel Trop-ical Research Foundation and LeopoldoRomero of the Belize Association ofTraditional Healers recently helped toestablish the worldÕs Þrst ethnobiomed-ical forest reserve. It is an entity de-signed speciÞcally to ensure that me-dicinal plants will be available for localuse. The protected area, called the Ter-ra Nova Rain Forest Reserve, compris-ing almost 6,000 acres of forest in theYalbak region, is to be managed by theAssociation of Traditional Healers. Thereserve was created not only to provideneeded medicinal plants but also toteach young people about their uses. Itis hoped that such teaching will en-courage youth to preserve the knowl-edge of their elders. Researchers willalso be collaborating with the healersto devise ways to make sure that medi-cinal plants are harvested sustainably.

In spite of its apparent successes,ethnobotany is unlikely to ever becomea major force behind commercial drugdiscovery programs. Its application islimited by the paucity of properlytrained ethnobotanists who have the

time to conduct rigorous, long-termÞeldwork in remote areas of the world.Also, many funders of drug researchstill perceive the ethnobotanical ap-proach as archaic, unscientiÞc and un-worthy of attention.

Yet the demonstrated ability of eth-nobotany to generate exciting leads fordrugs suggests to us that, for the nearfuture at least, the approach will occu-py an expanding role in drug develop-ment. Those who understand the value

of such work are truly in a race againsttime. In Samoa, two healers who Þrstprovided Cox with information leadingto the discovery of prostratinÑEpene-sa Mauigoa and Mariana LiloÑdied in1993. Generations of accumulated med-ical wisdom died with them. Ethnobot-anists can capture much of the remain-ing knowledge, but only if the researcheÝort is expanded soon. Sadly, plantknowledge seems to be disappearingeven faster than the forests themselves.


FURTHER READING

SAMOAN HEALER MARIANA LILO, whodied in 1993, is shown cutting the stemof Homalanthus nutans. She went on tomacerate the wood and steep it in hotwater, producing a brew to be drunk bypatients with yellow fever (a viral dis-ease). The active ingredient, prostratin(above ), was isolated by the NationalCancer Institute, which plans to licenseit to a drug company for investigationas a possible treatment for HIV infec-tion. Prostratin is one of a range of po-tential drugs that have been discoveredthrough recent ethnobotanical research.

ISLANDS, PLANTS, AND POLYNESIANS: ANINTRODUCTION TO POLYNESIAN ETHNO-BOTANY. Edited by Paul A. Cox and San-dra A. Banack. Dioscorides Press, 1991.

A NONPROMOTING PHORBOL FROM THESAMOAN MEDICINAL PLANT HOMALAN-THUS NUTANS INHIBITS CELL KILLING BYHIV-1. K. R. Gustafson et al. in Journalof Medicinal Chemistry, Vol. 35, No. 11,pages 1978Ð1986; May 29, 1992.

RAINFOREST REMEDIES: 100 HEALINGHERBS OF BELIZE. R. Arvigo and M. Bal-ick. Lotus Press, Twin Lakes, Wis., 1993.

THE DEVELOPMENT OF AN ETHNOBIOMED-ICAL FOREST RESERVE IN BELIZE: ITS ROLEIN THE PRESERVATION OF BIOLOGICALAND CULTURAL DIVERSITY. M. J. Balick,R. Arvigo and L. Romero in Conserva-tion Biology, Vol. 8, No. 1, pages 316Ð 317; March 1994.

OH

H3C

O

H

O

OPROSTRATIN

H

OH

H

CH3

CH3

CH3

H3C

OH


EIGHT-CELL EMBRYOS like the one shown atthe left can now be screened for the presenceof some genetic illnesses. Baby Brittany NicoleAbshire (right) was tested in this way becauseboth of her parents are carriers of the fatalTay-Sachs disease gene.

TRENDS IN GENETICS

Grading the Gene Testsby John Rennie, staÝ writer

Almost nine months before she was born, Brittany Nicole Abshirepassed the most important test

she will ever take. Her parents, Reneeand David, are both healthy carriers ofthe trait for Tay-Sachs disease, a cruel-ly disabling and ultimately lethal in-heritable disorder. After they lost onedaughter to Tay-Sachs in 1989, theyswore they would never have anotherchild unless they could be sure that itwould be free of the disease. Genetictests could diagnose the condition be-fore birth, but the AbshiresÕ religiousbeliefs ruled out abortion as a way ofscreening for healthy fetuses.

There seemed to be no hope untilthe Abshires learned about a new tech-nology called preimplantation genetictesting. The experimental procedurehad already been used to screen morethan a dozen children for a diÝerentgenetic disorder, cystic Þbrosis. GaryD. Hodgen and specialists at the JonesInstitute for Reproductive Medicine atthe Eastern Virginia Medical Schoolcollected ova and sperm from the Ab-shires and successfully fertilized sev-en ova in vitro. After three days, whenthose seven had developed to aboutthe eight-cell stage, HodgenÕs teamplucked a cell from each pre-embryoand tried to analyze its DNA.

For four of the pre-embryos, the anal-ysis worked: one of them showed thedeadly combination of genes, but threewere not even carriers. Those three pre-embryos were implanted in Renee, andone survived to become Brittany, whowas born this past January. Courtesyof genetic testing, Brittany is the Þrstchild ever certiÞed to be free of Tay-


From just a snippet of DNA, geneticists can sometimes forecast a patient’s health. But ethical problems surrounding

this testing are as ominous as the diseases themselves


Sachs disease before entering her moth-erÕs womb.

As the era of genetic testing dawns,miracles such as Brittany could becomecommonplace. Genetic testing is thefastest-growing area in medical diag-nostics: according to the OÛce of Tech-nology Assessment (OTA), the numberof genetic tests will increase 10-foldover the next decade. ÒPotential new ge-netic tests roll oÝ the conveyor belt ofthe Human Genome Project almost oncea week,Ó remarks Norman Fost of theUniversity of WisconsinÐMadison Medi-cal School. Hundreds of thousands offetuses are already being tested everyyear by techniques such as amniocen-tesis and chorionic villus sampling.

Tests are not just for the unborn:many can also be used to diagnose ill-nesses more accurately in children andadults. In the past year alone, research-ers have found genes associated withAlzheimerÕs disease, HuntingtonÕs dis-ease and colon cancer; they expect toÞnd a breast cancer gene almost anyday now. Tests based on those and oth-er discoveries could warn people thatthey are at special risk for those diseas-es. And used in conjunction with pro-spective therapies that replace defectivegenes with working ones, genetic testscould lead to real cures.

But many human geneticists and oth-er observers are concerned that the rap-id growth of genetic testing is already

posing ethical, legal, social and scientif-ic quandaries for which there will be noeasy answers. They fear that new testsare proliferating without adequate su-pervision. Because the genetics of dis-ease is proving to be more complexthan anticipated, the predictive powerand utility of some tests are in question.Yet states are enthusiastically institutingprograms for screening newborns thatmay be unnecessary. Meanwhile inves-tigators are beginning to see evidencethat Ògenetic discriminationÓ is costingsome people jobs and insurance.

ÒWe need to proceed cautiously, be-cause there is a potential for doingharm with this technology,Ó warns ge-neticist Michael M. Kaback of the Uni-versity of California at San Diego, a pio-neer in population screening. When weknow more about a humanÕs geneticmakeup than ever before, will we knowwhat to do with all that information?

A Mythological Model

Genetic testing is not a single technol-ogy. Rather it refers to a broad range ofmethods for gauging the presence, ab-sence or activity of genes in cells. At therelatively low-tech end, researchers cancount the chromosomes in a patientÕscells or measure the amount of telltaleproteins in his or her blood. At the mostsophisticated level, researchers assay acellÕs DNA with molecular probes that

can Þnd a speciÞc genetic sequenceamong the three billion base pairs thatmake up human DNA. Some tests costas little as $50, whereas others are morethan $1,000. With these tests, medicalgeneticists can try to predict the courseof a patientÕs health.

Unfortunately, the more that research-ers have learned about human genet-ics, the more they have come to appre-ciate that even seemingly straightfor-ward diseases are complicated. Notionsof genetic illness have often been builtaround the single-gene model, in whicha defect in a gene causes a particularhealth deÞcit. Some diseases do workthis way: the deformed blood cells ofsickle cell anemia are caused by a genethat makes an abnormal form of hemo-globin; the fatal miseries of Tay-Sachsresult from the lack of an enzyme thatbreaks down fatty substances in neu-rons. But this model is turning out tobe an oversimpliÞcation.

No more than about 3 percent of allhuman diseases are caused by defectsin a single gene, and none of those aremajor killers, as are heart disease andcancer. The more complex conditionsinvolve a host of genes that merelynudge a personÕs predisposition to de-velop an illness. According to most es-timates, everyone carries at least Þve to 10 genes that could make him or her sick under the wrong circumstanc-es or could adversely aÝect children.ÒWeÕre all mutants,Ó Kaback summariz-


GENES ARE SHARED by members of a family. If one person carries a gene for a dis-ease, then each of his or her parents, siblings and children has a 50 percent chanceof carrying the same gene. That fact aÝects the privacy of genetic data.

0.5PARENTS

0.5

0.5SIBLINGS

0.5 10.5PATIENT

0SPOUSE

0.5CHILDREN

0.5

Numbers represent the proportion of genes that individuals share on average with the patient.


es. ÒEverybody is genetically defective.ÓThe truth is that the rules for what

constitutes a genetic disease are notclear-cut. If researchers someday Þnd agene that confers a 60 percent predis-position for gross obesity, is that a ge-netic defect? What about a gene thatgives a 25 percent predisposition forcardiovascular disease at age 55? OrÑmoving into an even more ambiguousareaÑa gene that predisposes to anti-social behavior?

Furthermore, even some diseases thatonce appeared to Þt the single-genemodel on the basis of their hereditarypatterns are more variable than hadbeen assumed. Cystic Þbrosis, one ofthe most common hereditary disordersamong people of European descent, isa useful example. Its symptoms includethe accumulation of suÝocatingly thickmucus in the airways and often severedigestive problems. Today drug treat-ments allow half of all cystic ÞbrosissuÝerers to live to about age 30, butjust a couple of decades ago patientsrarely survived into their twenties, andsome still die in infancy.

Researchers had often prayed for agenetic test that could Þnd carriers ofthe disease in the general population.Then, in 1989, investigators at the Uni-versity of Michigan and the Universityof Toronto found the gene responsiblefor cystic Þbrosis on chromosome 7; itencodes a protein in cell membranesthat aÝects the intracellular balance ofchloride ions. DNA-based tests for mu-tations commonly associated with thedisease soon appeared.

But those tests have revealed a newclass of cystic Þbrosis patientsÑpeoplewith relatively minor symptoms, such

as asthma or bronchitis, who neverthought of themselves as genetically ill.Some male patients are perfectly healthyexcept that they are infertileÑfor rea-sons not yet understood, they lack avas deferens, the tube in the reproduc-tive system that conducts sperm cellsfrom the testicles. Basically, cystic Þ-brosis is not a single disease after all.

Moreover, molecular biology has re-vealed that cystic Þbrosis is not causedby a single type of mutation. Althoughone mutation is associated with 70 per-cent of all cases, and two others withanother 15 to 20 percent, more than360 mutations have been linked to cys-tic Þbrosis so far. No one has yet beenable to correlate Þrmly the severity ofthe disease with diÝerent mutations.And DNA tests designed to catch onemutation will miss others.

All these discoveries make it muchharder to interpret the results of genet-ic tests for cystic Þbrosis. A positive testresult does not indicate how severelyaÝlicted a patient will be, and a nega-tive test result could be misleadinglyreassuring. Thus, the DNA tests usuallyneed to be conÞrmed by biochemicalassays and monitoring for symptoms.

The problem confounding the single-gene disease model and the straight-forward interpretation of the tests iswhat many observers call the myth ofgenetic determinism. ÒGenetic determin-ism is one of these simpleminded errorsthat we were prone to commit when wethought genes linked to diseases in akind of inevitable, ineluctable fashion,Óexplains Thomas H. Murray, director ofthe Center for Biomedical Ethics at theCase Western Reserve University Schoolof Medicine and former head of a task

force for the Human Genome Project.ÒIt invites you to think that Ôgenes equalfate.Õ Ó Environmental circumstances, inthe form of modiÞed diets, therapeuticdrugs, behavioral changes and otherinßuences, can avert many disastrousoutcomes foretold in the DNA. Con-versely, because cystic Þbrosis, heartdisease, cancer, autoimmune disorders,multiple sclerosis and other conditionsarise from an unfortunate conßuenceof genetic and environmental factors,genetic tests for those illnesses cannever by themselves predict an individ-ualÕs future with perfect clarity.

Damned by DNA

Against this backdrop of genetic in-terpretation (and misinterpretation),the drama of population screening isbeing played. In recent decades, severalscreening programs aimed at detectinggenetic diseases in large groups of peo-ple have been attempted, some withgood results, some with bad.

One nightmarish example of well-intentioned testing gone wrong is thescreening campaign for sickle cell ane-mia during the early 1970s. In re-sponse to a groundswell of demandsthat something be done about the dis-ease, which is a scourge of the African-American community, the federal gov-ernment funded a screening programto detect carriers of the sickle cell gene.The program was easy to implementbecause carriers could be identiÞedfrom just one drop of blood throughan inexpensive test. At Þrst, the screen-ing enjoyed popular support. Someministers conducted tests on their con-gregations; some members of the Black


CHROMOSOME MAPS of the genetic traits for a variety of ill-nesses are still being compiled. Disease-related genes have al-ready been located on each of the chromosomes, and new

ones are constantly being discovered. The relation betweengenes and diseases is rarely simple, however. Almost all dis-eases are likely to have some genetic component.


Panthers were oÝering the tests door-to-door in black communities.

But soon things turned ugly. Becausepeople were rarely educated about themeaning of the tests, many perfectlyhealthy carriers of the trait were led tobelieve they were sick. This ignoranceextended to some state governments aswell. The Massachusetts legislature, forinstance, passed a law requiring that allchildren at risk for Òthe diseasesÓ sick-le cell anemia and sickle cell trait bescreened before enrollment in school.ÒBy legislative Þat, sickle cell trait be-came a disease,Ó Kaback moans.

Some insurance companies began todeny coverage to black carriers on thegrounds that they had a preexistingmedical condition or that their childrenwere bad risks. The U.S. Air Force Acad-emy rejected black applicants who werecarriers. Some commercial airlines re-fused to hire carriers as ßight atten-dants because of the erroneous beliefthat such individuals were particularlylikely to faint at high altitudes. Promi-nent scientists suggested on networktelevision that the best solution to theanemia problem would be for blackscarrying the gene to forgo breedingÑasuggestion that naturally fed fears thatthe screening was really genocidal inintent. In short, the testing did moreharm than good by becoming a tool oflong-standing prejudices.

The Tay-Sachs program, which Ka-back organized, is a shining example ofwhat can go right with such an eÝort.Tay-Sachs disease is especially preva-lent among Jews of eastern Europeandescent. (The Abshires are not Jewish,but they come from a small communityin Louisiana where the disease is alsocommon.) Since the early 1970s morethan a million Jews throughout theworld have volunteered for testing tolearn whether they are carriers of therecessive Tay-Sachs trait. When coupleswho both carry the mutation decide tohave children, they typically elect tohave prenatal testing. If a fetus has thedisease, they usually abort it ratherthan give birth to a child who wouldsuccumb within Þve years to a horriblyslow, painful death. More important,however, the tests also set at ease theminds of fearful couples who mightotherwise never risk having children.

Why did the Tay-Sachs program suc-ceed where the one for sickle cell failed?

Kaback and others credit the care thatwent into its implementation. Before thepilot programs in Baltimore and Wash-ington, D.C., began, 14 months werespent establishing contacts within theJewish community and educating poten-tial patients about the tests and theirimplications. People received extensivegenetic counseling both before and afterthe tests. Unlike the screening for sicklecell anemia, the Tay-Sachs program wasalways voluntary, which meant that peo-ple had the opportunity to prepare forthe consequences of the testing.

Another important diÝerence, Kabackargues, lies in the diseases themselves.Because Tay-Sachs is always so uniform-ly hideous in its progression, extreme-ly few people believe an aÝected childshould be brought into the world. Be-cause testing can prevent that tragedy,it carries a clear beneÞt. The beneÞts ofsickle cell testing are less distinct. Theseverity of the disease is variable andintermittent, and with prompt medicalattention, immunizations and antibiot-ic treatments (which alleviate the fre-quent bacterial infections that accom-pany the anemia), patients can live formany decades. Relatively few peoplebelieve aborting an anemic fetus on thebasis of a genetic test is humane, so thetest has less obvious utility.

Learning the Lessons

The Tay-Sachs model for screeninghas been successfully adapted for someother diseases. For instance, prudenttesting has greatly cut the incidence ofthalassemia, a genetic blood disordercommon among many people of Medi-terranean descent: in Sardinia, the rateof thalassemia has declined from onein 250 births to one in 1,200 birthsduring the past 20 years.

Applying the lessons learned fromthe Tay-Sachs and sickle cell experienc-es is not always easy, however, especial-ly for diseases in which the populationat risk is extremely large. Once againcystic Þbrosis serves as a useful exam-ple. In families with histories of the dis-ease, screening is usually accurate andbeneÞcial. But 80 percent of the chil-dren with cystic Þbrosis are born tofamilies without such histories, so near-ly all couples in the U.S. would need to


MASS SCREENING for sickle cell anemiawas at Þrst popular among African-Americans during the early 1970s. BoxerJoe Frazier (center ) is shown promotingone screening drive. Because the publicwas poorly educated about the mean-ing of the tests, the results were some-times misused to discriminate againsthealthy carriers of the trait.

TESTING FOR CARRIERS of cystic Þbro-sis has mushroomed in recent years.

GROWTH IN THE USE OF CYSTIC FIBROSIS CARRIER TESTS

(YEAR / NUMBER OF TESTS)

SOURCE: Office of Technology Assessment, 1992

1989 / 1,854

1991 / 9,310

1992 / 63,000

1990 / 6,114


be screened to prevent most cases.ÒHow do you get four million pregnantwomen a year into an educational set-ting?Ó Fost asks. Given that the numberof professional genetic counselors inthe U.S. is barely more than 1,000, cysticÞbrosis screening alone would swampthe countryÕs counseling resources.Most patients would have to get coun-seling information from their primaryphysiciansÑmany of whom have littleor no training in genetics, according tosurveys of the medical profession.

Yet many researchers believe, as Fostputs it, cystic Þbrosis screening is Òme-tastasizing, despite the lack of any evi-dence that it works.Ó For the past eightyears, investigators in Wisconsin havebeen conducting a double-blind studyto determine whether biochemicallyscreening newborns for cystic Þbrosisis beneÞcial. So far they have found noevidence that identifying the childrenat birth is better than waiting for anysymptoms to emerge. Nevertheless, in

1989 the states of Colorado and Wyo-ming both made cystic Þbrosis screen-ing mandatory for all newborns.

Because all genetic tests have somemargin of error, excessive newborn test-ing has the capacity to do harm, if onlyby worrying parents unnecessarily. Onestudy looked at families in which a childwas initially diagnosed as having cysticÞbrosis but was later shown to behealthy; one Þfth of the parents none-theless continued to fear that their chil-dren had the disease. Geneticists andethicists often express concern abouthow stigmatizing children with a diseaselabel may warp personal development.

In a report issued last November, theInstitute of Medicine ( IOM) of the Na-tional Academy of Sciences oÝered nu-merous recommendations about howgenetic testing should be conducted tominimize its potentially adverse eÝects.All testing, it suggests, should be vol-untary, and the results should be keptconÞdential to prevent misuse. All test-

ing should be linked to genetic counsel-ing, so that the patient understands theresults and their implications. For themost part, tests should be restricted tothose conditions for which some bene-Þcial intervention is possible, either astherapy or as reproductive planning.Tests should not usually be performedpurely for informationÕs sake; rather thepatient should be able to use the resultto make an informed decision about anissue of immediate relevance, such ashaving a child or electing a medicaltreatment. The IOM also strongly be-lieves an individual should decide freelywhether or not to be tested, without so-cial pressure or Þnancial inducement.

Unregulated Testing

As good as those guidelines may be,abiding by them will be diÛcult. Oneproblem with trying to set any limits isthat genetic testing is so easy to do; an-other is that the Þeld is largely unregu-lated. ÒCurrently there is very little tostop someone from implementing ge-netic tests on a population basis with-out the sort of institutional review andinformed consent required for othernew technologies,Ó Fost says.

Neil A. Holtzman, a health policy ex-pert at the Johns Hopkins UniversityHospital and an editor of the IOM re-port, notes that most companies in thebusiness of genetic testing oÝer it as alaboratory service and not as a kit forphysicians to use. They are thereforenot obliged to submit their methods forappraisal and approval by the Food andDrug Administration. The companies dofall under the jurisdiction of the Clini-cal Laboratory Improvement Amend-


WHO SHOULD KNOW the results of ge-netic tests is an unsettled ethical and le-gal issue. Because the results of genetictests may carry health and Þnancial im-plications for others, a patient or physi-cian may be obliged to divulge that in-formation. In a 1992 poll, a majority ofAmericans said the privacy of test re-sults should not be absolute.PATIENT’S FAMILY

PATIENT PHYSICIAN

PATIENT’SEMPLOYER

PATIENT’SINSURER

? ?

? ?

?

? ?

NO41 PERCENT

YES57 PERCENT

WHO SHOULD KNOW?

OPINIONS OF THOSE WHO BELIEVE SOMEONE ELSE DESERVES TO KNOW (PERCENT)

0 10 20 30 40 50 60 70 80 90 100

NOT SURE2 PERCENT

THE RIGHT TO KNOWIf someone is a carrier of a defective geneor has a genetic disease, does someone else deserve to know?

SPOUSE OR FIANCÉ

OTHER IMMEDIATE FAMILY MEMBERS

INSURER

EMPLOYER

SOURCE: March of Dimes/Lou Harris, 1992


ments of 1988. This law ensures thatlaboratories conducting interstate com-merce in Pap smears and other biomed-ical tests meet certain standards of re-liability. Unfortunately, Holtzman says,the Health Care Financing Administra-tion, the agency empowered with en-forcing those rules, Òhas really draggedits feet on setting up guidelines forthese new genetic tests.Ó

According to the IOM report, only 10states have established any licensingrequirements for genetic testing labo-ratories, and only New York State hascomprehensive regulations that pertainto DNA tests. This lack of oversight wor-ries Holtzman and the rest of the IOMcommittee about the possible marginof error in the test results. ÒI think weÕvealready seen a couple of examples ofcompanies seemingly getting tests outthere without any regulatory brakes be-ing put on,Ó Holtzman charges. ÒIf wedonÕt get the proper regulatory author-ity in place, itÕs going to become a prob-lem of too little, too late.Ó

Critics also point out that many ge-netics researchers have Þnancial ties tocompanies in the business of testing.ÒWhat I see my colleagues doing is iso-lating a gene, Þnding a single mutationand then jumping into populationscreening. ThatÕs not the way it shouldhappen, in my opinion,Ó Kaback says.ÒWeÕve got entrepreneurial interests in-ßuencing judgments that people aremaking about when tests are ready tobe deployed in the population.Ó

Kaback and the rest of the IOM com-mittee also worry about pressures be-ing put on physicians to use genetictests. ÒThere are private companiessending letters to doctors all over thecountry, telling them they should beoÝering cystic Þbrosis carrier testing toall their pregnant couples,Ó he says. Inthese litigious times, many doctors mayfeel it is safer to order a test than toface charges of negligence.

Because of such concerns, severalprofessional groups have issued state-ments that emphasize the experimentalstatus of most genetic tests. The Amer-ican Society of Human Genetics hastwice announced that oÝering screen-ing for cystic Þbrosis carriers to the

general public should not be consid-ered the standard in medical practice.This past March, the National AdvisoryCouncil for Human Genome Researchat the National Institutes of Healthwarned that screening for cancer shouldnot be performed widely until more re-search on its reliability and consequenc-es could be determined.

Genetic Privacy

Even if the accuracy and utility of ge-netic testing are assured, maintainingthe privacy of that information will re-main a problem. Your genes are not ex-clusively your own: you share half ofthem with each of your parents, sib-lings and children. If you discover thatyou carry a worrisome gene, you mayhave an ethical, if not legal, obligationto tell them. ÒThere are ripples from ge-netic testing that donÕt have analoguesin most other kinds of medical testing,Óremarks Arthur Caplan, a bioethicist atthe University of Minnesota.

A 1992 March of Dimes poll reportedthat 57 percent of the public thinkssomeone other than a patient deservesto know that he or she carries a defec-tive gene. Of those who believed so, 98percent thought a spouse or betrothedshould know. More surprisingly, how-ever, 58 percent thought insurance com-panies should also be informed, and 33percent thought an employer shouldbe told. In a diÝerent study, physiciansthemselves disclosed a willingness toviolate patient-doctor conÞdentiality insome cases: 54 percent said that, evenover a patientÕs objections, they wouldtell relatives at risk about the results ofa test for HuntingtonÕs disease (a lethalneurodegenerative disorder that usual-ly manifests in middle age). Twenty-four

percent said they would tell the pa-tientÕs employer, and 12 percent wouldtell an insurance company.

Industry has also revealed an appe-tite for individualsÕ genetic information,although the limited predictive abilityand high cost of the current tests seemto have restricted their appeal. A sur-vey by the OTA released in 1991 triedto determine whether employers wereconducting genetic tests as conditionsof employment; such tests could theo-retically reveal workers who would beparticularly bad (or expensive) healthrisks. It found that in 1989 only 12 of330 Fortune 500 companies were mon-itoring or screening for any reason. Butmore than half of the polled compa-nies found the idea of monitoring ac-ceptable, and 40 percent admitted thata personÕs health insurance costs mightaÝect his or her chance of employment.

Bioethicist Thomas Murray has alsofound interest in genetic data amonginsurance companies. ÒIf you ask thequestion narrowly, ÔAre insurers requir-ing DNA tests of customers?Õ, the an-swer is no,Ó he says. ÒThe tests are tooexpensive, not cost-eÝective, there arenÕtenough of them yetÑthere are a lot ofreasons. But are insurers interested ingenetic information, and will they be inthe future? The answer to that is a re-sounding yes. The best genetic infor-mation right now comes not from ge-netic testing but from oneÕs personalhealth historyÑwhat did your parentsdie of?Ó

The insurance industry itself conÞrmsthat view. A joint report issued in 1991by the American Council of Life Insur-ance and the Health Insurance Associa-tion of America stated that althoughinsurers would not be requiring tests inthe foreseeable future, they should be


BREE WALKER LAMPLEY, a televisionand radio celebrity in Los Angeles, hasectrodactyly, a genetic condition thatcauses a deformity of the hands andfeet. She was publicly criticized for hav-ing a baby in 1991 because her childhad a 50 percent chance of inheritingthat trait. (Her son did indeed inheritthat gene.) Her case illustrates the so-cial pressures that can be imposed onindividuals who carry genetic defects.


entitled to know the results of any ge-netic tests or other evidence of possibledisease in a policyholderÕs record. In-surers maintain that they need this in-formation not to discriminate geneti-cally but to set fair rates for all policy-holders. It would be wrong, they argue,to make healthy people pay higher pre-miums because they had been lumpedin with those at higher risk. They alsodo not want to exempt genetic testingresults from scrutiny because they saythis precedent might someday precludethem from using other medical and sta-tistical indicators of risk.

These arguments do not persuadeMurray. ÒInsurers are applying a modelof Ôactuarial fairnessÕ and justice theydeveloped in the realm of commercialinsurance,Ó he replies. Those principles,in his opinion, are not relevant to healthinsurance, which serves social and hu-manitarian functions that deserve con-sideration. Because no one can controlhis or her genetic makeup, it seemswrong to penalize individuals for it.

Many observers also doubt the abili-ty of insurance companies and otheragencies to interpret genetic informa-tion wisely. Paul R. Billings, a geneticistnow at the Palo Alto Veterans Adminis-tration Medical Center, has been col-lecting examples of people who wereapparently discriminated against be-cause of genetic information that be-came known to insurers, employers,health maintenance organizations(HMOs) and adoption agencies. He andhis colleagues Þrst publishedsome of their results in theAmerican Journal of Human Ge-

netics in 1992. Several of the cas-es they described concerned peo-ple who were healthy carriers ofgenetic diseases or had extreme-ly mild symptoms yet were stilldenied jobs or insurance cover-age. One woman who applied tobecome an adoptive parent wasrejected because her family his-tory of HuntingtonÕs diseasemade her Òtoo great a risk.Ó

In another case, parents whohad one child with cystic Þbrosisconceived again, and prenataltesting conÞrmed that this sec-ond child, too, would have the

disease. When the familyÕs HMO learnedthat the couple intended to proceedwith the pregnancy, it moved to with-draw or limit the entire familyÕs healthcoverage. Only after threats of a law-suit did the HMO change its mind.

Billings says a second report, listingabout 100 cases of genetic discrimina-tion, is almost done. With funding fromthe Human Genome Project, he and hisco-workers have also been conductinga further survey of 30,000 people withgenetic conditions. ÒOur preliminaryview of that study is that it conÞrmsgenetic discrimination is occurring,Ó henotes. A 1992 OTA report also foundthat about 15 percent of genetic coun-selors said some of their clients had ex-perienced discrimination in insurance.

Outside the Law

Anyone looking to the law for pro-tection from genetic discriminationwould Þnd a thin shield at best. At thefederal level, most experts agree, thestrongest statute for barring discrimi-nation in the workplace is the 1990Americans with Disabilities Act. But therelevance of that act to genetic infor-mation is still uncertain. The Equal Em-ployment Opportunity Commission,which enforces the act, has already of-fered the opinion that healthy carriersof genetic diseases would probably notqualify as having a disability.

Some states have laws that protectthe carriers of genes for sickle cell ane-

mia and a few other speciÞc traitsfound disproportionately among Afri-can-Americans and other groups whosemembers have been historical targetsof discrimination. Somewhat broaderlegislation has been passed in Wiscon-sin and a few other states and is beingdebated in a number of others. In 1991the California legislature passed a lawthat prevented any form of genetic dis-crimination, but it was subsequentlyvetoed by Governor Pete Wilson.

Because most concern about discrim-ination revolves around insurance cov-erage, insurance reform may be the keyto a solution. ÒOur task force recom-mended that all individual risk infor-mation be excluded from decisionsabout who gets insured, what they getinsured for and how much they getcharged,Ó Murray says. ÒWe see no oth-er practical, sustainable plan for healthcare coverage than community rating.ÓCommunity rating, which was the basisfor the Þrst health insurance programsduring the 1930s, is a system in whicha customerÕs premiums are determinedby the health proÞle of his or her com-munity. Genetic information about in-dividuals would be irrelevant.

The IOM committee and others alsofavor community rating for that rea-son. Insurers disagree, arguing that in-dividual risk rating serves the publicwelfare more equitably at less expense.Nevertheless, for many reasons, thepopularity of individual risk rating hasbeen declining. In recent years, New

York, Maine and a few otherstates have shifted to programsbased at least partially on com-munity rating.

Both insurers and their criticsagree that reforms in health careÞnancing are likely to rendermoot someÑbut not allÑof thedisputes about genetic discrimi-nation in insurance. The detailsof the new health care systemwill be important. For example,some insurance plans that wouldmeet the Clinton administrationÕsstandard of universal coveragemight still be able to disallowcoverage for certain genetic con-ditions. People with those traitsmight then have to buy addition-al insurance or pay some costsout of pocket. Under the currentsystem, insurance companies andHMOs often resist paying formany genetic tests or for thecosts of genetic counseling. It re-mains to be seen how reforms inhealth care Þnancing will affectthose policies.

Universal, comprehensive cov-erage could also become the in-


DOWN SYNDROME patients, whocarry three copies of chromosome21, exhibit a range of mental andphysical impairments. But giventhe proper education and medicalattention, many of them are ac-complishing much more than wasonce believed possible. Predict-ing the limits of people with ge-netic handicaps is uncertain.


strument of a demon that has longhaunted genetic technology: eugenics.The word immediately summons mem-ories of the Nazi genocidal horrors andother atrocities, such as the forcedsterilization of 30,000 people as Òmen-tal defectivesÓ in the U.S. before WorldWar II. Yet eugenics can arise through aseemingly more benign movement toration social resources. Last December,China announced a policy of discourag-ing people at risk for hereditary dis-eases from having children, on thegrounds that the genetically ill imposetoo much of a burden on society.

A Eugenic Democracy

The U.S. is not immune from suchthinking. ÒIf we get universal healthcare, people are going to want to knowwhy they should be paying for the Ôge-netically irresponsible,Õ Ó Caplan be-lieves. ÒIn this country, eugenics is notgoing to come from a Hitlerian dictatorsaying, ÔYou must do this.Õ ItÕs probablygoing to come from a society saying,ÔYou can have a kid like that if youwant, but IÕm not paying.Õ Ó

A counterargument might be that ina democracy, the public should be enti-tled to set limits on what costs individ-uals can impose on everyone. Yet sucharguments, like those surrounding in-surance, invariably depend on a deÞni-tion of fairness. Moreover, whatever mo-tivations lie behind the pressures beingput on people making reproductivechoices, their net eÝect is the same: so-ciety inßuences who will or will not beborn. As Billings says, ÒWhether wewant to dress it up in economic incen-tives, if you have to sell your house andgo broke to have a child of a certaintype, thatÕs eugenics.Ó

Because of its history, ÒeugenicsÓ isperhaps too loaded a word to describesome applications of genetic testing.Holtzman and many others prefer toreserve it to describe circumstances inwhich a government or other agency in-tervenes in reproductive decisions. Thatis why, Holtzman says, the IOM com-mittee so emphasized the importanceof preserving personal autonomy: cou-ples should weigh for themselveswhether the risks of having a sick childoutweigh other considerations.

ÒI think the goal of eliminating dis-eases and disabilities is a good one,ÓCaplan aÛrms. ÒI donÕt think thereÕsanything wrong with encouraging wom-en at risk to be screened for spina biÞ-da or to ask Jewish couples of easternEuropean descent to be screened forTay-Sachs. ItÕs wrong to confuse thegoal of eliminating disease with themoral problem of coercion.Ó

The troublesome fact remains, how-ever, that the line between genetic dis-eases and undesirable traits is blurred.Caplan cites the example of albinism,which is not a disease but is associatedwith a higher risk for skin cancer, faultyvision and social stigma. ÒI think thestandard that medicine wants to go withis, is there clear dysfunction or disor-der? If not, I would argue that medicineought not to be testing for it and coun-seling for it. If weÕre in the gray zone, Isuspect we should try to stay out ofthose areas, because thereÕs so muchelse of clear value that we could be do-ing,Ó he suggests.

Yet Caplan admits thatÑas much ashe would like toÑhe cannot frame aconvincing standard that would allowparents to select some of their chil-drenÕs traits but not their sex, height orother cosmetic features, presumingtechnological feasibility. Distaste forabortion will stop many parents fromexercising that veto for now. But suchchoices could be circumvented byemerging technologies for screeningembryos before they are implantedÑoreven before conception. ÒSo I think thestance that we will deal only with clear-cut disorders will last about Þve min-utes,Ó Caplan answers ruefully. ÒOnceyou can actually do that testing, the in-terest will swamp my objections. Theability to choose the traits of your childwill roar through with a whoosh.Ó

The consequences of those choicesmay be very hard to predict. Murray re-calls that at a meeting of the AmericanSociety for Human Genetics a few yearsago, he heard about a deaf couple who

had asked one geneticist whether thehearing ability of a fetus could be de-termined prenatally: they wanted to gothrough with the pregnancy only ifthey knew that the child would be deaf.

As Brittany Abshire and millions ofother healthy children and adults prove,genetic testing can immeasurably im-prove the quality of life for individuals,even entire families. To ignore the goodit can do would be an act of immoralblindness and cowardice. But usingthese technologies wisely will demandforesighted social and legal policies.The record is discouraging: in the past,when genetic discoveries have madetests possible, policymakers have ofteneither encouraged them prematurely oracted too late.

ÒThe paradigms of eugenics are pro-grams of unsurpassed evil. TheyÕre notgoing to get any less evil just becauseour genetics got better,Ó Murray muses.ÒWe need to be very conscious of whatweÕre dabbling in.Ó


INSURERSÕ ATTITUDES toward people who carry the genes for disease traits mayshape policies concerning the eligibility of those persons and their oÝspring forhealth care coverage. Critics worry that genetic discrimination could occur.

FURTHER READING

CYSTIC FIBROSIS AND DNA TESTS: IMPLI-CATIONS OF CARRIER SCREENING. Con-gress of the United States, Office ofTechnology Assessment. Report OTA-BA-532, August 1992.

MOLECULAR BIOLOGY: IMPACT ON HU-MAN DISEASE. Theme issue of FASEBJournal, Vol. 6, No. 10; July 1993.

ASSESSING GENETIC RISKS: IMPLICATIONSFOR HEALTH AND SOCIAL POLICY. Editedby Lori B. Andrews, Jane E. Fullarton,Neil A. Holtzman and Arno G. Motul-sky. National Academy Press, 1994.

STATEMENT: THE CARRIER OF A GENETIC DISEASE TRAIT HAS A PREEXISTING CONDITION

SOURCE: Office of Technology Assessment, 1992.

SURVEY RESPONDENTS (PERCENT)400 10 20 30 50 60 70 80

COMMERCIAL INSURERS

HEALTH MAINTENANCE ORGANIZATIONS

BLUE CROSS/BLUE SHIELD PLANS

AGREE

DISAGREE

NO RESPONSE


SCIENCE AND BUSINESS

Cyberspace CadetsGates and McCaw takenew venture into orbit

The proposal by Teledesic in Kirk-land, Wash., to launch an 840-sat-ellite communications network

bears a peculiarly American imprima-tur: an appeal to the utopian strain inthe national character and a conÞdencethat no vision is beyond the reach of in-novation and hard work. But like manysuch utopian visions, it also manifestsa grandiosity that pushes sanityÑin thiscase, the engineering and Þscal vari-etiesÑto its limit.

The plan calls for spending $9 billionto circulate some 900 satellites (840 ac-tive and up to 84 spares) in 435-mile-high orbits to deliver telephone, videoand computer data services to the en-tire world. That armada outnumbers bya factor of at least two all the commer-cial communications satellites current-ly in orbit.

Teledesic must position the 900 sat-ellites precisely so that they hurtle in21 loops around the North and SouthPoles; each orbital plane would containup to 44 satellites. A satellite wouldcommunicate with eight other satellites:two to the front and two to the rear in

the same orbital path, and two othersto each side in adjacent orbits.

Each bird, weighing 1,700 pounds atlaunch and measuring about 40 feet indiameter when fully deployed, carriesan advanced packet switch for relayingvoice, video and data. These asynchro-nous transfer mode switches, as theyare known by techies, have only recent-ly entered commercial use in terrestrialtelephone networks.

To meet an ambitious 2001 deadline,Teledesic would have to launch about a rocket per week or more over a two-year period, each one carrying eight ormore satellites. The estimated numberof components required for the totalcomplement of Teledesic satelliteswould be suÛcient to create an indus-try: the startupÕs application to the Fed-eral Communications Commission citesthe need to make 500 million galliumarsenide chips, more than have everbeen manufactured commercially.

ÒThe state of the art is such that wedonÕt know how to launch them, wedonÕt know how to keep them in orbitand we donÕt know how to build themat that price,Ó says Burton I. Edelson, a satellite communications specialist at George Washington University. ÒWehave a lot of problems maintaining sys-tem viability with a small number ofsatellites, even weather satellites.Ó

Also perplexing is the market forthese services. Teledesic portrays itselfto the FCC as an Internet for Monta-gnard tribesmen. ÒToday, the costs tobring modern communications to poorand remote areas is so high that manyof the worldÕs people cannot participatein the global community,Ó reads Tele-desicÕs FCC application. ÒYet the bene-Þts of the communications revolutionshould be extended to all of the worldÕscitizens, including those. . .who do nothave ready access to doctors, hospitals,schools, or libraries, and who are at riskof being shunted aside.Ó Hamlets andwadis may be well served. Not so thehordes of laptop computer and hand-held device owners: the technology op-erates with a Þxed ground antenna toprovide large amounts of digital com-munications capacity for video and datausers. TeledesicÕs market, Edelson says,could be more economically serviced bya few big satellites in the much highergeostationary orbit where they wouldcover a broad swath of the planet.

Winning endorsements from dozensof national telecommunications author-ities will prove as diÛcult as raising theprojected $9 billion. Furthermore, theintended cost of each satellite, at lessthan $10 million, is a small fraction ofthe price tag for other advanced com-munications satellites. ÒI donÕt know


SATELLITE is one of 840 in a worldwide network envisaged bythe newly formed corporation Teledesic, in which Craig McCaw(left) and Bill Gates (right) have invested. S

YG

MA

TE

LED

ES

IC

SY

GM

A

that anybody Þnds the $9-billion Þgurecredible; it will cost at least $25 to $35billion,Ó says Alan L. Parker, presidentof Orbital Communications, a Virginia-based company that plans to begin thelaunching of 26 small satellites for datatransmission this year.

A recipe for a technologically impos-sible and Þnancially dubious programon the scale of the Strategic Defense Ini-tiative? Then why are two of the initialÞnancial backers, William H. Gates IIIand Craig O. McCaw, preeminent amongentrepreneurs in computer softwareand telecommunications?

Neither tycoon is betting a lifeÕs sav-ings on TeledesicÕs success. Within daysof the Teledesic announcement, Micro-soft, of which Gates is chairman and amajor shareholder, agreed to build awireless network for data transmissionwith Mobile Telecommunications Tech-nologies, the largest U.S. paging concern.

In fact, Gates and McCaw have eachinvested but $5 million in Teledesic. Asubsidiary of McCawÕs cellular telephonecompany has put down an additional$3.7 million. If Teledesic somehow be-

came a success, these relatively smallinvestments could yield large payoÝs.John Pike, an analyst with the Federa-tion of American Scientists, thinks Gatesand McCaw may be interested in play-ing in low-earth orbit for some of thesame reasons that the U.S. Departmentof Justice antitrust lawyers are interest-ed in Microsoft. The game is calledmonopoly.

Teledesic does not intend to supplantthe local telephone company; it will sellcommunications capacity on its satel-lites only to existing service providers,many of them in developing nations.Bolstered by the names Gates and Mc-Caw, Teledesic may try to convincestate-held telephone companies to signon as investment partners for this or-biting version of the Internet.

If governments cast their lot withTeledesic, the competitionÑcompanieslaying Þber-optic cable, for exampleÑmight Þnd entry into these emergingmarkets blocked by the coopted bu-reaucracies. Even if a national telecom-munications market is open to competi-tors, the early arrival of Òan 800-pound

gorilla,Ó PikeÕs characterization of Tele-desic, would complicate later attemptsby others to enter this business.

Gates and McCaw could emerge asthe Harriman and Vanderbilt of the in-formation age, Pike says. ÒMost of theother cyberspace projects are just a bet-ter way to watch GilliganÕs Island,Ó headds. ÒIf this thing happens like theyhope it will happen, it would be a bigdeal, literally one of the deÞning reali-ties of the third millennium, in the waythat Ma Bell deÞned reality for most ofthe 20th century in the U.S.Ó

GatesÕs and McCawÕs thoughts abouttheir critics are hard to know becausethey do not share them publicly. (SCI-ENTIFIC AMERICANÕs requests for inter-views were turned down.) In any event,the original idea for Teledesic camefrom neither man, but from a venturecapitalist, Edward F. Tuck. For the past10 years, Tuck has been involved instarting new companies, among theman enterprise named Calling Communi-cations, which was intended to becomea worldwide telephone service using anetwork of communications satellites.


When polymer chemists want to fashion a material towhich nothing will stick, they often design a molecule

in which a fluorine atom takes the place normally occupiedby that of hydrogen in the ubiquitous hydrocarbons. Fluorineforms a tight covalent bond with carbon, keeping it fromcombining with atoms from other molecules.

The best known of these antistick concoctions is polytetra-fluoroethylene, aka Teflon. Dow Chemical has come up witha new fluorocarbon formulation whose abhesive (as opposedto adhesive) properties may be better than ordinary Teflon andthat may be of use where its frying-pancousin may not. The fluorochemical might,for example, serve not only as a protectivecoating for subway walls but as a surfacethat guards against barnacles on ships,dirt on wallpaper and ice on aircraft.

Unlike hard-to-apply Teflon, the Dowcoating can be painted onto a surfacewith a brush or spray gun. Cured at a rel-atively tepid 100 degrees Celsius for anhour, it forms a hard, clear finish. Teflonmust often be baked on metal at temper-atures above 250 degrees C. Once cured,the hardness of the finish left by the Dowpolymer resembles an automobile lacquer.Less porous than Teflon, the fluorocarbonstops contaminants from getting caughtbelow the surface. It also prevents thesurface from being “wetted,” or covered,by common solvents used in paints, dyesor inks, a quality of use in the war againstgraffiti. A felt-tip marker leaves tiny inkbeads that stick poorly on a surface treat-

ed with the Dow chemical but remain in place on Teflon. Still,Teflon is likely to retain its longtime affiliation with scram-bled eggs. The Dow fluorochemical does not tolerate heat aswell as Teflon does. It does, however, reside in a mostly wa-ter-based solution and so emits little of the organic vaporsthat might warrant scrutiny from environmental regulators.

Dow’s creation, which has yet to garner a trade name, ismanufactured by making a polymer that contains both fluo-rocarbon and hydrocarbon molecules. The fluorocarbons on these long polymer chains show a marked distaste for water.

Good-bye, Taki 183

STICK TEST demonstrates how the ink from a felt-tip marker forms into tinybeads on Dow ChemicalÕs ßuorinated polymer (horizontal band), but the markerdye remains on the Teßon substrate on both the top and bottom.

Tuck shared his idea with McCaw,who suggested that the network be setup primarily to carry data and video in-stead of regular telephone calls. Earlierthis year Calling Communications be-came Teledesic, with McCaw as chair-man and Tuck as vice chairman. It wasMcCaw who convinced Gates to comeon board as an investor. Kinship Part-ners II, in which Tuck is a principal,holds a $1.5-million stake in Teledesic.Tuck has remained fascinated for yearswith combining satellites and entrepre-neurialism. He is best known as found-er of a company called Magellan, whichis now one of the leading suppliers ofthe receivers for the Navstar global po-sitioning system, a network of 24 satel-lites that provides airplanes, boats, au-tomobiles and backpackers with pre-cise location information.

Tuck tries to make TeledesicÕs seem-ingly insurmountable costs look some-what manageable. They can, he says, beaddressed through economies of scalerealized from building hundreds of sat-ellites. Teledesic, in fact, will utilizemethods developed for manufacturing

the Brilliant Pebble satellites proposedfor Star Wars. It will also tap that pro-gramÕs investigations into automatedcontrol of large satellite networks.

The shift to small satellites reßectsthe broader trend toward miniaturiza-tion in the computer industry. W. Rus-sell Daggatt, a lawyer and the presidentof Teledesic, contrasts the Telepebblesto the huge communications satellitesin geostationary orbit at a height of22,300 miles. ÒI make the analogy to thecomputer industry that evolved frombig mainframes,Ó Daggatt says. ÒIf youwere thinking of sending up 840 ofthose powerful mainframes, peoplewould think you were crazy. But if youwere thinking of sending up 840 note-book computers, it is a diÝerent matteraltogether.Ó

Daggatt denies that Teledesic wantsto monopolize markets in developingnations. ÒOur total system capacity ona worldwide basis is inÞnitesimal com-pared with the need,Ó he asserts.

The brouhaha surrounding Teledesichas served as free publicity for a spateof other ventures that have targeted thelower reaches of outer space. Motorolaand partners plan to spend $3.3 billionto send up Iridium, a 66-satellite con-stellation. Loral, Qualcomm and ninepartners envisage spending $1.8 billionon Globalstar to launch 48 satellites.These projects are just two of a num-ber of plans for orbiting networks thatwould give portable telephone usersthe ability to place calls from virtuallyanywhere around the globe.

By using much higher frequencies,Teledesic, which is not a mobile ser-vice, is able to provide the network ca-pacity needed for video and high-speeddata communications as well as for or-dinary telephone calls. It can supportmore than 20,000 simultaneous digitalconnections, each carrying 1.54 millionbits per second of video or data.

But even MotorolaÕs and LoralÕs moremodest schemes may not escape un-scathed. Financing and construction ofa multisatellite network must be com-pleted before even the Þrst call isswitched. In contrast, Þber-optic or cel-lular telephone networks can be builtpiece by piece. Herschel Shosteck, aMaryland economist who has studiedthe wireless market, believes diminish-ing costs and advancing technology forland-based cellular networks will makesatellite telephone links uncompetitive.

Even on the earth, some grand cyber-space alliances have faltered, broughtdown by unpalatable Þnancial and reg-ulatory realities. Nevertheless, the Þeldis full of dreamers. And if they succeedin building, can the rest of us aÝordnot to come? ÑGary Stix


They tend to float to the surface of anaqueous solution, as if they were tiny mo-lecular buoys. The hydrocarbon moleculesthat remain submerged on the lower partof a chain contain ions that bind to a sur-face underneath. The result is a perfectyin-yang combination: abhesive side up,adhesive down. These ionic groups alsobind to another polymer, an oxazoline,that joins each buoy chain and helps to at-tach them to a surface. This cross-linkagefixes the chains rigidly in place.

Donald L. Schmidt, a Dow polymer chem-ist, along with a few colleagues, first de-vised the material seven years ago. Thecompany will now leave it to others to findcommercial uses for the product. “We’re nota coating company,” says Benjamin M. De-Koven, a project collaborator. Last Septem-ber, Dow signed a nonexclusive licensewith 3M, which is investigating a range ofapplications from finishes for kitchen coun-tertops to a potential addition to its Scotch-gard line of fabric coatings. 3M is alsoworking on getting the chemical to cure atroom temperature so that it can be sold asan ordinary self-drying finish like paint.

Keeping the writing off the wall may takea while longer. That means that Cool Boo-gie Dancer, Taki 183 and even good oldKilroy may still be around to leave a me-mento or two at the millennial New Year’scelebration. —Gary Stix

Requiem for Alpha?If Rep. Brown loses his game ofchicken, the space station dies

The planned international spacestation has fallen into dire politi-cal diÛculties that may lead to

its demise, a result of budgets belowthe expected levels and apprehensionsabout RussiaÕs large role in the pro-gram. In March, Representative GeorgeE. Brown, Jr., of California, the inßuen-tial chairman of the House Committeeon Science, Space and Technology, an-nounced his ÒregretfulÓ (and perhapsMachiavellian) ÒdecisionÓ to oppose thespace station unless two conditions aremet. The Þrst is that the Clinton admin-istrationÕs proposed budget for the Na-tional Aeronautics and Space Adminis-tration in 1995Ñ$14.3 billionÑsurviveits appropriations battles unscathed.That stipulation alone would requireunusual cooperation among legislators.But BrownÕs second condition for sup-porting the station puts him in a directbattle of willsÑand perhaps witsÑwiththe White House.

Brown says he will oppose the spacestation, now provisionally known as In-

ternational Space Station Alpha, unlessfunding for NASA after 1995 is increasedabove the administrationÕs publishedplans. Those plans call for only a smallraise by 1999, to $14.6 billion. Brownfears that, without the increases heseeks, the space station will Òdrain thehealth, vitality and future of the rest ofNASA.Ó

Because Brown is a long-term sup-porter of the space program, his viewswill carry considerable weight if thereis a tight vote. Last year authorizationfor the station was approved by a mar-gin of just one vote, and the opponents,led by Representative Tim Roemer ofIndiana, are hoping they will succeed incanceling the program in 1994.

NASA estimates that the remainingcost of bringing the station to full oper-ation in 2002 is $17.4 billion. Accord-ing to Marcia S. Smith of the Congres-sional Research Service, $11.2 billionhas already been spent over the pastdecade on successive designs and re-designs. Alpha, unveiled in late 1993,represented the Þfth and biggest re-design since the space station programbegan in 1984. Now it is being modiÞedto accommodate Russian components.

Three weeks after his ultimatum, theCongressional Budget OÛce handedBrown ammunition in the form of a report that paints a gloomy view ofNASAÕs future. The budget oÛce ana-lysts doubt the agencyÕs ability to im-

prove eÛciency. They make it clear thatif the space station is to go forwardÑand budgets do not increase above cur-rent levelsÑplans for unmanned spaceexploration will have to be scaled back.

Disclosure earlier this spring of a dis-pute between NASA and the space sta-tionÕs prime contractor, Boeing, hasmuddied the waters further. The twosides were about $800 million apart atthe beginning of April. Both expressedoptimism that the gap could be bridged,but the fracas strengthens the hand ofAlphaÕs opponents.

A more substantial cause for alarmamong station supporters is the grow-ing misgivings about RussiaÕs politicalcourse. The purpose of bringing theU.S.Õs former cold war adversary intothe game was to save money and to per-suade Russia to comply with restric-tions on exports of missile technology.But the strategy may yet backÞre. Ac-cording to plans NASA was publicizingin April, Russia would launch seven ofthe Þrst 11 station assembly missions.Moreover, Russia would provide thepropulsion module that is essential tokeep Alpha in orbit.

The heavy reliance on Russian hard-ware has disturbed some longtime sta-tion supporters, including Brown andRepresentative F. James Sensenbrennerof Wisconsin. Sensenbrenner now sayshe will vote against the program unlessNASA can produce a credible backupplan so that the U.S., together with Eu-

rope, Canada and Japan, could com-plete Alpha without Russian help. Fur-thermore, Smith observes, the amountthat Russian involvement will save hasshrunk. It appears to be no more than$2 billion, half the Þgure announcedlast year by NASA administrator DanielS. Goldin.

The space station has survived eightprevious attempts to cancel it, and itmay survive another. Some develop-ments are in its favor : in late March theprogram, which is now being run by a new management team at JohnsonSpace Center in Houston, successfullypassed two hurdles when it completeda systems design review as well as aninspection by a panel chaired by CharlesM. Vest, president of the MassachusettsInstitute of Technology. VestÕs paneloutlined options for the space stationlast summer. Louis J. Lanzerotti, a phys-icist at AT&T Bell Laboratories and amember of the Vest panel, says he andother panel members were impressedby the management team and by the at-tention that NASA is giving to how thestation will ultimately be used.

The dream team may have taken tothe Þeld too late. Several well-placedcongressional aides believe sentimenthas now turned against Alpha. The newchairman of the House appropriationscommittee, who will hold decisive pow-er over the program, is RepresentativeDavid R. Obey of Wisconsin. Obey lam-basted the space station in 1993 as a

Òßying turkey.Ó And a prominent stationsupporter, Representative Jim Bacchusof Florida, has said he will retire, whichhas increased uncertainty still more.

If President Bill Clinton were to mounta full-court press for the space stationÑand if there are no more alarming de-velopments in RussiaÑthe programmight yet survive. But John H. Gibbons,ClintonÕs science adviser, told an ap-propriations subcommittee earlier thisyear that if he had to make a choice foradditional budget cuts, he would Òrath-er look harder at NASA than look atfunds for the National Science Founda-tion.Ó If GibbonsÕs remark is a clue tothe administrationÕs priorities, then Al-

pha may follow the route to oblivionthe Superconducting Super Collidertook a year ago. ÑTim Beardsley

Immuno-LogisticsMoving vaccines from thelab to the bush and the street

The six major vaccines adminis-tered every year to some 110 mil-lion of the worldÕs children who

are fortunate enough to get themÑwhether in Mogadishu or inner-city Mi-amiÑcost more to deliver than tomake. The United Nations ChildrenÕsFund, for example, spends a total of$1.50 per child on the vaccines pro-duced for diphtheria, pertussis, teta-nus, poliomyelitis, measles and tuber-culosis. That amount is a tenth of whata government then has to disburse forlabor, transportation, training and re-frigeration to get these vaccines to in-fants and young children.

Streamlining the way vaccinations areadministered would bring about moresavings than would measures to lowerthe cost of these commodity vaccines.ÒItÕs an almost perfect use for the sci-ence and technology coming out of thebiotechnology industry,Ó says AnthonyRobbins, who is helping to develop aClinton administration program to im-munize all children born in the U.S.against a battery of diseases.

Such optimism colors the planning ofthe ChildrenÕs Vaccine Initiative (CVI),an eÝort by leading international orga-nization programs to spur developmentof a supervaccine that could make im-munizations more eÛcient and lesscostly while extending coverage to indi-viduals exposed to such endemic dis-eases as malaria and shigellosis. TheCVI has set the ambitious goal of devis-ing a vaccine that can be trickled into ababyÕs mouth in a single dose shortlyafter birth. This ideal vaccine should re-


INTERNATIONAL SPACE STATION ALPHA is depicted in its latest incarnation, withelements from the U.S., Europe, Canada, Japan and Russia.

NA

TIO

NA

L A

ER

ON

AU

TIC

S A

ND

SP

AC

E A

DM

INIS

TR

AT

ION


sist spoilage in the heat of the tropics.It should also incorporate additionaldisease protectionÑmalaria and hepa-titis B, for example. And, of course, itshould be cheap.

The CVI vision so far is just that. Itsannual budget for research, develop-ment and other expenses totals a mere$4 or $5 million, despite fund-raisingplans to bring in more than $300 mil-lion. But even in its embryonic state, theprogram, shepherded since 1990 bythe U.N. ChildrenÕs Fund, the U.N. Devel-opment Program, the Rockefeller Foun-dation, the World Bank and the WorldHealth Organization, has provided fo-cus to disparate laboratory activities.

The Benezech-Simpson Company inVion, France, and the Pasteur Institutefound that a vaccine made by combin-ing a live but weakened poliovirus withdeuterium oxide (heavy water) couldlast for a week at tropical temperatureswithout refrigeration. ÒDeuterium formsstronger bonds to the virus than doesthe hydrogen in water normally used inthis process,Ó says Philip K. Russell, aprofessor at the School of Hygiene andPublic Health at Johns Hopkins Univer-sity and an adviser to the CVI. ÒThemore rigid structure presumably slowsdown the degradation of nucleic acidsand proteins in the virus.Ó

The Walter Reed Army Institute ofResearch and its partner, Virogenetics,have developed a unique vaccine againstmalaria. Most trials with malaria vac-cines try to guard against a single stageof the life cycle of the parasite that caus-es the disease: either when it is Þrst in-jected into the human bloodstreamthrough the proboscis of a mosquito,or once it changes and multiplies in theliver, or when it reenters the bloodfrom the liver. This new vaccine shouldsupply a complete defense.

It works by bioengineering into thevaccinia virus the genes for seven ma-larial antigens, one or more of whichcorrespond to a life stage of the para-site. The technique of studding a singlevirus with genes for various antigens isalso being explored to induce immuni-ty against several diseases at once. Thevirus serves as a vector that, once in-jected, allows the genes to express anti-gens that provoke an immune response.

OraVax, a Þrm in Cambridge, Mass.,is collaborating with the University of

MarylandÕs Center for Vaccine Develop-ment on human trials of an oral vaccinethat uses an attenuated strain of Shigel-

la to promote immunization against thedysentery-causing bacterium.

CVIÕs vision will still need moldingwith hard currency; many small bio-technology research companies lack theÞnancial resources to develop a vaccinefor the commercial market. For theirpart, major pharmaceutical companiesmay show scant interest in research

and development on these new agents.Government help may be needed. A

National Academy of Sciences report,issued last year, recommended estab-lishment of a National Vaccine Author-ity. The authority would be empoweredto subsidize purchasing or even assistin development of vaccines spurned bythe private sector. If the governmentcould take a vaccine through late-stageclinical trials, the costs of developmentwould diminish for a large pharmaceu-


MALARIA VACCINES, such as this exper-imental one about to be administered toColombian children, might eventually beadded to the regimen of childhood vac-cinations. That goal would help realizethe vision of the international ChildrenÕsVaccine Initiative.

TIM

OT

HY

RO

SS

J.B

. Pic

ture

s



Little Winners Bioprospectors reap rewardsfrom Third World microbes

Strolling through a Costa Ricanbazaar, you trip over a sick dog.EÜuent drips from its inßamed

eyes. Do you hop away in alarmÑor col-lect the ßuid? Apparently the latter, ifyou happen to be a bioprospector. Themicrobe Acremonium kiliense, deposit-ed with a governmental agency for pos-sible patent claim, was gleaned fromjust that source.

Since penicillin achieved commercialsuccess near the end of World War II,drug manufacturers have been scour-ing the nooks and crannies of the earthfor microorganisms that produce med-ically active chemicals. Successes havebeen manifold. Microbes in Spanish soilharbored an anticholesterol agent mar-keted as the Þnancial blockbuster Meva-cor. Others from Argentina, Venezuelaand elsewhere have yielded the strepto-mycins and chloromycetins of cornerdrugstores.

Such pharmaceutical treasures usedto be literally free for the taking. Now

tical producer. A company might thenbe willing to take over the manufactur-ing of an otherwise unproÞtable prod-uct. ÒThereÕs a problem making theleap between a good research idea andsomething licensable,Ó says Jerald C.SadoÝ, director of the division of com-municable diseases and immunology atthe Walter Reed institute.

But pharmaceutical companies havenot embraced the idea of Uncle Sam as vaccine developer. ÒThe governmentdoes not have the people, the facilitiesor the researchers to do it,Ó says R.Gordon Douglas, Jr., president of Merck& Co.Õs vaccine division. Douglas pre-fers the option of vending technologyto developing countries. MerckÕs sale ofthe intellectual know-how for makinghepatitis B vaccine to the government ofChina exempliÞes the process, he says.

Vaccines may never prove as enticingas drugs for pharmaceutical makers.They can be more expensive to develop,and they may produce less incomeÑrepeated reÞlls not required. Govern-ments and international aid agenciesalso cannot aÝord pharmacy prices.

By one estimate, the annual world-wide vaccine market is about $2 billion.At $1.2 billion in revenues and climb-ing, Prozac, that existential elixir, mightovertake the sale of potions that savethe lives of millions of children everyyear. ÑGary Stix


developing nations, waking up to thefortunes that they have unknowinglyexported, are asking for some returns.

To bioprospectors, microbes have aparticular advantage over plants: theyare small. A Þstful of soil or a ßaskfulof sewage may contain thousands of or-ganisms that can easily be cultured in alaboratory. Besides, one-celled creatures,which like plants and insects are muchmore diverse in tropical countries, havetraditionally attracted less political at-tention. Most nations did not really careif within their boundaries you clippedthe hair of a spiny rat or rooted in ze-bra dung.

Now the Convention on Biological Di-versity has changed all that. Althoughmicrobes may ßourish in tropical coun-tries, Pat Roy Mooney of the Rural Ad-vancement Foundation International inOttawa, Canada, estimates that 86 per-cent of microbial culture collections arein the industrialized world. The Ameri-can Type Culture Collection (ATCC)Ñwhere all microbes under U.S. patent ap-plication, and many more, are deposit-edÑlists more than 3,000 bacteria andfungi collected in nations such as Pana-ma, Ethiopia and India. When the con-vention came into force in December1993, these microorganisms became ineÝect the exclusive property of the de-positor. The country of origin will havean uphill battle claiming any royaltiesthat a microbe may earn.

Increasingly, Southern Hemispherenations are feeling that they never quiteÞgured out the value of what they gaveaway. Besides the microbes, the 68 per-

cent of seeds in Northern gene banksthat were originally bred by farmers indeveloping nationsÑas well as plantsyielding a quarter of the worldÕs pre-scription drugsÑnow belong to indus-trialized countries. Comments Amir Ja-mal, former ambassador to the UnitedNations from Tanzania: ÒWeÕre realiz-ing rather late in the day that we soldourselves short.Ó

Traditionally, biological resourceshave been considered Òthe commonheritage of mankind.Ó In practice, thathas meant that industrialized countriestook freely of those resourcesÑand soldthem back to the Third World as com-mercial products. As a result, nationsbarely decades past colonialism aredrawing some uneasy parallels withtheir earlier history. Moreover, biotech-nology is a strategic science for the nextcenturyÑwith developed countries al-ready in the lead.

The biodiversity convention, whileruling that biological resources are thesovereign property of a nationÑon thesame footing as coal or oilÑalso re-quires that these resources be madeavailable to prospectors. But guidelinesfor such access will take some time to draft. ÒMost Third World countriesdonÕt quite know what, or how, to con-trol,Ó Jamal says. Nor is there any con-sensus as to the intrinsic worth of abioprospectorÕs gleanings.

ÒUntil you Þnd an active chemical en-tity, this stuÝ has no value,Ó points outDavid J. Newman of the National CancerInstitute, which screens soil, spongesand other natural substances for anti-

tumor and antiviral agents. On the oth-er hand, Walter V. Reid of the World Re-sources Institute in Washington, D.C.,calculates that if 3 percent royalties arepaid (typical for samples of unknownactivity) and if one assumes that achemical has a one-in-40,000 chance ofbecoming a drug worth $10 million, a soil sample with 100 new chemicalscould be worth about $500 to thesource country. Yet companies are notabout to pay such amounts up front.

Microbes, at least, are relatively freeof the intellectual-property questionsthat plague work on plants and seedsoriginally identiÞed or bred by nativecultures. Those companies that Þnd auseful microorganism and develop amarketable productÑa process oftencosting millions of dollars and lastingover a decadeÑcan take credit for thediscovery. Sometimes, though, the cred-it has to be shared. A strain of fungusreportedly used by Brazilian farmers tokill Þre ants, which cause hundreds ofmillions of dollars in crop damage inthe U.S., was patented by a scientist atthe University of Florida.

While the question of how much adeveloping country should get for itsraw materials is debated, drug manu-facturers are sifting through the soil intheir own neighborhoods so many timesthat they keep turning up the same mi-crobes. Adding incentive to collectingin tropical climes, the Food and DrugAdministration has made it harder forÒme-tooÓ drugs to be approved. Theseare medications that are very similar toa popular one, like the many variations

of penicillin on the market.Radically new organismsare needed.

Those prospectors whoyearn for jungle adventuremay be disappointed. Mi-crobes are as easy to pickup as diarrhea. (The sourceof one acquisition in theATCC is listed as Òstool ofan Iowa man who had re-cently been in Bangladesh.Ó)But until the terms of ex-change between geneticallyrich countries and techni-cally rich countries are set-tled, the more ethically in-clined are unsure of howto proceed. ÒMeanwhile,Ósays James D. McChesneyof the University of Missis-sippi, Òvaluable microbialresources are disappearingbefore we get a chance tolook at them.Ó Tropical for-ests, after all, are even rich-er than tropical sewers.

ÑMadhusree Mukerjee


CALCUTTA COURTYARD: How many patents can you spot in this picture?

RA

GH

UB

IR S

ING

H



THE ANALYTICAL ECONOMIST

When new leaders came intopower in eastern Europe andthe states of the former Soviet

Union, they found themselves saddledwith an unwanted inheritance: tens ofthousands of businesses, from shopsto steel mills that encompassed entirecities. Central planning and governmentcontrol of these myriad enterprises hadled to the economic disaster that sweptcommunist governments out of power.Yet transferring these state-owned com-panies to private hands is proving to benearly as much of a headache as wastrying to run them. ÒItÕs not just a puÝof smoke and a wave of the wand, andthe companies belong to the citizens,Órues Anthony A. Repa, an economic ad-viser to the Polish government.

The problems of privatization aresimple to state but tricky to solve. Mostcitizens have no money with which topurchase corporate shares, and manyenterprises are in such perilous Þnan-cial condition that few people wouldwant to buy them. Even solvent busi-nesses come encumbered with now ir-relevant assetsÑapartment complexesand sports stadiums among themÑthatmake valuation diÛcult.

Countries have taken various ap-proaches to shedding state property,according to economists at the WorldBank. In the territories of the formerEast Germany, a federal agency manag-es businesses and properties while try-ing to arrange their sale or return toformer owners. In Russia and in theCzech and Slovak republics, in contrast,citizens have received coupons withwhich they can buy shares in state-heldconcerns. In Poland, privatization hasfollowed several strategies, includingfocusing on foreign investors and oninvestment funds that act as proxiesfor citizens with coupons. Most of theemerging market nations have also tak-en a more informal approach to priva-tizing small businesses by selling themto their managers or by auctioning them.

Coupon schemes are popular becausethey avoid many problems that arisewhen state property is sold for cash.Nemat ShaÞk of the World BankÕs Cen-tral Europe Department lists some ofthese drawbacks in a recent report onthe Czech experience: foreigners caneasily outbid citizens; the few people

who have money can acquire the lionÕsshare of formerly public assets; andthe low price paid for Þrms because ofgeneral lack of capital distorts futuremarket patterns. But coupons entail adiÝerent problem. Unlike cash sales,they generate no government revenueand no capital for modernizing a Þrm.

At the same time, unless special pre-cautions are taken, coupon privatiza-tion may diÝuse corporate ownershipso widely that eÝective governance isimpossible. In Russia, some kind of gov-ernance has been preserved by reserv-ing a percentage of shares in each en-terprise for managers and workers. TheCzechs and Slovaks took another ap-proach, according to ShaÞk. They en-couraged the development of mutualfunds to which citizens signed overcoupons. In theory, the funds can keepa closer watch on each company thancan individual shareholders.

In Poland, this form of privatization,when it is implemented, will be some-what more intricate. CitizensÕ couponsbuy shares only in investment funds;the funds, in turn, purchase companiesand manage them. The share pricesrise and fall depending on the marketÕsassessment of the companies they in-vest in, thus providing incentives forcareful oversight. Only about 450 Pol-ish enterprises, out of a total of about9,000, will be subject to mass privatiza-tion, Repa points out. Other companiesare being groomed for foreign sale orhave been acquired by their managersin leveraged buyouts.

Although such subtle manipulationsof market structure may help provideeÝective oversight of companies andgive the public access to the capitalistdream of shared ownership, they donot necessarily address the problemsof actual restructuring needed for sur-vival. Almost all these businesses needcapital for new investment, and manyare burdened by the debts that piled

up in the old days of central planning.Firms that cannot survive in their

present form may be subjected to whatis called asset liberation (on Wall Streetit is known as Òrape, loot, pillageÓ). Dur-ing this process, auditors strip out prop-erty or businesses that might fetch agood price on their own and discardthe remainder. This tactic is explainedby the rationale that it is far cheaper forthe economy to shut a plant down andpay workers for two or three years un-til they can Þnd other jobs than it is topay people to lose money indeÞnitely.

Asset liberation is, however, a com-plex undertaking. The commingling ofgovernment and business produced in-numerable company towns throughouteastern Europe and the former SovietUnion. In these sites an electric powercompany might own not only genera-tors and oil reÞneries but housing forits employees, day care centers and ca-sinos. Some of these adjuncts can besold for a proÞt; others must be takenover by a state or municipality if theunderlying company is to be soldproÞtably.

Ironically, remnants of the politicalliberalization that marked the earlystages of the transition to market econ-omies may be hindering the last stagesof privatization. In their Þnal years, cen-tral authorities ceded ownership ofmany enterprises to provincial or mu-nicipal governments. Towns and citiesare even more strapped for resourcesthan are national governments, so theyare unwilling to bear the brunt of clos-ings that would eventually beneÞt theeconomy as a whole.

In one instance, the economic transi-tion seems to be occurring most suc-cessfully independently of state-led ef-forts. The private sector in Poland hasbeen growing rapidly, contributing about50 percent of the gross national prod-uct last year. ÒPoland is looking poten-tially like the economic tiger of the re-gion,Ó Repa notes. ÒThe growth of theprivate sector has been spontaneous. Ithas not been the result of privatization,but the result of a couple of million newbusinesses.Ó In fact, Repa explains, theprivate sector is encouraged by ongoingineÛciencies in the vestigial state-runsector: ÒIn a competitive market, it givesthem an advantage.Ó Whether jealous,unfrocked apparatchiks will permit thatadvantage to persist remains to be seen.ÑPaul Wallich and Marguerite Holloway

For Sale: One Country, As Is

The problems of privatization are simple to state but tricky to solve.


THE AMATEUR SCIENTIST conducted by John Iovine

Cows that produce pharmaceuti-cals in their milk and plants thatresist pests are only two of the

myriad beneÞts promised by geneticengineering [see ÒGrading the GeneTests,Ó by John Rennie, page 88]. Where-as endowing the genomes of cows andplants with desirable traits can be diÛ-cult, genetically manipulating bacteriais fairly straightforward. You can inserta gene for resistance to penicillin intothe bacterium Escherichia coli by fol-lowing the steps outlined here. Similartransformations often occur naturally,as in hospitals where antibiotic-resis-tant bacteria sometimes proliferate.

Because E. coli is already present inyour gut, there is little to worry about.But remember to adhere to the sterileprocedures described. They will ensurethat you do not inadvertently ingest thestuÝÑand that it is E. coli, and notsomething else, that is blossoming onyour petri dishes.

The essential trick to manipulating E.

coli genetically is to get the creatureÕssingle cell to think that a foreign geneis one of its own. In this case, we willgraft the penicillin-resistance gene ontoa plasmid from E. coli. Plasmids are cir-cular loops of DNA that exist in cellsindependently of chromosomes. Afterbeing altered, the plasmids are able toreenter the cell and thus help us sneakour gene into the bacterium. They thenreplicate within the cell and produceproteins, including the ones that willbring about our desired trait. Further-more, plasmids are propagated throughsuccessive generations. Thus, the geneis transmitted to all oÝspring, confer-ring its resistant properties to the en-tire colony of E. coli grown from the al-tered individuals.

In this experiment, we will constructthree types of recombinant plasmids,called pAMP, pKAN and pAMP/pKAN,by a process called ligation. By injectingthese into E. coli, we will confer resis-tance for ampicillin (pAMP), for kana-

mycin (pKAN) or for both ampicillinand kanamycin (pAMP/pKAN) to thebacterium. To see the results of our en-gineering, we will then grow colonies ofgenetically altered E. coli on platestreated with ampicillin and kanamycin.

For the materials needed in this ex-periment, I suggest purchasing a DNArecombination kit. In addition, you willneed an aquarium (or some containerof similar volume), a Bunsen burnerand a few odds and ends listed in thebox on the top of page 110.

First, we need to build an incubatorin which to grow our E. coli colonies. A20-gallon glass aquarium will do nicely;if you do not have one, use any contain-er of the same volume, such as a card-board box. Place the aquarium on itsside with the open end facing you. Tapea piece of plastic to the top of the aquar-ium so that it drapes down, coveringthe open end. You will need to lift theplastic up and out of the way to workinside the aquarium.

The optimal temperature for E. coliÕsgrowth is that of the human body, 98.6

degrees FahrenheitÑnot surprising, giv-en where it lives. To warm the incuba-tor, put in a standard incandescent lampenclosed in a can or small pail. I neededa 75-watt bulb to heat the incubator to92 degrees F (this is less than the opti-mal temperature but works Þne). Startout using a low-power bulbÑsay, 40wattsÑand measure the temperature af-ter 12 to 24 hours; increase the wattageif necessary. If the incubator becomeshotter than 98.6 degrees F, lower thewattageÑa bit cooler is better than abit hotter. If changing the bulb does notdo the job, insert a light dimmer anduse it to adjust the power to the lampand consequently the temperature.

It is a good idea to have the incuba-tor ready when the kit arrives. Open thekit and refrigerate the culture plates up-side down, along with the vials of plas-mid pAMP, pKAN and calcium chloride(needed for conditioning the bacteria).Freeze the vial of ligase/ligation buÝer/ATP. The other materials can be storedat room temperature. Just before start-ing the experiment, wipe the incubatorand the work area with a 10 percentbleach solution (made by adding onepart of Clorox liquid bleach to nine


Genetically Altering Escherichia coli

JOHN IOVINE has published numer-ous books on science and electronics. Hecurrently teaches a workshop on holog-raphy for novices at the Art Lab on Stat-en Island, N.Y. ESCHERICHIA COLI blossom on petri dishes in an incubating aquarium. Culture

tubes, pipettes and other materials await the experimenter.


parts of water) or a disinfectant, suchas Lysol. Also disinfect after each ses-sion and wash your hands with an an-tibacterial soap both before and after.Keep your work area spotless and dis-infect all the hardware, such as tubes,pipettes and transfer loops, by placingthem in the bleach solution after use.

Our next step is to incubate and growE. coli bacteria on a culture plate. Theculture plates contain Luria broth, orLB, which provides nutrients on whichthe bacteria thrive. We will need theseE. coli colonies for the rest of the exper-iment. Light the alcohol lamp or Bun-sen burner. Take one culture plate la-beled ÒLBÓ and mark on its bottom ÒE.

coli.Ó Write the date on it, too.Select the inoculating wire loop from

the kit and sterilize it by putting it intothe ßame of the lamp or the burner. Al-low the wire to get red-hot, then re-move it from the ßame and hold it fora few seconds to cool. Do not put theloop downÑthat will contaminate it.

Place the vial of E. coli culture in yourother hand and remove its cap with thelittle Þnger of the hand holding the in-oculating loop. With the cap removed,brießy pass the mouth of the vialthrough the ßame to sterilize it.

Lift the top of the LB plate markedÒE. coliÓ just enough to insert the inoc-ulating loop. Push the loop into theside of the jelly on the plate to cool it.Next drag the loop a few times throughthe E. coli culture in the vial. Removethe loop, pass the mouth of the vialthrough the ßame again and recap it.

Drag the loop across the jellylike sur-face, making a Z shape in one quadrantof the plate. Replace the lid.

Turn the culture plate 90 degrees, re-heat the loop and cool it by stabbing atthe gel away from the Þrst streak. Then

pass the loop once through the Þrststreak and make another zigzag shape.Repeat the turning, reßaming andstreaking another two times so thatthere are four zigzag shapes, one ineach quadrant of the culture plate. Re-place the lid on the plate.

Reßame the loop before putting itdown, so as not to contaminate thework space. Make this a habit.

Place the smeared plate upside downin the incubator to prevent condensa-tion that may collect on the lid fromfalling into the E. coli colonies. Incubatethe plate for 12 to 24 hoursÑno more.Then remove it from the incubator andallow the colonies to grow for one ortwo days at room temperature.

The next step is to link antibiotic-re-sistant DNA fragments with the E. coli

plasmids. The vials of pAMP and pKANcontain the indicated DNA fragmentsas well as the plasmids; a reagent calledligase inserts the DNA fragments andreseals the plasmid loops. (ATP in theligation solution provides energy forthe joining reaction.) The procedure ac-tually makes many diÝerent types ofhybrid molecules; however, only thosethat are properly formed will be main-tained and expressed in the cells.

Take the three vials containing 20 mi-croliters each of ligase/ligation buÝer/ ATP from the freezer. Label one vialÒ+pAMP/KAN,Ó another Ò+pAMPÓ andthe last one Ò+pKAN.Ó Take out thetubes labeled ÒpAMPÓ and ÒpKAN.Ó Us-ing a sterile needle-nose pipetteÑonefor each reagentÑadd 10 microliters ofpAMP and 10 microliters of pKAN tothe +pAMP/KAN vial; 10 microliters ofpAMP and 10 microliters of distilledwater to the +pAMP vial; and 10 micro-


RECOMBINANT DNA is made when a plasmid from E. coli and a circular piece ofDNA are cut into strips by enzymes and then rejoined. The ends are sealed by li-gase; only those recombinants that are properly oriented will survive in a cell.

E. coli PLASMID

RECOMBINANT DNA

DNA CONTAINING DESIRED GENE

Supplies for Genetic Alteration ExperimentE-Z Gene Splicer DNA Recombination and Transformation Kit. PN# 21-1160; $35.67

The kit includes:

Available from:

In addition to the materials provided in the kit, you will need:

One vial of plasmid pAMPOne vial of plasmid pKANCultured E. coliThree LB platesTwo LB/AMP platesTwo LB/KAN platesTwo LB/AMP/KAN platesThree vials of ligase/ligation

Alcohol lamp or Bunsen burnerAntibacterial soapAquarium or boxBeakers or bowls (two)Crushed iceDistilled water (small quantity)

One vial of calcium chlorideTwelve needle-nose pipettesEighteen one-milliliter sterile

transfer pipettesFour 15-milliliter sterile culture tubesFive sterile inoculating loopsGlass cell spreaderManual

Ethanol alcohol (70 to 95 percent)Felt-tip marker for labeling Household bleach, such as Clorox

(small quantity)Sheet of plastic for covering aquariumThermometer

Carolina Biological Supply Company2700 York Road, Burlington, NC 27215 (910) 584-0381

liters of pKAN and 10 microliters ofdistilled water to the +pKAN vial. Closethe tops of the vials and gently tap theirbottoms on the table to mix the re-agents. Incubate the vials at room tem-perature for 12 to 24 hours. Now the+pAMP/pKAN vial contains plasmidsfor resistance to both ampicillin andkanamycin, the +pAMP vial only thosefor resistance to ampicillin and the+pKAN vial only those for resistance tokanamycin.

Now we must make the E. coli cellsÒcompetentÓ to absorb the recombi-nant plasmids. This requires suspend-ing the E. coli cells in a cold calciumchloride solution and subjecting themto a brief heat shock at 107.6 degreesF. Just how DNA is absorbed by compe-tent E. coli cells is not known.

Prepare a water bath for heat-shock-ing the bacteria; you will need the bathjust once for 90 seconds. You can usean aquarium heater to bring the waterin a container to 108 degrees F (a cou-ple of degrees more or less should notmatter). In a pinch, just run in tap wa-ter, adjusting the temperature.

Get four sterile 15-milliliter tubes.Label them Ò+pAMP/KAN,Ó ÒÐpAMP/ KAN,Ó Ò+pAMPÓ and Ò+pKAN,Ó respec-tively. With a sterile transfer pipette,add 250 microliters of cold calciumchloride to each tube. Place the tubesin a beaker or bowl with crushed ice.

Pick up one or two colonies of E. coli

from the LB starter plate using a sterileplastic inoculating loop. Be careful notto take any gel from the plate. Immersethe loop in the calcium chloride solu-tion in the +pAMP/KAN tube and tapagainst the side of the tube to dislodgethe cell mass. Suspend the cells in thesolution by repeatedly pipetting in andout with a sterile transfer pipette. Re-turn the +pAMP/KAN tube to the ice.Transfer cell colonies to the other threetubes in the ice in the same way.

For each of the following steps, use aa fresh needle-nose pipette. Transfer10 microliters of ligated plasmid+pAMP/KAN from the appropriate vialto the +pAMP/KAN culture tube. Add10 microliters of ligated +pAMP to the+pAMP culture tube and 10 microlitersof ligated +pKAN to the +pKAN culturetube. Do not transfer any material intothe ÐpAMP/KAN culture tube; this lasttube should contain only unaltered E.

coli. Place all the tubes back in the iceand let them sit for 15 minutes.

After the ice incubation, it is time toheat-shock the E. coli cells. Remove allthe tubes from the ice and immediatelyimmerse them in the 108-degree F wa-ter bath for 90 seconds. Then return allthe tubes directly to the ice. Keep themthere for three to four minutes.

Using a sterile transfer pipette, add250 microliters of Luria broth to eachtube, to give your bacteria somethingto eat. Gently tap the tubes with yourÞnger to mix in the broth and incubatethe tubes at 98.6 degrees F for three tosix hours. With luck, the E. coli shouldnow be suÛciently shocked to absorbthe plasmids in their environment.

Finally, we can check to see if our E.

coli have indeed acquired resistance toampicillin and kanamycin. Some of theculture plates in the kit already containthe antibiotics to be pitted against ourE. coli and are marked as such. Theplates marked ÒLBÓ contain only Luriabroth; label one such plate Ò+LBÓ andthe other ÒÐLBÓÑon these plates we willgrow the altered and unaltered E. coli,respectively. Label one LB/AMP/KANplate Ò+pAMP/KANÓÑto this plate, con-taining both ampicillin and kanamycin,we will be adding the E. coli cells resis-tant (we hope) to both antibiotics. Labelthe other LB/AMP/KAN plate ÒÐpAMP/ KANÓÑonly unaltered E. coli bacteriaare to be added here. Label one LB/AMPplate Ò+pAMPÓ and the other Ò+pKANÓÑto these (ampicillin-treated) plates wewill be adding E. coli resistant to ampi-cillin and kanamycin, respectively. La-

bel one LB/KAN plate Ò+pKANÓ and theother Ò+pAMP.Ó

Add 100 microliters of the cell sus-pension from the ÐpAMP/KAN culturetube to the ÐLB plate and also to the ÐLB/AMP/KAN plate, using a steriletransfer pipette. Before smearing thecells over the surface of the gel, steril-ize the glass spreader. Dip the spread-er in the ethanol alcohol and ignite thealcohol using the Bunsen burner or al-cohol lamp. After the alcohol burns oÝ,the spreader is ready. Use it to distrib-ute the E. coli cells evenly over the gel.

With another sterile transfer pipetteÑuse one for each cultureÑadd 100 mi-croliters of the culture from the +pAMP/KAN tube to the +LB plate and also tothe +LB/AMP/KAN plate. Spread the cellsuspension as before, sterilizing theglass rod each time. From the +pAMPculture tube, add 100 microliters eachof the cell suspension to the +pAMPand +pKAN plates, then spread. Fromthe +pKAN culture tube, add 100 mi-croliters of the cell suspension to the+pAMP and +pKAN plates, then spread.

Allow the plates to gel for 10 minutes,then stack them and tape them togeth-er. Place the plates upside down in the98.6-degree F incubator. Incubate themfor 12 to 24 hours. Note that if the col-onies are overincubated, they will over-grow and become indistinguishable.

Now you are ready to see the resultsof your experiment. On the +LB and ÐLBplates, both the transformed and thenatural E. coli grow well. For the LB/ AMP/KAN platesÑcontaining both am-picillin and kanamycinÑthere is agrowth on the one labeled Ò+AMP/ KANÓ and none on the ÒÐAMP/KAN,Óshowing that only the E. coli trans-formed with both antibiotic-resistantgenes can grow. The LB/AMP plates il-lustrate that the gene for pAMP, andnot that for pKAN, confers resistanceto ampicillin. The LB/KAN plates illus-trate that the gene for pKAN, and notthat for pAMP, confers resistance tokanamycin.

By measuring the growth of the col-onies, you can see that the ligation oftwo genes, pAMP and pKAN, is morediÛcult than a single ligation: the+pAMP/+pKAN colonies are sparser(Þve to 50 colonies) than are the +pAMPor the +pKAN colonies. Also, ligationof the pKAN gene is more diÛcult (50to 500 colonies) than that of the pAMPgene (100 to 1,000 colonies).

Now that you have genetically al-tered bacteria, you can see that in prin-ciple they are quite simple to produce.In practice, genetic engineering can besomewhat problematic. For instance,cows and plants will need much biggerincubators.


ANTIBIOTIC-RESISTANT DNA hitches aride into an E. coli cell by hiding in analtered plasmid. Plasmids cannot enterthe cell (top) until it is made receptive(middle). On entering, the recombinantDNA is accepted as part of the geneticmaterial of the cell (bottom). It can pro-duce proteins and also propagate.

RECOMBINANT DNA

E. coliPLASMID

E. coli CELL

MAIN E. coliGENOME


BOOK REVIEWS by Philip Morrison

Albert en Famille

EINSTEIN LIVED HERE, by Abraham Pais.Oxford University Press, 1994 ($25).

Not even Albert Einstein remainsunrevised these iconoclasticdays, the attackers by now more

than a little shoddy. Thus, very welcomeis this book by the most thorough andjudicious of physicist-biographers, whowas himself EinsteinÕs young compan-ion during nine Princeton years offriendly colloquy on all the world. Abra-ham Pais has just assembled a wonder-fully readable look (no math) at thatgreat man and at ourselves viewing him.He oÝers a four-page Òmini-brieÞng,Óbut no more, on EinsteinÕs physics inthis short, often riveting collection ofpersonal but always well-supportedstudies. For example, a dozen pagessample just one Þle from the volumi-nous Einstein archives, the one Einsteincalled ÒDie Komische Mappe,Ó perhapsbest translated as the Òcuriosity Þle.Ó Inthat Dickensian richness of letters tothe famous man are the touching, theenraged and the funny: ÒMy thirteen. . .talking cats have just told me what thefourth dimension is.Ó

Timeliest of the short pieces is theopener. It examines Einstein en famille,

with the facts that constrain a new Òmi-nor industryÓ that has arisen around aSerbian biography of Mileva Maric, Ein-steinÕs Þrst wife and mother of his threechildren. The book, In the Shadow of

Albert Einstein, promotes the claim thatit is Mileva, AlbertÕs fellow student ofphysics, to whom we really owe the rel-ativity that arrived in the wonderful year1905. Pais explains how a misunder-standing of one Russian citation mayhave been the taproot of the wistful as-cription so long after the fact. In 1905Einstein signed his startling manu-scripts with his new wifeÕs family nameadded after a hyphen, then a custom inSwitzerland. But there was only one au-thor. Mileva herself never made anyclaim whatever for a share of AlbertÕsfame during all three decades of openestrangement.

The delightful letters between theyoung pairÑmore are yet to be pub-lishedÑare full of pet names, largelyout of AlbertÕs native Swabian dialect.Their Þrst child, always called Lieserl,was born out of wedlock somewherenear MilevaÕs home in Serbia, sometimearound the turn of 1901 to 1902. Thecouple were married a year later in Bern.The infant Lieserl never lived with themthere; Ònothing I have seen indicates thatEinstein ever even saw her.Ó In 1903 Mi-

leva went east again to visit her family;she was pregnant with their Þrst son,Hans Albert. Einstein wrote her fromBern in concern over some report aboutLieserlÕs coming down with scarlet fe-ver. No later reference to the daughteris known in all the correspondence, andserious eÝorts on the spot have turnedup no other documentation.

The couple became estranged step bystep over 10 years. After Einstein be-came professor in Berlin, Mileva, ailingand melancholic, took her two boys oÝto Zurich, where she would live from1914 until her death. She and Einsteinwere Þrst formally separated, then di-vorced in 1918. Little doubt remainsthat the Þrst marriage had been dark-ened by AlbertÕs relationship with hiscousin, Elsa. Three years older than Al-bert, she became his second wife in1919, bringing her two daughters withher. It was something of a mismatch;the scholar who rejected all poses cameto resemble Òa bohemian guest in a mid-dle-class homeÓ; ElsaÕs apartment withits Þne carpets, furniture and pictureswas far better suited to her and hergirls than to her new husband. ÒI knowfor a fact that in the early 1920s Ein-stein developed a strong attachment toa young woman,Ó an extramarital aÝairthat endured for a couple of years. It ap-


ALBERT EINSTEIN talks to reporters about atomic energy. From the December 29, 1934, Pittsburgh Post-Gazette.

«

EM

ILIO

SE

GR

È V

ISU

AL

AR

CH

IVE

S, A

ME

RIC

AN

INS

TIT

UT

E O

F P

HY

SIC

S


pears, moreover, not to have been thelast of such aÝairs, as more than onenew book will eagerly magnify for you.

Elsa died in Princeton in 1936. ÒI . . .live like a bear in a cave,Ó Einstein wrotethen, the more because of Òthe death ofmy comrade, who was more attachedto people [than I].Ó Einstein had insightinto his marriages; he had not been agood husband, and he knew it. To livein Òlasting harmony with a womanÓwas Òan undertaking in which I twicefailed rather disgracefully.Ó

The Þrst son, Hans Albert, became asuccessful professional engineer in Eu-rope, then served from 1947 to 1971as a prominent professor of hydraulicengineering at the University of Califor-nia at Berkeley. Deeply attached to themother with whom he had grown up,Hans was not always on good termswith his father : ÒProbably the only proj-ect he ever gave up on was me. He triedto give me advice, but. . . I was too stub-born.Ó Near the end of his fatherÕs life,he visited in reconciliation. Hans diedof a heart attack at age 69 and liesburied on MarthaÕs Vineyard; the gravemarker recalls Òa life devoted to hisstudents, research, nature, and music.Ó

The younger son, Eduard, called Tete,born to Mileva and Albert in 1910, suf-fered from schizophrenia, his illnessclearly visible only after he had Þnishedhigh school admirably. Thereafter helived precariously in and out of institu-tions, until in 1965 he died in a psychi-atric center near Zurich. Poor Milevahad died in 1948 after long ill health.AÜicted for years by EduardÕs tragedy,she became lonely, even paranoid, dy-ing with a small fortune in Swiss francsstuÝed into her mattress. One Einsteindwells quietly in Bern still, Hans Al-bertÕs son, Bernhard.

Almost half the book is an engaginginquiry into the lifelong interaction be-tween Einstein and the media, whereÒthe show goes on.Ó Professor Paisseeks to answer how Einstein becameso singularly famous that for decadeshis mere presence ensured an intolera-ble crowd of gazers. He was a prisonerof that disturbance and could hardlyleave his Princeton neighborhood inany public way. There is no doubt theturmoil began right after World War I,when the British mounted an eclipseexpedition to seek EinsteinÕs predicteddeßection of starlight by the sun. TheeÝort glowed worldwide as a sign ofhealing after a terrible war and as apromise of a new order.

An air of mystery about EinsteinÕsÒspace-warpÓÑthe very words seemedat once simple and paradoxicalÑsoonsurrounded the man. In 1921 he told aDutch newspaper : ÒI am sure that it is

the mystery of non-understanding thatappeals to them.Ó A reader sees thepoint but is not quite persuaded. Any-way, it wasnÕt his photogenic looks: inthe 1920s Einstein was still a Òfriend-ly . . .pot-bellied gent, dressed in a . . .bourgeois way.Ó No sweatshirts yet.Among the two dozen images repro-duced, many are fresh and striking: hisÞrst childhood photograph, the 1903wedding picture, the two young sons.

That formula become a logo, E = mc2,has a more understandable rise, theconsequence of a pedagogic inventionby physicist Harry Smythe as WorldWar II ended. He cited it on the Þrstpages of his celebrated oÛcial 1945 re-port on the secret Manhattan Project,setting out EinsteinÕs relation as thekey to the release of nuclear energy.Good physics, though quite misleadinghistory, the idea penetrated beyond anyteacherÕs dream, until now, alas, it isless an explanation than a mantra.

One Þrm judgment demands repeat-ing. Einstein was deeply honest, unpre-tentious and caring for the future, buthe was never politically naive. His can-dor looks the better as time passes,bringing hope for compassion and rea-son here where Albert Einstein lived.

The Whale as Hot-Oil Balloon

AIR AND WATER: THE BIOLOGY AND

PHYSICS OF LIFEÕS MEDIA, by Mark W.Denny. Princeton University Press,1993 ($39.50).

In 1988 this Stanford biologist, atalented reverse-engineer of life,from spiderwebs to surf-swept kelp

forests, was enlisted to lecture his fel-low zoologists on the distinctions be-

tween water and air. Original, straight-talking and enthusiastic, Denny hasbuilt out of that experience an unusualbook for biology students and any seri-ous readers who have even a rustygrasp of freshman math.

ÒEvery day I Þnd a new.. .biologicalexample of physics in action.Ó The con-tent and heft of his book therefore dulyrecall a pretty full physics text. The ear-ly chapters open with units, NewtonÕslaws, stress and energy. The next fewelaborate on ßuid pressure and ßow.DiÝusion, less commonplace, receives agood account. Then we read on to heat,sound, optics, waves. There are plentyof equations, though not much math,for intermediate steps are fully present-ed. Graphs and tables of the most var-ied data abound.

All the results have one purpose, aquantitative understanding of the waysof life in two ßuids. Biologist Denny in-tends his readers to build coherentphysical intuition; formulas are indis-pensable but insuÛciently nourishingfor that. The book tilts to clarity and toexamples; derivations that are physicalare given pride of place, others omit-ted. The wave equation is absent; in itsstead a careful physical derivation ofsound speed is oÝered. Random walkand the physics of diÝusive transportare treated without the heavy statisticalweapons.

Life is many-faceted, its systems di-verse. That keeps interest high and as-sures the width of the account. Relativ-ity, quarks and other extensions ofphysics far beyond everyday experi-ence are forgone; you canÕt have it allin one book. But what you do have is awell-organized process that analyzes incontext about 100 substantial issues ofhow and why life works as it does inwater and in air, as a hopeful designermight address them. The action is inscale, form, motion, energy, Þeld, detec-tion and simple chemistry. Information-al issues are oÝstage; necessity is herewell ahead of chance. We are amongphysiologists, newly wide-ranging ones.A sample of topics follows. Denny worksthem out in Þrst approximation, nonumber crunching, unabashedly willingto invoke the spherical organism.

Since water-dwellers match waterdensity rather well, could they adjustto ßoat or sink by raising or loweringbody temperature slightly, on the mod-el of the hot-air balloon? A thin-walledsea ÒballoonÓ would need only a frac-tion of a degree rise to equal the densi-ty change its airborne counterpart gainsonly by heating internal air through ascorching 100 degrees Celsius. But theenergy cost is high, losses strong with-in dense water. So there are no living


CARTOON by Herblock appeared in theWashington Post several days after Ein-steinÕs death on April 18, 1955.


hot-air or hot-water balloons. There isone hot-oil balloon: surprisingly, it isthe huge sperm whale. Its unique sper-maceti organ holds two or three tons ofa special lipid mix, whose density fallsas its temperature rises 10 or 20 timesas much per degree as does the densityof water. Fifty-ton Moby Dick can ad-just buoyancy at will by 100 pounds orso, simply by heating or cooling slight-ly his headful of sperm oil, a task forwhich he has several means in place.

Corals deep in the reef feel strongbuoyant forces and little gravitationalload. Yet redwoods grow 10 times ashigh as corals. The answer is dynami-cal : steady ßow produces forces of liftand drag on solid bodies (well workedout here), but there are also reactionforces from accelerating ßow of ambi-ent ßuid. Reaction forces limit thelength of a rooted column in propor-tion both to the ßuid density and to itsmaximum acceleration. Accelerationsin the surf reach measured values near40 gravities. If trees endured the samereaction limit, they could withstand su-perhurricane accelerations of 3,000gravities. Plainly these never occur inthe atmosphere. Trees are not limitedin height as corals are by major un-steady ßow but mainly by their insta-bility to buckling under their ownweight. Talk of this with any sailor!

A physicist reader Þnds this workevokes pride for his science and envyfor the biologists who raised these newproblems and then found out how tosolve them simply. It deserves manyreaders, and even more it would gracethe shelves of long-term browsers.

Of Flies and Men

HUMAN LONGEVITY, by David W. E.Smith. Oxford University Press, 1993($35).

Pathologist and molecular biolo-gist, David W. E. Smith of North-western University has for most

of a decade centered his wide interestson human longevity. He synthesizeshis conclusionsÑand his doubtsÑinthis brief, firm and eye-opening book.The facts of life and death are neverstraightforward, even when they are ac-counted for numerically. Their mean-ing is treated widely, within medicineand human pathology, in the biology ofother forms of life, in the sciences ofhuman behavior and in human evolu-tion. Five chapters discuss these topics,one focused on the intrinsic and ubiq-uitous diÝerence between longevity inmen and in women. Everywhere (savein South Asia and Iran), the men die

younger even though women are sick-er, perhaps to a degree from medicalinattention. The three Þnal chapterstreat evolution and oÝer a forecast ofthe future.

Sooner or later we animals all die,but that is not so for all individual liv-ing cells. The Þssion of bacteria pro-ceeds indeÞnitely two for one. Then,the saying goes, Òwith sex came death.ÓCells like our own show clonal senes-cence: a cell and its progeny do notcontinually divide. They need a newdeal, some meiotic dance. Many humancell lines have been cultivated in tissueculture until growth stopped, after 50doublings more or less. But Òcancercells are immortal.Ó Tumor cell lineshave been maintained for decades, stilldoubling every day or two.

What dies is an organism, an interde-pendent society of cells, human or fruitßy. The little ßy passes egg and larvalstage, to reproduce and die after a fewweeks as adult. The causes of ßy deathare not fully known, but often it is thecells that line the active digestive tractthat stop working, clogged by a pig-ment derived from massive membranebreakdown. ( In mammals, unlike theßies, the digestive system cells are ca-pable of replacement.) Then the ßy neu-rons, never plentiful, drop graduallyout of action. ÒOld ßies die constipatedand largely immobilized.Ó

The curve of dying gives statisticsdecisive for ßy or human. Follow a sam-ple of our kind and count deaths yearby year. In developed lands today, mostlives come to an end within a rather nar-row band of ages, the rate peaked nearthe expected average lifetime. Smallbirds and rodents, as well as Òdrinkingglasses washed by machine in a restau-rant,Ó show an utterly contrasting pat-tern. They die by random causes, deathtaking a near-constant fraction of sur-vivors left at any age, until they haveall gone.

A more complex intermediate patternbegins with a high death rate in earlylifeÑinfant mortality. The rate thenfalls steeply, ßattens out, to rise againslowly, then more and more rapidly toa wider, earlier peak, until the sampleruns out. The change from intermedi-ate to modern form is seen in action inthe curves shown for the U.S. popula-tion at 20-year intervals since 1900. By1980 the infant death rate is very muchsmaller, and the high, narrow end peakdominates. The developed countries,and a few others as well, have enteredthe late-peak phase. (This is only partof the demographic transition; full pop-ulation studies must of course take ac-count of births, too.)

It is unlikely that our species ever had


a survival pattern as random as thatfor machine-washed tumblers. A natu-ral maximum age seems to be built inamong species that have larger brains,mature late and care for their young,particularly those, like primates, withfew multiple births. ÒThe human spe-cies is unique,Ó the farthest of all outalong that axis. A maximum life span isnot merely statistical; humans live twiceas long as other large mammals. Thepresent record human age was reachedby a man of 120 years (with only atouch of doubt about its validity).

In the U.S. now there are 30,000 cen-tenarians among the 1 percent of usÑ2.5 million peopleÑwho are older than85 years. These people are real outly-ers, not merely there on the tail of asmooth mortality curve but distinct.(One mouse from a particularly well-cared-for strain outlived all his kinfolk,too, by about 20 percent.) It is probablethat we have been a long-lived speciesfor 100,000 years, even if death camemuch earlier to most. Turtles and clamsand redwoods belong in another cate-gory. Indeed, the author does not men-tion the curious diÝerences even amongmammals in the rate of their life pro-cesses; a mouse heart beats during amouse lifetime about as many times asan elephantÕs does, but in only a fewyears. People much outlast them all intotal pulse count.

We all know that the chief agents ofdeath are no longer random infectionsthat may enter at any age, but long-de-layed functional failures in heart orbrain, where many cells cannot be re-placed, just as in the fruit-ßy gut. Doesour end resemble the aged ßyÕs? Hard-lyÑour failures show no steady loss ofcell by cell until too few are at work butrather an acute loss of whole clustersof cells whose blood supply has ceased.

If there is a simple picture to bepainted, it could be that the constella-tion we fearÑischemic heart disease,cerebrovascular calamity and the com-mon malignant tumorsÑmay dependon the varied human diet and our com-plex hormonal management of choles-terol. Wear and tear, essentially incre-mental functional loss, may bring theend in diverse ways, once the agingfeedback responses encounter challeng-es they can no longer meet.

But the story we know today is notcomplete. A ÒscientiÞc rationale mayemerge,Ó able not only to delay mortal-ity as we can now often do but even topostpone the senescence that causesthe disabilities of aging. That optimismends, a little unexpectedly, the authorÕssobering insights. Just where we shouldlook in the long genomic message hecannot tell.


At seminars and conferences, phys-icist Wolfgang Pauli was famous for dismissing work he particu-

larly disliked as Òganz falsch,Ó complete-ly false, or, yet more damning, Ònot even

false.Ó Pauli was not the only practition-er of the deft put-downs the British callpinking. As a group, scientists show asurprising historical talent for the in-sult. Beneath the impersonal surface oftheir formal scientiÞc discourse lies avast subtext of bloodletting.

Isaac Newton, for instance, was ascathing polemicist who honed his at-tacks on Robert Hooke and GottfriedWilhelm Leibniz to razor-sharpness insuccessive drafts, each of them, accord-ing to a biographer, more oÝensive thanthe last. ÒMr Hook thinks himselfe con-cerned to reprehend me for laying asidethe thoughts of improving Optiques byRefractions,Ó Newton wrote in one sal-vo. ÒBut he knows well yt it is not forone man to prescribe Rules to ye stud-ies of another, especially not withoutunderstanding the grounds on wch heproceeds.Ó When Leibniz published hiscalculus without acknowledging whathe knew of NewtonÕs progress, Newtonattacked, sometimes through John Keill(Johann Bernoulli referred to Keill asÒNewtonÕs toadyÓ), sometimes througha committee of the Royal Society. Heeven tucked one beautifully compressedaccusation of what was then known as plagiary into the mathematical foot-notes of Commercium epistolicum:

ÒThus the method which earlier he[Leibniz] wanted, asked for, received,and understood with diÛculty, he dis-covered forsooth. . . .Ó

Priority fuels the bitterest assaults,but ignorance, particularly when scien-tists move outside their area of exper-tise, also provides a rich source. For in-stance, the pugnacious Hermann Kolbe,a major 19th-century research chemistwith an unfortunate lack of interest inoptical activity, used the Journal f�rPraktische Chemie to ridicule JacobusHenricus vanÕt HoÝ, who had clariÞedthe relation between optical activity andmolecular structure in a brilliant pam-phlet. ÒA Dr J. H. vanÕt HoÝ of the veteri-nary school at Utrecht, Þnds, as it seems,no taste for exact chemical investiga-tion,Ó Kolbe wrote. ÒHe has thought itmore convenient to mount Pegasus (ob-

viously loaned at the veterinary school)and to proclaim in his La chimie dans

lÕ�space how during his bold ßight tothe top of the chemical Parnassus, theatoms appeared to him to have groupedthemselves throughout universal space.ÓVanÕt HoÝ had his revenge; he reprint-ed KolbeÕs jibes in the second editionof La chimie, guaranteeing KolbeÕs his-torical reputation as a fool.

Chauvinism, too, gives impetus tothe scientiÞc put-down. The French sci-entist Marie Jean Pierre Flourens saidin his review of The Origin of Species,

Ò . . . [W]hat unclear ideas, what falseideas! . . . O lucidity! O French stabilityof mind, where art thou?Ó And theSwedish chemist J�ns Jakob Berzelius,who took to studying textbooks otherthan his own when conÞned to bed bygout, wrote, ÒThe chemists of Englandlive in their own world. . . . There is agreat deal of litigation here about pri-ority in the most petty matters. . . . onecan regard them as one would puppieswho stand and snarl over their bones,from which the meat has been gnawedon the continent.Ó

Singularly unprophetic remarksabound in the history of science,from the comments of Charles

DarwinÕs doctor father to his son (ÒYoucare for nothing but shooting, dogs, & rat-catching, & you will be a disgraceto yourself & all your familyÓ) to those of Albert EinsteinÕs headmaster whenasked what profession Einstein shouldfollow (ÒIt doesnÕt matter; heÕll nevermake a success of anythingÓ). ElihuThomson, electrical engineer of note,visited Thomas A. EdisonÕs workshopand told the newspapers that he Òdidnot think very highly of the Edisonlamp and expected no great future forit.Ó J. Louis Agassiz, Harvard Universityprofessor and naturalist, maintainedthat he would Òoutlive this maniaÓ ofevolution, which he characterized inthe American Journal of Science andArts as Òa scientiÞc mistake.Ó

Press conferences eventually sped upthe tempo of scientiÞc sniping, inÞght-ing and backstabbing once conductedin the main through debates, letters,symposia and book reviews. Edisonused the newspapers in his well-publi-cized Þght to fend oÝ the new AC (al-

ternating current) distributing systemsthat were far more powerful than hislimited DC (direct current) stations. Inwhat became known as the ÒWar of theCurrents,Ó Edison invited reporters andguests in daily throughout 1887 towatch the stray cats and dogs of WestOrange, N.J., totter onto tin sheets andbe electrocuted by high-tension AC cur-rents. Red-lettered pamphlets warnedthat Òpatent piratesÓ like George West-inghouse were bent on introducing thesehazardous currents Òinto the Americanhome.Ó Despite WestinghouseÕs con-stant explanations that AC was receivedonly at low voltages because of Stanleytransformers, Edison and his colleagueswere so successful in arousing the pub-lic that the eponym Òto WestinghouseÓwas suggested as an alternative to Òtoelectrocute.Ó

In the present age of litigation, scien-tists are usually more circumspect thanEdison. They consult attorneys beforethey talk, as did Steven E. Koonin, thenchairman of the nuclear physics divi-sion of the American Physical Society,before speaking out on cold fusion at ameeting of that society in 1989. Ad-vised to avoid the ÒF-wordÓ (fraud),Koonin summed up Stanley Pons andMartin FleischmannÕs cold fusion workwith the deadly phrase Òincompetenceand perhaps delusion.Ó Nathan S. Lew-is, who also spoke, provided a list ofpointed questions the press might wantto ask Pons and Fleischmann, addingthat if the two scientists were going tohave publication by press conference,he would institute peer review by pressconference.

Not all put-downs are scathingÑsome small part of them manages tobe aÝectionate, as was the comment ofan Edinburgh professor on an eveningspent with Charles Babbage (ÒIt waswith the greatest diÛculty that I es-caped from him at two in the morningafter a most delightful eveningÓ). Someof them are even meant to be compli-ments, although, of course, they alsopink. For instance, Pauli, already a mas-ter of dis-ing when only a graduate stu-dent, attended a seminar given by Ein-stein and generously acknowledged,ÒYou know, what Professor Einsteinsays is not so stupid.Ó

ANNE EISENBERG teaches at Poly-

technic University in Brooklyn, N.Y.

ESSAY by Anne Eisenberg


The Art of the ScientiÞc Insult


Documents

The Eyethe-eye.eu/public/concen.org/Scientific American 1993...physician, and his mother devoted themselves to all aspects of culture. By his early teens Weil had become Òpas-sionately